The DOI matches its Crossref record; 218 of the 315 references it cites were found in a registry or an index, and 1 are not checked yet; 4 of the 278 sources its claims cite were checked for support.
Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks
Open accessAbstract
Artificial Intelligence (AI) techniques of deep learning have revolutionized the disease diagnosis with their outstanding image classification performance. In spite of the outstanding results, the widespread adoption of these techniques in clinical practice is still taking place at a moderate pace. One of the major hindrance is that a trained Deep Neural Networks (DNN) model provides a prediction, but questions about why and how that prediction was made remain unanswered. This linkage is of utmost importance for the regulated healthcare domain to increase the trust in the automated diagnosis system by the practitioners, patients and other stakeholders. The application of deep learning for medical imaging has to be interpreted with caution due to the health and safety concerns similar to blame attribution in the case of an accident involving autonomous cars. The consequences of both a false positive and false negative cases are far reaching for patients' welfare and cannot be ignored. This is exacerbated by the fact that the state-of-the-art deep learning algorithms comprise of complex interconnected structures, millions of parameters, and a 'black box' nature, offering little understanding of their inner working unlike the traditional machine learning algorithms. Explainable AI (XAI) techniques help to understand model predictions which help develop trust in the system, accelerate the disease diagnosis, and meet adherence to regulatory requirements. This survey provides a comprehensive review of the promising field of XAI for biomedical imaging diagnostics. We also provide a categorization of the XAI techniques, discuss the open challenges, and provide future directions for XAI which would be of interest to clinicians, regulators and model developers.
Verified 2026-10-07 · 1 reference not checked yet
Overview
AI snapshot
Written from the paper’s full text.
TL;DR
This survey reviews explainable AI (XAI) techniques for deep neural network–based biomedical imaging, categorizing methods and mapping them to imaging modalities and medical specialties. It finds Grad-CAM is the most widely used XAI method, while evaluation standards, definitions, and clinical validation remain open challenges.
Key findings
- The survey categorizes XAI techniques into visualization (perturbation-based and gradient/backpropagation-based) and explanation by example, further classified by scope (local/global), applicability (model-specific/model-agnostic), and usage (intrinsic/post hoc).
- Across biomedical imaging, Grad-CAM is the most common and preferred XAI method, followed by LRP, LIME, and SHAP; use is increased in neurology, respiratory, and ophthalmic studies, with MRI dominant in neurology, CXR in respiratory, and fundus images in ophthalmology.
- Reported modality-specific performance examples include COVID-19 CXR with EfficientNet-B7 at 95.01% accuracy and AUC 0.950; COVID-19 CT/CXR with VGG-19 at 95.61% accuracy; low-quality CT with ResNet-50 at 99.87% accuracy and 99.58% sensitivity; Parkinson’s disease 3D CNN at 95.29% accuracy and ROC-AUC 0.98; neonatal MRI brain dysmaturation at 98.5% accuracy; coronary CTA atherosclerosis AUC 0.91; and coronary artery calcium detection AUC 0.9627 with 0.9272 accuracy.
- Evaluation of explainability remains unsettled: the paper states there is no agreement on measuring or evaluating explainability, and it reviews quantitative metrics such as deletion/insertion, CLEVR, and MSFI alongside qualitative human evaluation.
Methods
- Literature survey; papers were selected by searching Web of Science, ScienceDirect, IEEE Explore, and Google Scholar for articles on explainable, interpretable, and trustworthy XAI techniques for biomedical applications.
- No primary experiments or new data are reported; the paper synthesizes XAI techniques, tools/frameworks, evaluation methods, and applications grouped by imaging modality and medical specialty.
- Techniques are categorized by scope, applicability, and usage/nature, with summaries provided in tables.
Limitations
- Stated: there is no universally agreed definition of “explanation”; no agreement on evaluation metrics; lack of algorithm standardization; most explainability techniques are visualization-based and qualitative; current definitions lack consistency and may represent a false hope for patient-level decision support; model-highlighted features may differ from human-identified features.
- Stated: the paper calls for inherently interpretable and accurate models, human-in-the-loop systems, regulatory benchmarks, and federated learning, implying these are not yet resolved.
- Evident from the text: the survey does not report a systematic review protocol, search dates, inclusion/exclusion counts, or formal quality appraisal, and it does not independently validate the XAI methods it reviews.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
250 Not checked yet · 30 Own interpretation · 9 No basis given · 4 Own result · 3 Partly supported · 1 Supported · accuracy not yet measured
Every source the paper cites for this claim was checked and supports it.
In the paper
“The biases in the algorithms, data, or humans can result in misdiagnosis of certain patient groups further compounding the inequalities [39].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.1. DNN in medical applications
What the sources say
- Supportedcites [39] Norori, 2021, Addressing bias in big data and AI for health care: a call for open science, Patterns source
The Bias section identifies imbalanced data and human prejudices as sources of algorithmic bias, and the AI and bias in medicine section says biased algorithms can increase health inequities. That section also reports lower diagnostic accuracy for Black patients and a risk of misdiagnosis, supporting the claim's link between bias, misdiagnosis of patient groups, and inequality. (LLM, conf=0.92)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“The image pre-processing using contrast-to-noise ratio, removing irrelevant regions etc. improved the DNN performance [267].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Partly supportedcites [267] Heidari, 2020, Improving the performance of CNN to predict the likelihood of COVID-19 using chest X-ray images with preprocessing algorithms, Int. J. Med. Inf. sourceExperimental — accuracy not yet measured
The Image preprocessing section describes removing the diaphragm region, reducing noise with bilateral filtering, and normalizing contrast with histogram equalization. The Discussion reports improved CNN classification accuracy from these preprocessing approaches. However, the paper does not describe using a contrast-to-noise ratio; that specific method is not supported. (LLM, conf=0.87)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The most used explainable technique for X-rays seems to be Grad-CAM, as evidenced by Table 4.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: unclear — It is unclear why the paper says it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Again, Grad-CAM seems to be the most favored approach as evidenced by Table 4.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
hedgedcorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“There are different XAI techniques used for different medical imaging modalities but Grad-CAM seems to be the most widely used as evidenced by Table 4.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
hedgedcorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Although use of Grad-CAM and other XAI techniques have been reported in Table 4 but there is not a predominant single XAI technique used across the various imaging modalities.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.4. Ultrasound imaging
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“However, the most common and preferred XAI method across the various medical imaging modalities seems to be Grad-CAM (Table 4), followed by LRP, LIME and SHAP and other methods.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.1. XAI applications in medical imaging
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“The explainability techniques are mostly based on visualizations that provide a qualitative indication by highlighting the significance of the image areas contributing to a model's decision.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.1. XAI applications in medical imaging
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: unclear — It is unclear why the paper says it holds.
In the paper
“Hence, false negative error may be costlier compared to a false positive in relation to medical image classification [37].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.1. DNN in medical applications
What the sources say
- Not checked yetcites [37] Singh, 2021
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Convolutional Neural Networks (CNN) are state-of-the-art models for image segmentation and classification [2], although encoder-decoder transformer architectures have also been proposed [3][4][5].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 1. Introduction
What the sources say
- Not checked yetcites [2] Barredo Arrieta, 2020, Explainable Artificial Intelligence (XAI): concepts, taxonomies, opportunities and challenges toward responsible AI, Inf. Fusion source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The model opaqueness makes it hard even for the model developers to understand how the model is functioning [14].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 1. Introduction
What the sources say
- Not checked yetcites [14] Tripicchio, 2020, Is deep learning ready to satisfy industry needs?, Procedia Manuf. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The Watson system could not provide justifications for its recommendations in clinical practice [38].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.1. DNN in medical applications
What the sources say
- Not checked yetcites [38] Durán, 2021
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“AI can aid biomedical imaging as it can “see” additional and superior detailed information regarding tumor size, texture, shape etc. compared to, for example, human radiologists [7].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 1. Introduction
What the sources say
- Not checked yetcites [7] Sorantin, 2021, The augmented radiologist: artificial intelligence in the practice of radiology, Pediatr. Radiol.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
None of the sources cited for this claim have been checked yet.
In the paper
“It was demonstrated that deep learning models on chest X-ray image datasets were not relying on medical pathology but instead on confounding factors for COVID-19 detection [41,43].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.1. DNN in medical applications
What the sources say
- Not checked yetcites [41] Ahmed, 2021
- Not checked yetcites [43] DeGrave, 2021, AI for radiographic COVID-19 detection selects shortcuts over signal, Nat. Mach. Intell. source
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The advances in post-hoc interpretability is helpful by approximating the black box models [58].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.2.5. Accuracy and interpretability trade-off
What the sources say
- Not checked yetcites [58] Rai, 2019, Explainable AI: from black box to glass box, J. Acad. Market. Sci. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This trade-off can limit the usability of the more accurate models.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.2.5. Accuracy and interpretability trade-off
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“An explainable capsule network, X-Caps was proposed to predict based on human interpretable features by encoding high-level image attributes within the capsule vectors [15].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.2. Visualization by perturbation
What the sources say
- Not checked yetcites [15] R. LaLonde, D. Torigian and U. Bagci, “Encoding visual attributes in capsules for explainable medical diagnoses,” in Medical Image Computing and Computer Assisted Intervention – MICCAI 2020. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Owing to a large variety of AI models and their intended purposes, a single strategy for explainability will probably not work well for the different stakeholders in the system [60].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.3. XAI stakeholders
What the sources say
- Not checked yetcites [60] Páez, 2019, The pragmatic turn in explainable artificial intelligence (XAI), Minds Mach. source
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The model was evaluated using the proposed automatic causal metrics, termed as deletion and insertion, showing better performance compared to Grad-CAM, Sliding Window, and LIME [72].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.2. Visualization by perturbation
What the sources say
- Not checked yetcites [72] V. Petsiuk, A. Das and K. Saenko, “RISE: Randomized Input Sampling for Explanation of Black-box Models,” [Online]. Available: https://doi.org/10.48550/arXiv.1806.07421. [Accessed 15 September 2022].
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Thus, there is an accuracy cost for intrinsic interpretable models [21].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.2.5. Accuracy and interpretability trade-off
What the sources say
- Not checked yetcites [21] Adadi, 2018, Peeking inside the black-box: a survey on explainable artificial intelligence (XAI), IEEE Access source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The method was shown to provide better results in highlighting the regions of interest in the saliency maps comparison with RISE and LIME [73].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.2. Visualization by perturbation
What the sources say
- Not checked yetcites [73] Vasu, 2020, Iterative and adaptive sampling with spatial attention for black-box model explanations
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“It remains a challenge to build models that are both accurate and interpretable [52].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.2.5. Accuracy and interpretability trade-off
What the sources say
- Not checked yetcites [52] Abeyrathna, 2021, Extending the tsetlin machine with integer-weighted clauses for increased interpretability, IEEE Access source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“During the model development, it is important to have data from unseen sources in the test set to ensure model generalizability [41].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.1. DNN in medical applications
What the sources say
- Not checked yetcites [41] Ahmed, 2021
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The evaluation of deep learning techniques was performed across different hospitals, and for pneumonia screening it was found that the models had better internal (of hospitals) results compared to external [42].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.1. DNN in medical applications
What the sources say
- Not checked yetcites [42] Zech, 2018, Variable generalization performance of a deep learning model to detect pneumonia in chest radiographs: a cross-sectional study, PLoS Med. source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The method was demonstrated to be better compared to Grad-CAM in both the objective evaluations and through human studies [80].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.3. Visualization by gradient/backpropagation
What the sources say
- Not checked yetcites [80] Chattopadhay, 2018, Grad-CAM++: generalized gradient-based visual explanations for deep convolutional networks
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“For the localization and severity grading of COVID-19, activation maps could be unsuitable for clinical use as these could provide inconsistent localization [159].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [159] Degerli, 2021, COVID-19 infection map generation and detection from chest X-ray images, Health Inf. Sci. Syst. source
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Using differences lets the information through even in case of zero gradient [86].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.3. Visualization by gradient/backpropagation
What the sources say
- Not checked yetcites [86] Shrikumar, 2017, Learning important features through propagating activation differences
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The model was used for COVID and pneumonia classification and achieved accuracies of 95% and 97% respectively [163].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [163] Ravi, 2020
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The method Eigen-CAM was proposed to use principal components in order to provide visual explanation in the presence of an adversarial noise [82].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.3. Visualization by gradient/backpropagation
What the sources say
- Not checked yetcites [82] Muhammad, 2020, Eigen-CAM: class activation map using principal components
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“MSFI metric had better results compared to the other chosen post-hoc heatmap algorithms for multi-modal imaging tasks [133].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.6. Evaluation of explainability
What the sources say
- Not checked yetcites [133] Jin, 2022, Evaluating explainable AI on a multi-modal medical imaging task: can existing algorithms fulfill clinical requirements?
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“It was formally proved by the authors that Grad-CAM generalized CAM for many CNN architectures [79].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.3. Visualization by gradient/backpropagation
What the sources say
- Not checked yetcites [79] Selvaraju, 2017, Visual explanations from deep networks via gradient-based localization
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The nature of explanations being concept-based and counterfactual, were found to be relevant for both experts and non-experts [103].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > Model Agnostic Globally Interpretable Explanations (MAGIX)
What the sources say
- Not checked yetcites [103] Akula, 2020, CoCoX: generating conceptual and counterfactual explanations via fault-lines
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“It was shown that the attribution maps of the counterfactual images help in interpreting the misclassified images [122].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > Model Agnostic Globally Interpretable Explanations (MAGIX)
What the sources say
- Not checked yetcites [122] Suzuki, 2021, Understanding the reason for misclassification by generating counterfactual images
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The model performance was more accurate on CT compared to CXR images, and had an accuracy of 95.61% in COVID-19 detection.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [158] Panwar, 2020
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The explanation space visualization with low-dimensional features was shown to improve the human machine communication [95].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.4. Explanation by example
What the sources say
- Not checked yetcites [95] Qi, 2020
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The sparse masks align well with human intuitions as compared to backpropagation-based techniques [75].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.2. Visualization by perturbation
What the sources say
- Not checked yetcites [75] Li, 2020, Neural image compression and explanation, IEEE Access source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The proposed method was compared with Grad-CAM, Grad-CAM++, and CNN Fixations and the results showed an improvement for the weakly supervised object localization [82].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.3. Visualization by gradient/backpropagation
What the sources say
- Not checked yetcites [82] Muhammad, 2020, Eigen-CAM: class activation map using principal components
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The results showed that a visual inspection of the explanations in itself may lack sensitivity to the model and data [107].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.6. Evaluation of explainability
What the sources say
- Not checked yetcites [107] Adebayo, 2018, Sanity checks for saliency maps, Adv. Neural Inf. Process. Syst.
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“It was shown that PIECE generated the best semi-factuals with a higher L 1 distance (test to explanation image) compared to other similar methods, such as Min-Edit [106].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > Model Agnostic Globally Interpretable Explanations (MAGIX)
What the sources say
- Not checked yetcites [106] Kenny, 2021, On generating plausible counterfactual and semi-factual explanations for deep learning
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The mask is used to subsample less important background pixels to a low resolution whereas the important pixels retain their resolution [75].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.2. Visualization by perturbation
What the sources say
- Not checked yetcites [75] Li, 2020, Neural image compression and explanation, IEEE Access source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The proposed method was evaluated and compared against LIME, Interpretable Decision Sets (IDS) and Bayesian Decision Lists (BDL) and the accuracy was found to be higher than IDS and BDL [97].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > Model Agnostic Globally Interpretable Explanations (MAGIX)
What the sources say
- Not checked yetcites [97] Lakkaraju, 2017
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“An image classification task was chosen for the human evaluation of the proposed model for 'justified trust' and 'explanation satisfaction', and demonstrated better performance compared to other techniques [103].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > Model Agnostic Globally Interpretable Explanations (MAGIX)
What the sources say
- Not checked yetcites [103] Akula, 2020, CoCoX: generating conceptual and counterfactual explanations via fault-lines
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The clusters of similar filters from one layer to another can highlight the filters responding to similar semantic concepts [91].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.3. Visualization by gradient/backpropagation
What the sources say
- Not checked yetcites [91] Tan, 2020, Locality guided neural networks for explainable artificial intelligence
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The approach was validated using VGG16 model and could derive counterfactual explanations for 44 out of 50 images [121].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > Model Agnostic Globally Interpretable Explanations (MAGIX)
What the sources say
- Not checked yetcites [121] J. Chandrasekaran, Y. Lei, R. Kacker and D. R. Kuhn, “A combinatorial approach to explaining image classifiers,” in IEEE International Conference on Software Testing, Verification and Validation Workshops (ICSTW), 20021.
hedgeddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The proposed system was used for deep medical image classification using CheXpert dataset and the results showed that the finer details of the input images were manifested in the generated symbols [252].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [252] Chowdhury, 2021, Emergent Symbolic language based deep medical image classification
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Although there are efforts to incorporate explainability while analyzing the CXRs and X-rays, the algorithms are often not able to explain or mimic the exact experience of a human grader or radiologists which they utilize during X-ray examinations.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The explanations using Bayesian Teaching on CXR were obtained for pneumothorax diagnosis and showed that it can help medical experts to predict AI decisions [8].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [8] Folke, 2021, Explainable AI for medical imaging: explaining pneumothorax diagnoses with Bayesian teaching, Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications III
plaindirectn = 8- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The study showed that the ensemble provided better results compared to transfer learning and achieved an AUC of 0.92 [167].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [167] Liz, 2021, Ensembles of Convolutional Neural Network models for pediatric pneumonia diagnosis, Future Generat. Comput. Syst. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The model activations were visualized using LIME and Grad-CAM and found to highlight the discriminative regions in the CXR [171].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [171] Rajaraman, 2019, Visualizing and explaining deep learning predictions for pneumonia detection in pediatric chest radiographs, SPIE Medical Imaging
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The multiclass classification (COVID, normal, viral or bacterial pneumonia) achieved an accuracy of 90.2%.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [160] Mahmud, 2020, CovXNet: a multi-dilation convolutional neural network for automatic COVID-19 and other pneumonia detection from chest X-ray images with transferable multi-receptive feature optimization, Comput. Biol. Med. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The method provided an average AUC score of 0.824 across all 14 pathologies.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [168] Chen, 2019, Lesion location attention guided network for multi-label thoracic disease classification in chest X-rays, IEEE journal of biomedical and health informatics source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The lesion detection was assessed using CAM and was found to be reliable [169].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [169] Liu, 2019, SDFN: segmentation-based deep fusion network for thoracic disease classification in chest X-ray images, Comput. Med. Imag. Graph. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The model achieved agreement with visual emphysema scores in the test cohort [178].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [178] Humphries, 2020, Deep learning enables automatic classification of emphysema pattern at CT, Radiology source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“LIME was used to help identify the specific features that helped to differentiate CT images for COVID-19 and normal images using 8 CNN models [157].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [157] Ahsan, 2020, COVID-19 symptoms detection based on NasNetMobile with explainable AI using various imaging modalities, Machine Learning and Knowledge Extraction source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The framework was tested on low-quality CT images, and ResNet-50 performed best amongst the selected DNNs with an accuracy of 99.87% and a sensitivity of 99.58% [183].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [183] Ko, 2020, COVID-19 pneumonia diagnosis using a simple 2D deep learning framework with a single chest CT image: model development and validation, J. Med. Internet Res. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“CNN model for COVID-19 classification of CT images used t-SNE (dimensionality reduction) and Grad-CAM showing well separated clusters for COVID-19 and non-COVID-19 cases [173].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [173] Alshazly, 2021, Explainable COVID-19 detection using chest CT scans and deep learning, Sensors source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Binary classification for renal cell carcinomas used large feature set with Relational Functional Gradient Boosting (RFGB) and provided better model performance in accuracy, F1-Score and AUC, compared to other techniques and a shallow neural network [219].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [219] Kunapuli, 2018, A decision-support tool for renal mass classification,” journal of digital imaging, J. Digit. Imag. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Grad-CAM was used for explanations of the classification results and the proposed system was found to identify the small size lesions [225].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [225] Chen, 2019, A dual-attention dilated residual network for liver lesion classification and localization on CT images
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“It was shown that the lung and lobe segmentation improved the classification results by over 6% [174].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [174] Pennisi, 2021, An explainable AI system for automated COVID-19 assessment and lesion categorization from CT-scans, Artif. Intell. Med. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The ensembles were useful for improving the classification results for both bacterial and viral pneumonia [167].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [167] Liz, 2021, Ensembles of Convolutional Neural Network models for pediatric pneumonia diagnosis, Future Generat. Comput. Syst. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Heatmaps were used for visualisations and performed well for smaller lesion detection [226].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [226] Shapira, 2020, Liver lesion localisation and classification with convolutional neural networks: a comparison between conventional and spectral computed tomography, Biomed. Phys. Eng. Express source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The study showed that the ensemble provided better results compared to transfer learning and achieved an AUC of 0.92 [167].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [167] Liz, 2021, Ensembles of Convolutional Neural Network models for pediatric pneumonia diagnosis, Future Generat. Comput. Syst. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“CAM was used to generate the heatmaps and the system provided good accuracy results despite a small dataset [235].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [235] Yi, 2019, Automated semantic labeling of pediatric musculoskeletal radiographs using deep learning, Pediatr. Radiol. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The results were incorporated to increase the radiologists' confidence level for clinical diagnosis [180].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [180] Lei, 2020, Shape and margin-aware lung nodule classification in low-dose CT images via soft activation mapping, Med. Image Anal. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Five deep CNN models were used for automatic classification of paediatric musculoskeletal radiographs of the five anatomical area (shoulder, elbow, pelvis, hand, and knee) and achieved a classification rate of 33 radiographs per second [235].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [235] Yi, 2019, Automated semantic labeling of pediatric musculoskeletal radiographs using deep learning, Pediatr. Radiol. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The system was evaluated and showed an AUC of 0.91 [239].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [239] Candemir, 2020, Automated coronary artery atherosclerosis detection and weakly supervised localization on coronary CT angiography with a deep 3-dimensional convolutional neural network, Comput. Med. Imag. Graph. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“A Dual-attention Dilated Residual Network (DADRN) was proposed for liver lesion classification on a CT image dataset and achieved an accuracy comparable to RoI based methods [225]”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [225] Chen, 2019, A dual-attention dilated residual network for liver lesion classification and localization on CT images
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The method achieved better classification performance compared to the other approaches based on ROC [169].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [169] Liu, 2019, SDFN: segmentation-based deep fusion network for thoracic disease classification in chest X-ray images, Comput. Med. Imag. Graph. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“CT images of COVID-19 were classified with U-Net achieving better performance (accuracy, precision, specificity, and AUC) than other state-of-the-art methods, such as COVID-Net and COVNet [141].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [141] Yang, 2022, Unbox the black-box for the medical explainable AI via multi-modal and multi-centre data fusion: a mini-review, two showcases and beyond, Inf. Fusion source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The malignancy prediction in a lung cancer dataset were derived from capsule vector weights indicating the contribution of a visual attribute to nodule's malignancy [15].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [15] R. LaLonde, D. Torigian and U. Bagci, “Encoding visual attributes in capsules for explainable medical diagnoses,” in Medical Image Computing and Computer Assisted Intervention – MICCAI 2020. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Grad-CAM++ was used for visualization and the model was found to be useful for predicting stroke outcome [136].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [136] Hilbert, 2019, Data-efficient deep learning of radiological image data for outcome prediction after endovascular treatment of patients with acute ischemic stroke, Comput. Biol. Med. source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The localization used Grad-CAM for deriving heatmap visualisations helpful to the physician in locating atherosclerosis location and likelihood [239].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [239] Candemir, 2020, Automated coronary artery atherosclerosis detection and weakly supervised localization on coronary CT angiography with a deep 3-dimensional convolutional neural network, Comput. Med. Imag. Graph. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The method performed better compared to DeepTaylor, LRP, and DeepSHAP in terms of image quality and semantic meaning [176].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.4. Ultrasound imaging
What the sources say
- Not checked yetcites [176] Katzmann, 2021, Explaining clinical decision support systems in medical imaging using cycle-consistent activation maximization, Neurocomputing source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“CNN feature maps activations were used to understand the CNN model activations in dermoscopic lesion images to increase the dermatologists' trust in the system [215].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.5. Optical imaging 4.5.1. Dermatology
What the sources say
- Not checked yetcites [215] Molle, 2018, Visualizing convolutional neural networks to improve decision support for skin lesion classification
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Fetal brain planes were classified as normal or abnormal using a CNN model on ultrasound images, with Grad-CAM for lesion localization and diagnosis [155].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.4. Ultrasound imaging
What the sources say
- Not checked yetcites [155] Xie, 2020, Computer-aided diagnosis for fetal brain ultrasound images using deep convolutional neural networks, Int. J. Comput. Assist. Radiol. Surg. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The method was validated on a weighted MRI dataset and provided better performance over Regional Mean Volume (RMV), and Hierarchical Feature Fusion (HFF) methods using the four metrics (accuracy, sensitivity, specificity, and AUC) for the binary and multiclass classification [145].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [145] Lee, 2019, Toward an interpretable Alzheimer's disease diagnostic model with regional abnormality representation via deep learning, Neuroimage source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“A method aimed at increasing the clinical accuracy for the early stages of Alzheimer's disease was proposed and showed that the voxel relationships in different regions can provide discriminative information for clinical diagnosis [145].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [145] Lee, 2019, Toward an interpretable Alzheimer's disease diagnostic model with regional abnormality representation via deep learning, Neuroimage source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“MRI and CT images for hydrocephalus segmentation achieved significant improvement in thin-slice MRI and CT images [141].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [141] Yang, 2022, Unbox the black-box for the medical explainable AI via multi-modal and multi-centre data fusion: a mini-review, two showcases and beyond, Inf. Fusion source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The 3D, T1-weighted multi-echo, gradient-echo images were obtained from an MRI scanner and it was concluded using relevance analysis that relevant voxels were around venous vessels [153].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [153] Lopatina, 2020, Investigation of deep-learning-driven identification of multiple sclerosis patients based on susceptibility-weighted images using relevance analysis, Front. Neurosci. source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The results were compared with DenseASPP, RefineNet and DeepLabv3+, and showed comparable results [211].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.5. Optical imaging 4.5.1. Dermatology
What the sources say
- Not checked yetcites [211] Gu, 2021, CA-net: comprehensive attention convolutional neural networks for explainable medical image segmentation, IEEE Trans. Med. Imag. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“It was found that the age and location metadata improved the classification performance.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.5. Optical imaging 4.5.1. Dermatology
What the sources say
- Not checked yetcites [213] Li, 2020, Fusing metadata and dermoscopy images for skin disease diagnosis
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The method also provided representations of regional abnormalities for prediction interpretability [145].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [145] Lee, 2019, Toward an interpretable Alzheimer's disease diagnostic model with regional abnormality representation via deep learning, Neuroimage source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The proposed approach was shown to be effective for generating counterfactual explanations, which are an important tool to explain the model predictions [216].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.5. Optical imaging 4.5.1. Dermatology
What the sources say
- Not checked yetcites [216] Narayanaswamy, 2021, Using deep image priors to generate counterfactual explanations
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“CNN was used to create the biomarkers for Parkinson's disease from Neuromelanin Sensitive Magnetic Resonance Imaging (NMS-MRI) and achieved better test accuracy compared to contrast ratio-based classification [147].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [147] Shinde, 2019, Predictive markers for Parkinson's disease using deep neural nets on neuromelanin sensitive MRI, Neuroimage: Clinical
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“3D CNN was used for tumor grading on BraTS dataset for expediting tumor assessment for the whole brain and a defined tumor region, and showed that the results with defined RoI achieved better accuracy [150].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [150] Pereira, 2018, Automatic brain tumor grading from MRI data using convolutional neural networks and quality assessment
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Automated skin lesion segmentation can help with an early diagnosis of skin disease [211].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.5. Optical imaging 4.5.1. Dermatology
What the sources say
- Not checked yetcites [211] Gu, 2021, CA-net: comprehensive attention convolutional neural networks for explainable medical image segmentation, IEEE Trans. Med. Imag. source
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Grad-CAM was used for visual explanations of the wrap-around model in low-field brain MRI images on two datasets showing an agreement between the radiologists and the models using Cohen's kappa values [154].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [154] Jimeno, 2022, ArtifactID: identifying artifacts in low-field MRI of the brain using deep learning, Magn. Reson. Imag. source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“A framework used volumetric neonatal MRI for classification of brain dysmaturation with 3D CNN and CAM reporting a classification accuracy of 98.5% [242].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [242] Ceschin, 2018, A computational framework for the detection of subcortical brain dysmaturation in neonatal MRI using 3D Convolutional Neural Networks, Neuroimage source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The model used CAM for finding the relevant features and the final visualizations demonstrated how the system reached its decision on classifying tumors [149].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [149] Kumar, 2021, Doctor's dilemma: evaluating an explainable subtractive spatial Lightweight convolutional neural network for brain tumor diagnosis, ACM Trans. Multimed Comput. Commun. Appl source
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“MSFI was shown to perform better in comparison to the other chosen post-hoc heatmap algorithms for multi-modal imaging tasks [133].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [133] Jin, 2022, Evaluating explainable AI on a multi-modal medical imaging task: can existing algorithms fulfill clinical requirements?
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The generalization and representation of the model were enhanced by reducing the model uncertainty [141].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [141] Yang, 2022, Unbox the black-box for the medical explainable AI via multi-modal and multi-centre data fusion: a mini-review, two showcases and beyond, Inf. Fusion source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“A Deep CNN based model was proposed for simultaneous skin lesion segmentation and classification and found to perform better compared to the other models on Jaccard index and AUC on ISIC 2017 and PH2 datasets [214].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.5. Optical imaging 4.5.1. Dermatology
What the sources say
- Not checked yetcites [214] Xie, 2020, A mutual bootstrapping model for automated skin lesion segmentation and classification, IEEE Trans. Med. Imag. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Ultrasound images have not been used as extensively in clinical AI analysis as the other imaging modalities [253].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.4. Ultrasound imaging
What the sources say
- Not checked yetcites [253] Komatsu, 2021, Towards clinical application of artificial intelligence in ultrasound imaging, Biomedicines source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The evaluation of the proposed scheme showed significant performance improvement over U-Net, through the use of comprehensive attention guidance with spatial, channel, and scale attention modules [211].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [211] Gu, 2021, CA-net: comprehensive attention convolutional neural networks for explainable medical image segmentation, IEEE Trans. Med. Imag. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The study concluded that the framework and LRP provided transparency in the model's decision [142].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [142] Eitel, 2019, Uncovering convolutional neural network decisions for diagnosing multiple sclerosis on conventional MRI using layer-wise relevance propagation, Neuroimage: Clinical
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Various methods were compared and DeepLIFT was determined to be preferable to LRP [153].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [153] Lopatina, 2020, Investigation of deep-learning-driven identification of multiple sclerosis patients based on susceptibility-weighted images using relevance analysis, Front. Neurosci. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The interpretability was evaluated with Grad-CAM and Guided Backpropagation for validating the trustworthiness [150].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [150] Pereira, 2018, Automatic brain tumor grading from MRI data using convolutional neural networks and quality assessment
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
None of the sources cited for this claim have been checked yet.
In the paper
“Some studies reporting DNN medical imaging performance exceeding the radiologists were often using small datasets and can be considered a significant limitation [6,254].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“White-box modelling methods like XNN, scalable Bayesian rule lists, and monotonic GBM, provide little accuracy penalty for interpretability [264].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.2. Understanding and improving explainability
What the sources say
- Not checked yetcites [264] Hall, 2018
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The lack of data quality can also hinder the results of explanation by examples [120].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [120] Kim, 2016, Examples are not enough, learn to criticize! criticism for interpretability
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The reduction of noise in the data can make it easier for the model to explain or interpret [268].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [268] Paka, 2021
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Deep learning models require sufficient training data to yield reproducible and useful results [37].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [37] Singh, 2021
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“This can be a challenge for imaging datasets where obtaining the diseased image samples is challenging due to limited number of samples compared to normal cases [6].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [6] Chan, 2020, Deep learning in medical image analysis, Adv. Exp. Med. Biol. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Denoising methods can improve the robustness of the DNN against small perturbations and quality degradations by using denoising methods to enhance the edges and outlines [266].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [266] Yim, 2017, Enhancing the performance of convolutional neural networks on quality degraded datasets
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
None of the sources cited for this claim have been checked yet.
In the paper
“The data quality can be affected by noise and incorrect labelling [10,254].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [10] G. Ras, M. van Gerven and P. Haselager, “Explanation Methods in Deep Learning: Users, Values, Concerns and Challenges,” Explainable And Interpretable Models in Computer Vision and Machine Learning, pp. 19-36. source
- Not checked yetcites [254] Le, 2019, Artificial intelligence in breast imaging, Clin. Radiol. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The model's highlighted features might also differ in relevance to human identified features and decisions [19].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.2. Understanding and improving explainability
What the sources say
- Not checked yetcites [19] Ghassemi, 2021, The false hope of current approaches to explainable artificial intelligence in health care, The Lancet Digital Health source
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
None of the sources cited for this claim have been checked yet.
In the paper
“Using the XAI techniques to assist the clinician can help to produce better diagnosis than the model or clinician alone [204,212,233,234].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.1. XAI applications in medical imaging
What the sources say
- Not checked yetcites [204] Qian, 2021, Prospective assessment of breast cancer risk from multimodal multiview ultrasound images via clinically applicable deep learning, Nat. Biomed. Eng. source
- Not checked yetcites [212] Stieler, 2021, Towards domain-specific explainable AI:model interpretation of a skin image classifier using a human approach
- Not checked yetcites [233] Kiani, 2020, Impact of a deep learning assistant on the histopathologic classification of liver cancer, NPJ Digital Med. source
- Not checked yetcites [234] Heinemann, 2019, Deep learning enables pathologist-like scoring of NASH models, Sci. Rep. source
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The quality and quantity of the training data plays a major role in the trained model's performance and resultantly to the explainability that can be obtained from such a model.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Data collection for breast cancer covering all image variations is difficult considering other factors such as age, race, ethnicity etc. [6].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [6] Chan, 2020, Deep learning in medical image analysis, Adv. Exp. Med. Biol. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
None of the sources cited for this claim have been checked yet.
In the paper
“The use of multiple scales can increase the localization and explainability [167,196,200,221].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.1. XAI applications in medical imaging
What the sources say
- Not checked yetcites [167] Liz, 2021, Ensembles of Convolutional Neural Network models for pediatric pneumonia diagnosis, Future Generat. Comput. Syst. source
- Not checked yetcites [196] Liao, 2019, Clinical interpretable deep learning model for glaucoma diagnosis, IEEE journal of biomedical and health informatics source
- Not checked yetcites [200] El Adoui, 2020, Multi-input deep learning architecture for predicting breast tumor response to chemotherapy using quantitative MR images, Int. J. Comput. Assist. Radiol. Surg. source
- Not checked yetcites [221] Uehara, 2020, Multi-scale exaplainable feature learning for pathological image analysis using convolutional neural networks
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
None of the sources cited for this claim have been checked yet.
In the paper
“The incorporation of the supplementary information in the model training can help to improve the model's prediction and explainability [156,217,226].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.1. XAI applications in medical imaging
What the sources say
- Not checked yetcites [156] Tang, 2019, Interpretable classification of Alzheimer's disease pathologies with a convolutional neural network pipeline, Nat. Commun.
- Not checked yetcites [217] Barata, 2021, Explainable skin lesion diagnosis using taxonomies, Pattern Recogn. source
- Not checked yetcites [226] Shapira, 2020, Liver lesion localisation and classification with convolutional neural networks: a comparison between conventional and spectral computed tomography, Biomed. Phys. Eng. Express source
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
None of the sources cited for this claim have been checked yet.
In the paper
“DNNs enable automated RoI selection without the time-consuming manual annotations [225,226].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.1. XAI applications in medical imaging
What the sources say
- Not checked yetcites [225] Chen, 2019, A dual-attention dilated residual network for liver lesion classification and localization on CT images
- Not checked yetcites [226] Shapira, 2020, Liver lesion localisation and classification with convolutional neural networks: a comparison between conventional and spectral computed tomography, Biomed. Phys. Eng. Express source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Developing an interpretable model from the very beginning will make explanation and fairness evaluation easier [264].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.2. Understanding and improving explainability
What the sources say
- Not checked yetcites [264] Hall, 2018
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: unclear — It is unclear whether it could be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“A systematic understanding of XAI and humans as a system is lacking and remains a challenge [8].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.5. Human-in-the-loop
What the sources say
- Not checked yetcites [8] Folke, 2021, Explainable AI for medical imaging: explaining pneumothorax diagnoses with Bayesian teaching, Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications III
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“New types of information and novel insights into the data are possible by combining the data from diverse sources [272].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [272] Holzinger, 2018, From machine learning to explainable AI
hedgeddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The radiologist's accuracy in characterising lesions can be much improved using a DNN [6].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.5. Human-in-the-loop
What the sources say
- Not checked yetcites [6] Chan, 2020, Deep learning in medical image analysis, Adv. Exp. Med. Biol. source
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“LRP pruning was shown to provide good results for transfer learning with small datasets [56].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [56] Yeom, 2021, Pruning by explaining: a novel criterion for deep neural network pruning, Pattern Recogn. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This can facilitate a global repository of benchmark datasets for biomedical imaging.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Many DNN models meet the explainability requirements of the system developers and specialists but not of the end users of those systems [59].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.6. Incorporating stakeholders in XAI
What the sources say
- Not checked yetcites [59] Bauer, 2021, Expl(AI)n it to me – explainable AI and information systems research, Business & Information Systems Engineering source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The imaging data captured in clinical settings with the exception of radiology may only be housed in the local department's systems [13].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [13] McCarthy, 2021, Enterprise imaging and big data: a review from a medical physics perspective, Phys. Med. source
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Counterfactual algorithms have the potential to uncover biases in both the models and data [106].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
What the sources say
- Not checked yetcites [106] Kenny, 2021, On generating plausible counterfactual and semi-factual explanations for deep learning
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The predictions of DNN with the same architecture may be different due to data quality and the model initialization (random) for MRI images [269].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [269] Thibeau-Sutre, 2020, Visualization approach to assess the robustness of neural networks for medical image classification
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“The DNN model can also be susceptible to an adversarial attack.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Explainable systems that involve the human user in the decision process of the DNN are useful to determine answers to the 'why' and 'how' questions regarding the model's predictions [93].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.5. Human-in-the-loop
What the sources say
- Not checked yetcites [93] Dong, 2017, Improving interpretability of deep neural networks with semantic information
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption: The ultimate beneficiary of model explainability is a human user.
None of the sources cited for this claim have been checked yet.
In the paper
“The information fusion and multimodal system can have a significant performance over the systems working with a single type of data [36,100].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.7. Hybrid/fusion imaging for XAI
What the sources say
- Not checked yetcites [36] Panagiotis, 2021, Artificial intelligence: deep learning in oncological radiomics and challenges of interpretability and data harmonization, Phys. Med. source
- Not checked yetcites [100] Holzinger, 2021, Towards multi-modal causability with Graph Neural Networks enabling information fusion for explainable AI, Inf. Fusion source
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Current AI success is attributable to models that provide better performance but are opaque or black box [45].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
What the sources say
- Not checked yetcites [45] Gunning, 2019
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The clinicians are more likely to use a DNN for diagnosis if the model provides insights into its decision and can be understood to be not being driven by any artifacts [283].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.6. Incorporating stakeholders in XAI
What the sources say
- Not checked yetcites [283] Gilvary, 2019, The missing pieces of artificial intelligence in medicine, Trends Pharmacol. Sci. source
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“In medical imaging context, it could not only include the different image modalities for diagnosis but increasingly includes patient attributes or time-series data [13].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.7. Hybrid/fusion imaging for XAI
What the sources say
- Not checked yetcites [13] McCarthy, 2021, Enterprise imaging and big data: a review from a medical physics perspective, Phys. Med. source
hedgedcorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The use of XAI for a binary decision regarding presence or absence of a disease is of great help, allowing more complex cases which are difficult to classify to be determined in a clinical setting by a human expert [281].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.5. Human-in-the-loop
What the sources say
- Not checked yetcites [281] Mursch-Edlmayr, 2020, Artificial intelligence algorithms to diagnose glaucoma and detect glaucoma progression: translation to clinical practice, Translat. Vision Sci. Technol. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption: Trained DNN models can provide predictions comparable to human experts in a fraction of the time.
In the paper
“The study of adversarial attacks can also help to develop better interpretation models that are more robust against such attacks [278].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
What the sources say
- Not checked yetcites [278] Le, 2021, Robust adversarial attack against explainable deep classification models based on adversarial images with different Patch sizes and perturbation ratios, IEEE Access source
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Explainability can also be used for model improvement.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.9. Explainability to improve models
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“The results on four datasets demonstrated that the Guided Zoom improved the model's classification accuracy [290].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.9. Explainability to improve models
What the sources say
- Not checked yetcites [290] Bargal, 2021, Guided Zoom: zooming into network evidence to refine fine-grained model decisions, IEEE Trans. Pattern Anal. Mach. Intell. source
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“It was shown that interpretability of the representation learnt by the hidden units can be significantly affected by the different training techniques, thereby helping to understand the CNN model characteristics [291].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.9. Explainability to improve models
What the sources say
- Not checked yetcites [291] Bau, 2017, Network dissection: quantifying interpretability of deep visual representations
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Motivated by the need to develop hybrid intelligent systems, knowledge graphs can represent domain knowledge with XAI to develop trustworthy explanations [117].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.8. Mathematical and graph based XAI approaches
What the sources say
- Not checked yetcites [117] Tiddi, 2021
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Hybrid imaging provides additional information from the different image modalities that can significantly increase the investigation and identification of a disease [35].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.7. Hybrid/fusion imaging for XAI
What the sources say
- Not checked yetcites [35] O'Sullivan, 2020, Developments in AI and machine learning for neuroimaging
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Multimodality technique was observed to provide better specificity compared to each of the modalities alone [143].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.7. Hybrid/fusion imaging for XAI
What the sources say
- Not checked yetcites [143] Kamal, 2021, Alzheimer's patient analysis using image and gene expression data and explainable-AI to present associated genes, IEEE Trans. Instrum. Meas. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Model pruning using LRP was shown to be helpful to determine which filters and weights in a model were important [56].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.9. Explainability to improve models
What the sources say
- Not checked yetcites [56] Yeom, 2021, Pruning by explaining: a novel criterion for deep neural network pruning, Pattern Recogn. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- It may state two separate claims as one. 81
- Probably not a substantial claim. 102
- The quote only partly states this claim. 4
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“It was shown that LRP was useful for explainability of Alzheimer's classification predictions and applicable to similar disease diagnosis using MRI data [144].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Partly supportedcites [144] Böhle, 2019, Layer-wise relevance propagation for explaining deep neural network decisions in MRI-based Alzheimer's disease classification, Front. Aging Neurosci. sourceExperimental — accuracy not yet measured
The Discussion and Conclusion show that LRP helped explain individual CNN decisions in MRI-based Alzheimer’s classification. The Conclusion says the framework *might* be useful for other disease-classification studies, but the paper does not demonstrate its applicability to diagnosing other diseases. (LLM, conf=0.85)
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“The clinical acceptance of AI techniques would be aided by multi-modality and information fusion of both imaging and non-imaging data [34,170].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.7. Hybrid/fusion imaging for XAI
What the sources say
- Not checked yetcites [34] Karthik, 2021, AI for COVID-19 detection from radiographs: incisive analysis of state of the art techniques, IRBM source
- Partly supportedcites [170] Dunnmon, 2019, Assessment of convolutional neural networks for automated classification of chest radiographs, Radiology sourceExperimental — accuracy not yet measured
The Introduction and Discussion report that combining a chest-radiograph CNN score with a clinician’s assessment improved triage performance and could be useful in clinical practice. They do not assess multimodal imaging, fusion with non-imaging patient data, or whether either would increase clinical acceptance of AI. (LLM, conf=0.82)
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“The increased use of XAI techniques is seen in neurology, respiratory and ophthalmic studies.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.1. XAI applications in medical imaging
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“CXR images are most favored for respiratory studies due to the ease with which these can be obtained, with Grad-CAM as the most prevalent XAI methods.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.1. XAI applications in medical imaging
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“In ophthalmic studies, fundus images seem to be the dominant modality.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.1. XAI applications in medical imaging
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“In neurology, the most prevalent image modality is MRI.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.1. XAI applications in medical imaging
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
None of the sources cited for this claim have been checked yet.
In the paper
“Due to the paucity of the medical data, some research is focused on addressing the associated challenges [140,227].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.1. XAI applications in medical imaging
What the sources say
- Not checked yetcites [140] Choi, 2020, Cognitive signature of brain FDG PET based on deep learning: domain transfer from Alzheimer's disease to Parkinson's disease, Eur. J. Nucl. Med. Mol. Imag. source
- Not checked yetcites [227] Rajpurkar, 2020, AppendiXNet: deep learning for diagnosis of appendicitis from a small dataset of CT exams using video pretraining, Sci. Rep. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“The problem of bias can be addressed by an open sharing of the research process, methods, data and results, with such open sharing across institutions and funding agencies enabling open science [39].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.1. DNN in medical applications
What the sources say
- Not checked yetcites [39] Norori, 2021, Addressing bias in big data and AI for health care: a call for open science, Patterns source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“AI can potentially transform the image-based diagnosis in radiology [8].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 1. Introduction
What the sources say
- Not checked yetcites [8] Folke, 2021, Explainable AI for medical imaging: explaining pneumothorax diagnoses with Bayesian teaching, Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications III
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.
In the paper
“Neural networks require a lot of training data and time before they can produce acceptable results but are like a 'black-box' providing little or no insight into the decision-making process [13].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 1. Introduction
What the sources say
- Not checked yetcites [13] McCarthy, 2021, Enterprise imaging and big data: a review from a medical physics perspective, Phys. Med. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
None of the sources cited for this claim have been checked yet.
In the paper
“Convolutional Neural Networks (CNN) are state-of-the-art models for image segmentation and classification [2], although encoder-decoder transformer architectures have also been proposed [3][4][5].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 1. Introduction
What the sources say
- Not checked yetcites [3] Ahamed, 2022, Joint learning with local and global consistency for improved medical image segmentation
- Not checked yetcites [4] Dosovitskiy, 2010
- Not checked yetcites [5] Strudel, 2021, Segmenter: transformer for semantic segmentation
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“The open models can be trained with local data from, for example, hospitals across the globe to mitigate bias and boost sample size [39].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.1. DNN in medical applications
What the sources say
- Not checked yetcites [39] Norori, 2021, Addressing bias in big data and AI for health care: a call for open science, Patterns source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“Algorithm or model bias can be a risk in AI as the models can generate a wrongful output based on an intentional or unintentional bias [40].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.1. DNN in medical applications
What the sources say
- Not checked yetcites [40] Meske, 2020
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Disease diagnosis through these techniques is inevitable to help overcome the shortage of, for example, 2000 radiologists in UK [9].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 1. Introduction
What the sources say
- Not checked yetcites [9]
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.
In the paper
“The performance of IBM Watson system for oncology had very successful performance and agreements with oncologists in some areas but poor in other areas, indicating adaptations to local practices and not extensible at a global scale [38].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.1. DNN in medical applications
What the sources say
- Not checked yetcites [38] Durán, 2021
hedgeddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“Clinicians are the primary users of XAI systems and they may have differing views but all have a need for the explainability for the patient [27].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.3.3. Medical practitioners
What the sources say
- Not checked yetcites [27] Antoniadi, 2021, Current challenges and future opportunities for XAI in machine learning-based clinical decision support systems: a systematic review, Appl. Sci. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“Most XAI work is by the AI experts which might help engineers to build better AI systems but may be limited for other stakeholders especially the end-user [8].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.3.1. AI experts and application developers
What the sources say
- Not checked yetcites [8] Folke, 2021, Explainable AI for medical imaging: explaining pneumothorax diagnoses with Bayesian teaching, Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications III
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The regulators would most likely be interested in an overall performance and a combination of explanations of the model rather than specific cases of individual patients [25].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.3.2. Regulators
What the sources say
- Not checked yetcites [25] W. Samek and K. R. Müller, “Towards explainable artificial intelligence,” in Explainable AI: Interpreting, Explaining and Visualizing Deep Learning. Lecture Notes in Computer Science, vol. vol. 11700, Springer, Cham..
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“The models with an intrinsic interpretability provide better explainability at lower predictive performance compared to post-hoc methods that in contrast have less explainability but a higher predictive performance [59].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.2.5. Accuracy and interpretability trade-off
What the sources say
- Not checked yetcites [59] Bauer, 2021, Expl(AI)n it to me – explainable AI and information systems research, Business & Information Systems Engineering source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“It was demonstrated that the proposed 2D capsule network system provided better visualizations and malignancy predictions compared to 3D CNN and achieved better diagnostic accuracy [15].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.2. Visualization by perturbation
What the sources say
- Not checked yetcites [15] R. LaLonde, D. Torigian and U. Bagci, “Encoding visual attributes in capsules for explainable medical diagnoses,” in Medical Image Computing and Computer Assisted Intervention – MICCAI 2020. source
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“For example, VGG-16 model requires 15.5 billion floating-point operations for a single image classification with approximately 138 million parameters [56].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.2.5. Accuracy and interpretability trade-off
What the sources say
- Not checked yetcites [56] Yeom, 2021, Pruning by explaining: a novel criterion for deep neural network pruning, Pattern Recogn. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Failure to identify and address the concerns of the system users can lead to another “AI Winter”, where a small market segment could change the research and development focus from XAI [61].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.3.3. Medical practitioners
What the sources say
- Not checked yetcites [61] A. Preece, D. Harborne, D. Braines, R. Tomsett and S. Chakraborty, “Stakeholders in Explainable AI,” [Online]. Available: https://doi.org/10.48550/arXiv.1810.00184. [Accessed 15 September 2022].
hedgeddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Natural Image Compression and Explanation (NICE) uses a mask over the entire image for each pixel in the image that captures the saliency based on its contribution to the prediction.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.2. Visualization by perturbation
What the sources say
- Not checked yetcites [75] Li, 2020, Neural image compression and explanation, IEEE Access source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“While trustworthiness should be a property of every explainable system, this does not mean that every trustworthy system should also be explainable [2].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.2.4. Trustworthy AI
What the sources say
- Not checked yetcites [2] Barredo Arrieta, 2020, Explainable Artificial Intelligence (XAI): concepts, taxonomies, opportunities and challenges toward responsible AI, Inf. Fusion source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“Self-explaining AI is proposed that produces a decision and an explanation.It provides a human understandable explanation and confidence level for each decision and which was deemed to motivate and lead towards trustworthy AI [55].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.2.4. Trustworthy AI
What the sources say
- Not checked yetcites [55] Elton, 2020, Self-explaining AI as an alternative to interpretable AI
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“The heatmaps are commonly used to explain AI models and the quality of heatmaps can be appreciated using perturbation analysis [25].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3. Explainability techniques for DNN
What the sources say
- Not checked yetcites [25] W. Samek and K. R. Müller, “Towards explainable artificial intelligence,” in Explainable AI: Interpreting, Explaining and Visualizing Deep Learning. Lecture Notes in Computer Science, vol. vol. 11700, Springer, Cham..
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“The introduction of a hierarchy excluded the regions with less important features [74].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.2. Visualization by perturbation
What the sources say
- Not checked yetcites [74] Monroe, 2021, HihO: accelerating artificial intelligence interpretability for medical imaging in IoT applications using hierarchical occlusion, Neural Comput. Appl. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Similar to ISO 13485:2016 Medical devices standard for Quality management systems that prescribes the medical devices regulatory requirements, there is a need for certifications of DNN models for clinical adoption [64].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.3.2. Regulators
What the sources say
- Not checked yetcites [64] 2018
plain- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.
In the paper
“Local Interpretable Modelagnostic Explanations (LIME) was proposed to explain any classifier's predictions by approximating it locally around the prediction [70].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.2. Visualization by perturbation
What the sources say
- Not checked yetcites [70] Ribeiro, 2016, Why should i trust you?': explaining the predictions of any classifier
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“The patients require simplified explanations for the clinical decision as the DNN models are very complex and cannot be explained to a lay user [49].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.3.4. Patients
What the sources say
- Not checked yetcites [49] Gryz, 2021, Black box algorithms and the rights of individuals: no easy solution to the ‘explainability’ problem, Internet Policy Review source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“Similarly, formal mathematical notations and techniques can unambiguously provide explanations for domain experts but are not targeted to the end users [63].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.3.1. AI experts and application developers
What the sources say
- Not checked yetcites [63] Heuillet, 2021, Explainability in deep reinforcement learning, Knowl. Base Syst. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“The application developers would like to use the XAI techniques to understand the inner working and model improvement [62].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.3.1. AI experts and application developers
What the sources say
- Not checked yetcites [62] Wang, 2020, CNN hyperparameter optimization based on CNN visualization and perception hash algorithm
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Certification frameworks for a formal validation of models could testify the model operation statistically and ensure that the model predictions meet the required regulatory standards such as ISO/IEC CD 23053 (2020) for establishing a framework for use of AI systems [49].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.3.2. Regulators
What the sources say
- Not checked yetcites [49] Gryz, 2021, Black box algorithms and the rights of individuals: no easy solution to the ‘explainability’ problem, Internet Policy Review source
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“Randomized Input Sampling for Explanation of Black-box Models (RISE) generates an importance map based on the importance of each pixel to the prediction [72].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.2. Visualization by perturbation
What the sources say
- Not checked yetcites [72] V. Petsiuk, A. Das and K. Saenko, “RISE: Randomized Input Sampling for Explanation of Black-box Models,” [Online]. Available: https://doi.org/10.48550/arXiv.1806.07421. [Accessed 15 September 2022].
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“A novel hierarchical occlusion algorithm is proposed and is based on two observations; that the important features are localized in an image, and features collection in multiple scales can affect the prediction outcome [74].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.2. Visualization by perturbation
What the sources say
- Not checked yetcites [74] Monroe, 2021, HihO: accelerating artificial intelligence interpretability for medical imaging in IoT applications using hierarchical occlusion, Neural Comput. Appl. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim.
In the paper
“Iterative and Adaptive Sampling with Spatial Attention (IASSA) was proposed that used an iterative and adaptive sampling module to determine the saliency map of each pixel for the model prediction [73].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.2. Visualization by perturbation
What the sources say
- Not checked yetcites [73] Vasu, 2020, Iterative and adaptive sampling with spatial attention for black-box model explanations
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“An interpretable model can obtain transparency by providing the reasons behind their output [52].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.2.2. Interpretable AI
What the sources say
- Not checked yetcites [52] Abeyrathna, 2021, Extending the tsetlin machine with integer-weighted clauses for increased interpretability, IEEE Access source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“A novel hierarchical occlusion algorithm is proposed and is based on two observations; that the important features are localized in an image, and features collection in multiple scales can affect the prediction outcome [74].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.2. Visualization by perturbation
What the sources say
- Not checked yetcites [74] Monroe, 2021, HihO: accelerating artificial intelligence interpretability for medical imaging in IoT applications using hierarchical occlusion, Neural Comput. Appl. source
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim.
In the paper
“There is little agreement on what is an interpretable ML system and how it should be evaluated [51].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.2.2. Interpretable AI
What the sources say
- Not checked yetcites [51] Doshi-Velez, 2017
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“There is a need to have a more formalized definition of interpretability in the context of DNNs [51].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.2.2. Interpretable AI
What the sources say
- Not checked yetcites [51] Doshi-Velez, 2017
plain- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.
In the paper
“A lack of consensus in the meaning of explainability originates due to different stakeholders, each having their own understanding of the terms [61].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.3. XAI stakeholders
What the sources say
- Not checked yetcites [61] A. Preece, D. Harborne, D. Braines, R. Tomsett and S. Chakraborty, “Stakeholders in Explainable AI,” [Online]. Available: https://doi.org/10.48550/arXiv.1810.00184. [Accessed 15 September 2022].
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
None of the sources cited for this claim have been checked yet.
In the paper
“There is a trade-off between the accuracy and interpretability of the models, as shown in Fig. 2, where non-linear models with better performance, such as DNNs are less explainable, and linear models which are least accurate, such as decision trees, are more explainable [2,21,45].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.2.5. Accuracy and interpretability trade-off
What the sources say
- Not checked yetcites [2] Barredo Arrieta, 2020, Explainable Artificial Intelligence (XAI): concepts, taxonomies, opportunities and challenges toward responsible AI, Inf. Fusion source
- Not checked yetcites [21] Adadi, 2018, Peeking inside the black-box: a survey on explainable artificial intelligence (XAI), IEEE Access source
- Not checked yetcites [45] Gunning, 2019
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The model during training ideally should learn the discriminating image features of the various classes but instead can also learn extraneous features [42].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.1. DNN in medical applications
What the sources say
- Not checked yetcites [42] Zech, 2018, Variable generalization performance of a deep learning model to detect pneumonia in chest radiographs: a cross-sectional study, PLoS Med. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“For domains with a lack of knowledge about the underlying causal process, such as healthcare, it is prudent not to prefer explainability over diagnostic or predictive accuracy as this may affect the health and well-being of patients with some who get additional investigations and treatment and similarly those who are mjs-or undiagnosed [57].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 2.2.5. Accuracy and interpretability trade-off
What the sources say
- Not checked yetcites [57] London, 2019, Artificial intelligence and black-box medical decisions: accuracy versus explainability, Hastings Cent. Rep. source
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.
In the paper
“The proposed technique exploited learned attention for better interpretability, well-aligned to human intuition on the datasets with METEOR and BLEU metrics [92].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.4. Explanation by example
What the sources say
- Not checked yetcites [92] Xu, 2015
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“The LRP has some limitations such as the heatmaps were noisy and could be non-discriminative [84].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.3. Visualization by gradient/backpropagation
What the sources say
- Not checked yetcites [84] Jung, 2021, Explaining CNN and RNN using selective layer-wise relevance propagation, IEEE Access source
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“Among the evaluated XAI methods on the evaluation task, Deconvnet and Grad-CAM were found to be least accurate, whereas LRP, Integrated Gradients and Guided Backpropagation were found to be the most accurate methods [130].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.6. Evaluation of explainability
What the sources say
- Not checked yetcites [130] Arras, 2022, CLEVR-XAI: a benchmark dataset for the ground truth evaluation of neural network explanations, Inf. Fusion source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The evaluation of proposed technique was found to be more consistent with human understanding and intuition with an improved computational efficiency compared to the other selected approaches [90].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.3. Visualization by gradient/backpropagation
What the sources say
- Not checked yetcites [90] Lundberg, 2017, A unified approach to interpreting model predictions, Adv. Neural Inf. Process. Syst.
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“Gradient-weighted Class Activation Mapping (Grad-CAM) was proposed for producing visual explanations by using the gradients in the final model layer to identify and highlight the important image regions [79].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.3. Visualization by gradient/backpropagation
What the sources say
- Not checked yetcites [79] Selvaraju, 2017, Visual explanations from deep networks via gradient-based localization
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“It was demonstrated that the proposed method could work in the presence of adversarial noise and irrespective of the model accuracy [82].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.3. Visualization by gradient/backpropagation
What the sources say
- Not checked yetcites [82] Muhammad, 2020, Eigen-CAM: class activation map using principal components
boosted- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The method was motivated by Grad-CAM and RISE, however, unlike Grad-CAM it was gradient-free, and unlike RISE, it used a combined final mask instead of a random mask [76].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.2. Visualization by perturbation
What the sources say
- Not checked yetcites [76] S. M. Muddamsetty, N. S. J. Mohammad and T. B. Moeslund, “SIDU: similarity difference and uniqueness method for explainable AI,” in IEEE International Conference on Image Processing (ICIP).
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The model was evaluated through a comparison with LRP, Contrastive LRP (CLRP), and Softmax Gradient LRP (SGLRP), and found to provide less noise, better class discrimination and held entire target objects [84].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.3. Visualization by gradient/backpropagation
What the sources say
- Not checked yetcites [84] Jung, 2021, Explaining CNN and RNN using selective layer-wise relevance propagation, IEEE Access source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“It was shown that the combined approach achieved better attention maps and a higher classification accuracy [114].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.4. Explanation by example
What the sources say
- Not checked yetcites [114] Kashyap, 2020, Looking in the right place for anomalies: explainable ai through automatic location learning
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The interface was evaluated using surveys for both experts of DNN and non-experts and showed the approach to be effective [126].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.5. Libraries, tools, and frameworks
What the sources say
- Not checked yetcites [126] Vyas, 2020, An interactive graphical visualization approach to CNNs and RNNs
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“It was argued that the Gradient-Based methods are lacking as the neurons' gradients do not relate to the top-down saliency [85].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.3. Visualization by gradient/backpropagation
What the sources say
- Not checked yetcites [85] Zhang, 2018, Top-down neural attention by excitation Backprop, Int. J. Comput. Vis. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Grad-CAM++ was proposed as a method to build on Grad-CAM and provided better visual explanations compared to it and was better in explaining multiple occurrences of the same class objects in a single image [80].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.3. Visualization by gradient/backpropagation
What the sources say
- Not checked yetcites [80] Chattopadhay, 2018, Grad-CAM++: generalized gradient-based visual explanations for deep convolutional networks
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The best performing individual DNN architectures were VGG-19 and DenseNet-161 and the heatmaps generated by Grad-CAM++ were the most reliable in comparison with Grad-CAM and LRP [162].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [162] Karim, 2020, Deepcovidexplainer: explainable COVID-19 diagnosis from chest X-ray images
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“A Hierarchical Feature Network (HFnet) was proposed that showed that the low-level layer features are superior in comparison to low-level semantic questions [115].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.4. Explanation by example
What the sources say
- Not checked yetcites [115] Hong, 2019, Exploiting hierarchical visual features for visual question answering, Neurocomputing source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“It was found that any foreign substance in the lung cavity resulted in misclassification of images, and sometimes a correct classification was due to the wrong reasons [164].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [164] Lee, 2020, Evaluation of scalability and degree of fine-tuning of deep convolutional neural networks for COVID-19 screening on chest X-ray images using explainable deep-learning algorithm, J. Personalized Med. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The model performance was more accurate on CT compared to CXR images, and had an accuracy of 95.61% in COVID-19 detection.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [158] Panwar, 2020
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The quote only partly states this claim.
In the paper
“Thus, the gradient based methods lack in generalizability and robustness [85].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.3. Visualization by gradient/backpropagation
What the sources say
- Not checked yetcites [85] Zhang, 2018, Top-down neural attention by excitation Backprop, Int. J. Comput. Vis. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Compared to qualitative approaches through visual observations and comparisons, quantitative approaches can provide more objectivity and reliability in the process.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.6. Evaluation of explainability
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“The framework was qualitatively evaluated on MNIST dataset with different types of human users and received very positive feedback [124].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.5. Libraries, tools, and frameworks
What the sources say
- Not checked yetcites [124] Spinner, 2020, IEEE Trans. Visual. Comput. Graph.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“It was shown through visualizations that the features are not random or uninterpretable patterns but rather have class discrimination and increasing invariance between layers [87].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.3. Visualization by gradient/backpropagation
What the sources say
- Not checked yetcites [87] Zeiler, 2014, Visualizing and understanding convolutional networks
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“A visual explainable system Similarity Difference and Uniqueness Method (SIDU) was proposed to estimate pixel saliency by extracting feature maps from the last CNN layer, and creating a difference mask to form the visual explanation of the prediction [76].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.2. Visualization by perturbation
What the sources say
- Not checked yetcites [76] S. M. Muddamsetty, N. S. J. Mohammad and T. B. Moeslund, “SIDU: similarity difference and uniqueness method for explainable AI,” in IEEE International Conference on Image Processing (ICIP).
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“The proposed scheme was evaluated on clinical and general datasets and showed better visual explanations with a greater trust for the human expert [76].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.2. Visualization by perturbation
What the sources say
- Not checked yetcites [76] S. M. Muddamsetty, N. S. J. Mohammad and T. B. Moeslund, “SIDU: similarity difference and uniqueness method for explainable AI,” in IEEE International Conference on Image Processing (ICIP).
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“COVID-19 detection on CXR images with a fine tuned EfficientNet-B7 achieved an accuracy of 95.01% and AUC scores of 0.950, which were better in comparison to many other DNN approaches for COVID-19 detection [161].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [161] Chetoui, 2021, Deep efficient neural networks for explainable COVID-19 detection on CXR images
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Despite the large number of existing explainable models, there is no agreement on techniques for measuring or evaluating the explainability performance of an AI model.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 3.6. Evaluation of explainability
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“The two explainability approaches are complementary, in that, saliency map highlights the pixels with significant contribution to the model prediction and explanation-by-examples selects the training images likely for prediction in Bayesian Teaching [8].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > Model Agnostic Globally Interpretable Explanations (MAGIX)
What the sources say
- Not checked yetcites [8] Folke, 2021, Explainable AI for medical imaging: explaining pneumothorax diagnoses with Bayesian teaching, Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications III
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“An adapted ResNet architecture was used on MR CLEAN Registry dataset of CT angiography images and determined that the automated DNN method was better than radiological image biomarkers for stroke outcome prediction and was helpful in the treatment selection [136].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [136] Hilbert, 2019, Data-efficient deep learning of radiological image data for outcome prediction after endovascular treatment of patients with acute ischemic stroke, Comput. Biol. Med. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The feature map showed positive values for the portions with significant influence on the disease diagnosis [236].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [236] Yoo, 2021, Diagnosis support model of cardiomegaly based on CNN using ResNet and explainable feature map, IEEE Access source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“The CT images from multiple sources were mixed to avoid overfitting and bias [175].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [175] Wu, 2021, JCS: an explainable COVID-19 diagnosis system by Joint classification and segmentation, IEEE Trans. Image Process. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The XAI can help in identifying all such factors which form the basis of a particular decision.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“The method provided insights of the high-level feature importance in the trained CNN model [224].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [224] Couteaux, 2019, Towards interpretability of segmentation networks by analyzing deepdreams
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The quote only partly states this claim.
In the paper
“The proposed architecture provided better insights into the model's decisions and achieved better performance compared to the considered approaches.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [182] Kumar, 2019, SISC: end-to-end interpretable discovery radiomics-driven lung cancer prediction via stacked interpretable sequencing cells, IEEE Access source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The ensemble was found to improve the classification of the system and achieved an AUC of 90.29% in predicting nodules that would probably be diagnosed as cancerous in two years.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [179] Paul, 2020, Convolutional Neural Network ensembles for accurate lung nodule malignancy prediction 2 years in the future, Comput. Biol. Med. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“However, the dataset was small with low-resolution images [167].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [167] Liz, 2021, Ensembles of Convolutional Neural Network models for pediatric pneumonia diagnosis, Future Generat. Comput. Syst. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“Emergent Symbolic language are the cornerstones of Good Old-Fashioned AI (GOFAI) systems which are explainable [252].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [252] Chowdhury, 2021, Emergent Symbolic language based deep medical image classification
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“CXR images were used for lung cancer prediction using the proposed X-Caps XAI technique generating malignancy scores with explanations at human-level that were better interpretable by the radiologists [15].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [15] R. LaLonde, D. Torigian and U. Bagci, “Encoding visual attributes in capsules for explainable medical diagnoses,” in Medical Image Computing and Computer Assisted Intervention – MICCAI 2020. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The JSC identified and segmented the opacification areas, a basic CT feature of COVID-19 [175].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [175] Wu, 2021, JCS: an explainable COVID-19 diagnosis system by Joint classification and segmentation, IEEE Trans. Image Process. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Emergent languages involve multi-agent communication games with a sender/ receiver coordinate to complete a task and have been shown to enhance the neural network capabilities [252].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
What the sources say
- Not checked yetcites [252] Chowdhury, 2021, Emergent Symbolic language based deep medical image classification
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“The model performance was found comparable to similar 3D CNN models and had a high diagnostic performance for acute appendicitis detection [227].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [227] Rajpurkar, 2020, AppendiXNet: deep learning for diagnosis of appendicitis from a small dataset of CT exams using video pretraining, Sci. Rep. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The proposed radiomic sequencer achieved correct predictions that can potentially improve clinical adoption [182].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [182] Kumar, 2019, SISC: end-to-end interpretable discovery radiomics-driven lung cancer prediction via stacked interpretable sequencing cells, IEEE Access source
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“The proposed system achieved an AUC of 0.9627 and a detection accuracy of 0.9272 [240].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [240] Huo, 2019, Coronary calcium detection using 3D attention identical dual deep network based on weakly supervised learning, Med. Imaging Image Process.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This can also help in overcoming the challenges associated with the limited information present in an X-ray and actually highlighting how the density, shape, size and location of abnormalities affect the final diagnosis.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.1. Conventional imaging/X-rays
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may state two separate claims as one.
In the paper
“RFGB found simpler rules for modelling the features and could be easily compared to a large single highly accurate model [219].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [219] Kunapuli, 2018, A decision-support tool for renal mass classification,” journal of digital imaging, J. Digit. Imag. source
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“An average classification accuracy of 94.78% was achieved.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.2. Computed Tomography (CT)
What the sources say
- Not checked yetcites [177] Xu, 2019, Pulmonary textures classification via a multi-scale attention network, IEEE journal of biomedical and health informatics source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The quote only partly states this claim.
In the paper
“It was shown that the hierarchical diagnosis performed better than the other models for classification, and identified the relevant image regions with attention maps [217].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.5. Optical imaging 4.5.1. Dermatology
What the sources say
- Not checked yetcites [217] Barata, 2021, Explainable skin lesion diagnosis using taxonomies, Pattern Recogn. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The ABCD-rule of dermoscopy is known to outperform other methods [212].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.5. Optical imaging 4.5.1. Dermatology
What the sources say
- Not checked yetcites [212] Stieler, 2021, Towards domain-specific explainable AI:model interpretation of a skin image classifier using a human approach
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“The study concluded that by embedding model explainability in the early model training can save time and help clinicians in using a model's prediction for the clinical decisions [152].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [152] Windisch, 2020, Implementation of model explainability for a basic brain tumor detection using convolutional neural networks on MRI slices, Neuroradiology source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may state two separate claims as one.
In the paper
“Four CNN models were used and visualizations by heatmaps determined the important features which were compared against the expert's judgement [187].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.5.2. Diabetic retinopathy
What the sources say
- Not checked yetcites [187] Reguant, 2021, Understanding inherent image features in CNN-based assessment of diabetic retinopathy, Sci. Rep. source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“From the aggregated explanations it was concluded that cisterns and ventricles governed the age predictions [146].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [146] Levakov, 2020, From a deep learning model back to the brain—identifying regional predictors and their relation to aging, Hum. Brain Mapp. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“It can be difficult to control the image quality due to manual operation, operator dependency and acoustic shadows [253].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.4. Ultrasound imaging
What the sources say
- Not checked yetcites [253] Komatsu, 2021, Towards clinical application of artificial intelligence in ultrasound imaging, Biomedicines source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“CA-Net also provided better explainability by visualizing the attention weight maps, although it had around 15 times smaller model size [153].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [153] Lopatina, 2020, Investigation of deep-learning-driven identification of multiple sclerosis patients based on susceptibility-weighted images using relevance analysis, Front. Neurosci. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The significant features for the prediction were identified by isolating the neuron activation patterns [192].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.5.2. Diabetic retinopathy
What the sources say
- Not checked yetcites [192] Niu, 2021, Explainable diabetic retinopathy detection and retinal image generation, IEEE journal of biomedical and health informatics
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Grad-CAM was used for visualisations and helped to identify the tumor location.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [152] Windisch, 2020, Implementation of model explainability for a basic brain tumor detection using convolutional neural networks on MRI slices, Neuroradiology source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The quote only partly states this claim.
In the paper
“The video data was converted to a barcode type timeline to aid the explainability which made it easier to identify the substructures [243].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.4. Ultrasound imaging
What the sources say
- Not checked yetcites [243] Komatsu, 2021, Detection of cardiac structural abnormalities in fetal ultrasound videos using deep learning, Appl. Sci. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The datasets EyePACS and DIARETDB1 were used and although different CNN models had a consistent classification performance, the image feature disagreement between different architectures was 70% as the highest value [187].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.5.2. Diabetic retinopathy
What the sources say
- Not checked yetcites [187] Reguant, 2021, Understanding inherent image features in CNN-based assessment of diabetic retinopathy, Sci. Rep. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The choice of the MRI slice/image is important and the authors proposed selecting the best MRI slice for improved model performance [220].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [220] Hassan, 2022, Prostate cancer classification from ultrasound and MRI images using deep learning based Explainable Artificial Intelligence, Future Generat. Comput. Syst. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“Breast ultrasound and MRI can provide additional information for the mammographic diagnosis [254].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.4. Ultrasound imaging
What the sources say
- Not checked yetcites [254] Le, 2019, Artificial intelligence in breast imaging, Clin. Radiol. source
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“The system achieved an accuracy of 95.29% and Receiver Operating Characteristics (ROC)-AUC of 0.98 for both classes.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.3. Magnetic resonance imaging (MRI)
What the sources say
- Not checked yetcites [148] Chakraborty, 2020, Detection of Parkinson's disease from 3T T1 weighted MRI scans using 3D convolutional neural network, Diagnostics source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“These postulates provided the association between the disease and the pathogens [192].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.5.2. Diabetic retinopathy
What the sources say
- Not checked yetcites [192] Niu, 2021, Explainable diabetic retinopathy detection and retinal image generation, IEEE journal of biomedical and health informatics
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“The results demonstrated a high classification accuracy and explainability [191].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.5.2. Diabetic retinopathy
What the sources say
- Not checked yetcites [191] Quellec, 2021, ExplAIn: explanatory artificial intelligence for diabetic retinopathy diagnosis, Med. Image Anal. source
boosted- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“Similarly, the use of frameworks combining the automated segmentation and classification could for example, can use the segmentation results to guide the classification process [214].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.1. XAI applications in medical imaging
What the sources say
- Not checked yetcites [214] Xie, 2020, A mutual bootstrapping model for automated skin lesion segmentation and classification, IEEE Trans. Med. Imag. source
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This can provide a degree of confidence in the model but without a standard way of evaluating the different models and their outcomes cannot be trusted for medical imaging.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.1. XAI applications in medical imaging
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“Large and high-quality data is required by DNN for better results without which the model may overfit, that is, becomes closely aligned to the limited set of datapoints [265].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [265] Tayarani-N, 2020
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The XAI techniques can help in understanding the predictions of a deep learning model yet there are challenges to be overcome before these techniques become mainstream for healthcare applications.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5. Challenges and future directions
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper may give no reason why it holds. It may state two separate claims as one.
In the paper
“Most of the studies relating to cancer detection use a post-hoc explainability technique whereas a development of intrinsic explanation techniques would be advantageous for adoption and explainability [258].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.2. Understanding and improving explainability
What the sources say
- Not checked yetcites [258] Gulum, 2021, A review of explainable deep learning cancer detection models in medical imaging, Appl. Sci. source
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The system was considered useful for mass screening as it could be run without an Internet connection on a mobile device [193].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.5.2. Diabetic retinopathy
What the sources say
- Not checked yetcites [193] Martins, 2020, Offline computer-aided diagnosis for Glaucoma detection using fundus images targeted at mobile devices, Comput. Methods Progr. Biomed. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The adoption of AI in clinical practice would thus take more effort in making the XAI techniques to be more comprehensible.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.2. Understanding and improving explainability
hedgeddirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The creation of quantitative metrics for model comparison for explainability will give rise to the successful application of relevant explainability techniques to different image modalities.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.1. XAI applications in medical imaging
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“XAI techniques using multimodal DNN models can help with the hidden patterns related to each modality in segmentation and classification tasks.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.1. XAI applications in medical imaging
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“Without sufficient data, the AI models can suffer from overfitting and the data from different hospitals differ due to the variety of data collection policies and protocols [265].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [265] Tayarani-N, 2020
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Ophthalmology is one of the most widely used application of AI and there are efforts towards explainability in this area but are still at very basic level.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.5.2. Diabetic retinopathy
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
In the paper
“Explanatory Interactive Learning (XIL) which adds the human in the model training loop [259] can also help to increase the adoption of the XAI techniques.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.1. XAI applications in medical imaging
What the sources say
- Not checked yetcites [259] Schramowski, 2020, Making deep neural networks right for the right scientic reasons by interacting with their explanations, Nat. Mach. Intell. source
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The quality of images improves the performance of DNN algorithms in correctly predicting a disease and therefore can reduce the explainability gap.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
None of the sources cited for this claim have been checked yet.
In the paper
“The term 'explanation' as used in XAI is not very well defined [60,260,261].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.2. Understanding and improving explainability
What the sources say
- Not checked yetcites [60] Páez, 2019, The pragmatic turn in explainable artificial intelligence (XAI), Minds Mach. source
- Not checked yetcites [260] Payrovnaziri, 2020, Explainable artificial intelligence models using real-world electronic health record data: a systematic scoping review, J. Am. Med. Inf. Assoc. source
- Not checked yetcites [261] Gillmann, 2021, Ten open challenges in medical visualization, Comput. Graphics Appl. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“A universally agreed definition of explanations is lacking and there has been some efforts in this direction by developing mathematical methods [25].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.2. Understanding and improving explainability
What the sources say
- Not checked yetcites [25] W. Samek and K. R. Müller, “Towards explainable artificial intelligence,” in Explainable AI: Interpreting, Explaining and Visualizing Deep Learning. Lecture Notes in Computer Science, vol. vol. 11700, Springer, Cham..
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“This differs for the different perspectives of a patient, clinician and engineer and the current definitions lack consistency with explainable AI representing a false hope to provide a patient level decision support [19].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.2. Understanding and improving explainability
What the sources say
- Not checked yetcites [19] Ghassemi, 2021, The false hope of current approaches to explainable artificial intelligence in health care, The Lancet Digital Health source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“There is no agreement on what is meant by 'explanation' [262].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.2. Understanding and improving explainability
What the sources say
- Not checked yetcites [262] Guidotti, 2018, A survey of methods for explaining black box models, ACM Comput. Surv. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The development of DNN and then having attempts to identify their internal workings and decision-making process has resulted in development of many useful techniques.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.2. Understanding and improving explainability
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Such metrics can also meet the needs for the regulations and standards in assessing and approving the XAI techniques for the clinical practice.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.1. XAI applications in medical imaging
plain- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
None of the sources cited for this claim have been checked yet.
In the paper
“Various techniques can be employed to address the lack of data, such as, data augmentation [34,37].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [34] Karthik, 2021, AI for COVID-19 detection from radiographs: incisive analysis of state of the art techniques, IRBM source
- Not checked yetcites [37] Singh, 2021
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“The ultimate responsibility would still be with the human radiologists [5].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.5. Human-in-the-loop
What the sources say
- Not checked yetcites [5] Strudel, 2021, Segmenter: transformer for semantic segmentation
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Rests on an assumption: The final decision for image-based disease diagnosis would rest with the specialist, whereas XAI model insights should be used only to aid the decision.
In the paper
“It is aimed at translating the algorithmic model research to the actual practice in domains such as healthcare [275].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
What the sources say
- Not checked yetcites [275]
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Such automated workflows are expected to make the process of incorporating XAI easier into clinical decisions by making it possible to select the best visualizations in response to a few simple questions.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“DNN systems have sensitivity similar to radiologists but often times these DNN can detect different lesion characteristics [6].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.5. Human-in-the-loop
What the sources say
- Not checked yetcites [6] Chan, 2020, Deep learning in medical image analysis, Adv. Exp. Med. Biol. source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The funders and governments are emphasizing an open sharing of data for AI model training but most data are not representative, interoperable, and sufficiently diverse [39].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [39] Norori, 2021, Addressing bias in big data and AI for health care: a call for open science, Patterns source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“The model architecture has many hyperparameters that need to be tried and re-tried for designing an architecture with an optimum performance.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The automation will not only make these XAI techniques widely available but will enable medical practitioners to make better and informed decisions by leveraging the insights through XAI.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The use of sufficient, representative, and diverse type of data for DNN models can help to reduce bias and enhance generalizability.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“A framework for trustworthy and XAI combined selected features of trusted oracles, blockchain and smart contracts to reduce adversarial attacks and bias [11].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
What the sources say
- Not checked yetcites [11] M. Nassar, K. Salah, M. ur Rehman and D. Svetinovic, “Blockchain for explainable and trustworthy artificial intelligence,” Wiley Interdisciplinary Reviews: Data Min. Knowl. Discov., vol. 10, no. 1. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“End to end automated XAI applications, such as, combining segmentation and classification, will help XAI adoption in clinical practice.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“Certification of DNN models by agencies external to the developer or user could help to eliminate risk of unauthorized changes to how the system works and can prove beneficial in case of litigation to prove the correctness of the system [49].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.6. Incorporating stakeholders in XAI
What the sources say
- Not checked yetcites [49] Gryz, 2021, Black box algorithms and the rights of individuals: no easy solution to the ‘explainability’ problem, Internet Policy Review source
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“In order to use AI successfully requires a multi-disciplinary effort encompassing guarantees of ethical, legal and social aspects of AI and increasing awareness [5].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.6. Incorporating stakeholders in XAI
What the sources say
- Not checked yetcites [5] Strudel, 2021, Segmenter: transformer for semantic segmentation
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.
In the paper
“Few-shot learning can work with very few training samples [13].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [13] McCarthy, 2021, Enterprise imaging and big data: a review from a medical physics perspective, Phys. Med. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“The CNN fixation method was applied to detect an adversarial attack on sets of images with a clean and perturbed image, and even despite the labels being changed, the proposed method could still localize the regions correctly in both images [69].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
What the sources say
- Not checked yetcites [69] Mopuri, 2019, An unraveling approach to visualize the discriminative image regions, IEEE Trans. Image Process. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“This can in future help towards models' certifications and standardizations such as by Food and Drug Administration (FDA) [274].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
What the sources say
- Not checked yetcites [274] Proposed Regulatory Framework for Modifications to Artificial Intelligence/Machine Learning (AI/ML)-Based Software as a Medical Device (SaMD) -Discussion Paper and Request for Feedback”.
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.
In the paper
“The decisions affecting patients' lives cannot be entrusted fully to an algorithm”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.5. Human-in-the-loop
What the sources say
- Not checked yetcites [261] Gillmann, 2021, Ten open challenges in medical visualization, Comput. Graphics Appl. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence.
- Rests on an assumption: DNN predictions comparable to human experts are not sufficient to fully delegate medical diagnosis to an algorithm.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“The adversarial inputs comprising a modified image can drive a model to misclassify and is termed as an adversarial attack.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Biomedical imaging datasets sometime suffer from noise, whereas CT images are commonly affected by noise [265].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [265] Tayarani-N, 2020
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“A counterfactual approach, Global Explanations for Bias Identification (GEBI) was proposed for identification of bias in the data for image classification although it can be applied to other problems [279].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
What the sources say
- Not checked yetcites [279] Mikołajczyk, 2020
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“AutoML is a step towards the direction of automatically creating an optimum architecture and making it possible for the architectures to be reproducible and comparable.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This would provide opportunities for measuring the accuracy and explainability of DNN against the same dataset and mitigate the different biases in the model and data.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
hedgeddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“Undetectable adversarial attacks can be devised to change the importance maps dramatically and interpretations of breast ultrasound images were shown to be susceptible to adversarial attacks similar to other image modalities [203].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
What the sources say
- Not checked yetcites [203] Rasaee, 2021, Explainable AI and susceptibility to adversarial attacks: a case study in classification of breast ultrasound images
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The use of XAI techniques such as visualizations are helpful to identify the model bias and an experienced clinicians can identify any unsupported or bad DNN decisions.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
In the paper
“Similarly, Google's What-If Tool can be used for optimizing the model's AI fairness [276].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
What the sources say
- Not checked yetcites [276]
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Most ML studies on medical images use trained models that are often based on a single unit or hospital which brings in a selective bias and sometimes inclusion of low-quality scans [273].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.4. Model centric XAI
What the sources say
- Not checked yetcites [273] Hampe, 2019
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
None of the sources cited for this claim have been checked yet.
In the paper
“Additional realistic data could be generated through GANs using the latent data distributions [13,34,37].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.3. Data-centric XAI
What the sources say
- Not checked yetcites [13] McCarthy, 2021, Enterprise imaging and big data: a review from a medical physics perspective, Phys. Med. source
- Not checked yetcites [34] Karthik, 2021, AI for COVID-19 detection from radiographs: incisive analysis of state of the art techniques, IRBM source
- Not checked yetcites [37] Singh, 2021
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Explainable AI coupled with deep neural network models can help with both the detection and diagnosis of disease, by providing additional insights to the clinician.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 6. Conclusion
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim.
- Rests on an assumption: Trained DNN models can provide predictions comparable to human experts in a fraction of the time.
In the paper
“The black box and opaque nature of deep neural networks precludes their usefulness despite excellent performance of multimodal methods [288].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.7. Hybrid/fusion imaging for XAI
What the sources say
- Not checked yetcites [288] Joshi, 2021, A review on explainability in multimodal deep neural nets, IEEE Access source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Two significant issues faced for multimodal imaging are the class imbalances and lack of sufficient labelled data [35].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.7. Hybrid/fusion imaging for XAI
What the sources say
- Not checked yetcites [35] O'Sullivan, 2020, Developments in AI and machine learning for neuroimaging
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We see a lack of algorithm standardization for biomedical imaging with diverse and independent efforts for improvements.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 6. Conclusion
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper may give no reason why it holds.
- Rests on an assumption: Currently, image-based diagnosis and prognosis in biomedical imaging require expert opinion of practitioners who are already overburdened.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The overcoming of challenges to clinical adoption of XAI will foster growth of the field, and result in better health and safety measures for the patients through timely interventions by building trust in the model's predictions.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 6. Conclusion
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper may give no reason why it holds. It may state two separate claims as one.
- Rests on an assumption: Increasingly, there are health and legal regulations that mandate approval and compliance for deep learning techniques.
In the paper
“It was shown with the proposed deconvnet that the model visualizations can be used to determine model problems and can help improve the models [87].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.9. Explainability to improve models
What the sources say
- Not checked yetcites [87] Zeiler, 2014, Visualizing and understanding convolutional networks
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“Graph Neural Networks can learn from graph structures and can represent information capturing different types of relationships rather than just an ordered grid structure [100].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.8. Mathematical and graph based XAI approaches
What the sources say
- Not checked yetcites [100] Holzinger, 2021, Towards multi-modal causability with Graph Neural Networks enabling information fusion for explainable AI, Inf. Fusion source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“The mathematical approaches can improve the interpretability of models, however, may not be suitable as concrete examples for the enduser [63].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.8. Mathematical and graph based XAI approaches
What the sources say
- Not checked yetcites [63] Heuillet, 2021, Explainability in deep reinforcement learning, Knowl. Base Syst. source
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Federated learning is also promising for biomedical applications with different hospitals contributing their data to train and improve a deep learning model without violating data privacy.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 6. Conclusion
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“Influence functions have basis in statistics and were demonstrated to be useful for generation of visually similar training set attacks, determining dataset errors, understanding, and debugging the model behavior [104].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 5.9. Explainability to improve models
What the sources say
- Not checked yetcites [104] Koh, 2017
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We consider a human-in-the-loop as a pre-requisite for any XAI application based clinical decision.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 6. Conclusion
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper may give no reason why it holds.
- Rests on an assumption: Trained DNN models can provide predictions comparable to human experts in a fraction of the time.
- Rests on an assumption: DNN predictions comparable to human experts are not sufficient to fully delegate medical diagnosis to an algorithm.
- Rests on an assumption: Human expertise cannot be dispensed, and only simple and time-consuming tasks should be directed to AI.
- Rests on an assumption: The final decision for image-based disease diagnosis would rest with the specialist, whereas XAI model insights should be used only to aid the decision.
The paper relies on this without support.
XAI techniques help to understand model predictions which help develop trust in the system, accelerate the disease diagnosis, and meet adherence to regulatory requirements.
“Explainable AI (XAI) techniques help to understand model predictions which help develop trust in the system, accelerate the disease diagnosis, and meet adherence to regulatory requirements.”
The paper relies on this without support.
There is often an explainability gap between the model's prediction and its explanations being understandable to a human.
“However, there is often an explainability gap between the model's prediction and its explanations being understandable to a human.”
The paper relies on this without support.
Incorporating explainability into a model takes time and effort, and may affect its prediction accuracy.
“Incorporating explainability into a model takes time and effort, and may affect its prediction accuracy.”
The paper relies on this without support.
The use of XAI techniques can reduce the workload of the clinicians, help to make better diagnosis, and reduce the complications associated with a false positive and a false negative outcome.
“The use of XAI techniques can reduce the workload of the clinicians, help to make better diagnosis, and reduce the complications associated with a false positive and a false negative outcome.”
The paper relies on this without support.
Medical practitioners are more likely to employ and trust a system if they can understand the reasoning of a prediction by a model.
“They are more likely to employ and trust a system if they can understand the reasoning of a prediction by a model.”
The paper relies on this without support.
Patients require assurance that the XAI methods can be trusted, and are providing a valid and consistent prediction.
“Patients require assurance that the XAI methods can be trusted, and are providing a valid and consistent prediction.”
The paper relies on this without support.
DNN can provide very accurate predictive performance but are hard to interpret or explain.
“DNN can provide very accurate predictive performance but owing to a large number of trainable parameters, layers, and the performance determined through the hyperparameters, these are hard to interpret or explain.”
The paper relies on this without support.
CXR is the most commonly used imaging modality for detecting COVID-19 or lung infections, but it is less accurate than CT.
“CXR is the most commonly used imaging modality for detecting COVID-19, or lung infections, however, in comparison to CT, it is less accurate.”
The paper relies on this without support.
MRI is the most favored imaging modality for neurology and is particularly useful for neurodegenerative diseases.
“MRI is the most favored imaging modality for neurology and particularly useful for the neurodegenerative diseases.”
The paper relies on this without support.
Ophthalmologists are interested in finding the parameters responsible for a specific finding, rather than just knowing the final diagnosis.
“The ophthalmologists are interested in finding out the parameters responsible for a specific finding, rather than just knowing the final diagnosis.”
The paper relies on this without support.
In fundus images, the actual variations in landmarks (blood vessels, optic disc, fovea/macula) along with appearances of any new lesion are important.
“In fundus images, the actual variations in landmarks i.e. blood vessels, optic disc and fovea/macula region along with appearances of any new lesion are important.”
The paper relies on this without support.
For OCT images, it is important to look into layer variants and lesions appearing in those layers.
“For OCT images, it is important to look into layer variants and lesions appearing in those layers.”
The paper relies on this without support.
The labelled datasets used to train DNNs are often from multiple sources or maintained by different organisations owning the data.
“The labelled datasets used to train DNNs are often from multiple sources or maintained by different organisations owning the data.”
The paper relies on this without support.
One of the stated advantages of DNN is automated learning of image features and overcoming problems associated with hand-crafted features.
“One of the stated advantages of DNN is that of an automated learning of the image features and overcoming the problems associated with the hand-crafted features.”
The paper relies on this without support.
Trained DNN models can provide predictions comparable to human experts in a fraction of the time.
“The trained DNN model can provide predictions comparable to human expert in a fraction of the time”
The paper relies on this without support.
DNN predictions comparable to human experts are not sufficient to fully delegate medical diagnosis to an algorithm.
“however, this is not sufficient to fully delegate the medical diagnosis to an algorithm”
The paper relies on this without support.
Human expertise cannot be dispensed, and only simple and time-consuming tasks should be directed to AI.
“However, human expertise cannot be dispensed and only simple, and time-consuming tasks should be directed to AI.”
The paper relies on this without support.
The ultimate beneficiary of model explainability is a human user.
“The ultimate beneficiary of the model explainability is a human user.”
The paper relies on this without support.
The final decision for image-based disease diagnosis would rest with the specialist, whereas XAI model insights should be used only to aid the decision.
“Similar to the role of an autonomous car driver, the final decision for image-based disease diagnosis would rest with the specialist whereas the XAI model insights should be used only to aid the decision.”
The paper relies on this without support.
Currently, image-based diagnosis and prognosis in biomedical imaging require expert opinion of practitioners who are already overburdened.
“Currently, the image-based diagnosis and prognosis in the different disciplines of biomedical imaging such as radiology and ophthalmology require expert opinion of practitioners who are already overburdened.”
The paper relies on this without support.
Increasingly, there are health and legal regulations that mandate approval and compliance for deep learning techniques.
“Increasingly, there are health and legal regulations that mandate approval and compliance for the deep learning techniques which can be helped using XAI techniques.”
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“The “black box” nature of DL remains a fundamental barrier to clinical trust [153,154,155,156,157].”
From Artificial Intelligence in Clinical Oncology: From Productivity Enhancement to Creative Discovery (2025) · cites this paper as [153]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.6. Nuclear imaging
“SHAP was found to generate better heatmaps out of the six interpretation methods for the four selected DNN models.”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.6. Nuclear imaging
“The Guided backpropagation performed best amongst the six selected interpretation methods [138].”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 4.7. clinical/observational data
“Transthoracic Echocardiography images were used for an automatic Carpentier's functional classification of mitral valve diseases with eight deep CNN models [246].”
§ Survey of explainable artificial intelligence techniques for biomedical imaging with deep neural networks > 6. Conclusion
“| | ECG video data | Deep CNN, Grad-CAM | Morphological classification of mitral valve disease using eight | [246] |”
Citations
Cites
An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., et al. (2020) · arXiv · doi:10.48550/arxiv.2010.11929 · Cited by 47,472
A Unified Approach to Interpreting Model Predictions
Lundberg, S., & Lee, S.-I. (2017) · arXiv · doi:10.48550/arxiv.1705.07874 · Cited by 23,707
Proceedings of IEEE Conference on Computer Vision and Pattern Recognition
(1993) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.1993.341010 · Cited by 14,475
Gauriat, C.-M., Pencolé, Y., Ribot, P., et al. (2024) · HAL · doi:10.4230/oasics.dx.2024.27 · Cited by 12,623
Barredo Arrieta, A., Díaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, S., Barbado, A., Garcia, S., Gil-Lopez, S., Molina, D., Benjamins, R., et al. (2020) · Information Fusion · doi:10.1016/j.inffus.2019.12.012 · Cited by 10,423 · cited 7×
“2, where non-linear models with better performance, such as DNNs are less explainable, and linear models which are least accurate, such as decision trees, are more explainable [2,21,45].”
BackgroundInfluentialThe graph neural network model
Scarselli, F., Gori, M., Tsoi, A. C., Hagenbuchner, M., & Monfardini, G. (2009) · IEEE Transactions on Neural Networks · doi:10.1109/tnn.2008.2005605 · Cited by 9,582 · cited 3×
“The model and the learning al-gorithm, output functions, comparison with Random Walks and Recursive Neural Networks, and complexity were derived mathematically [99].”
BackgroundMethodInfluentialShow, Attend and Tell: Neural Image Caption Generation with Visual Attention
Xu, K., Ba, J., Kiros, R., Cho, K., Courville, A., Salakhutdinov, R., Zemel, R., & Bengio, Y. (2015) · arXiv · doi:10.48550/arxiv.1502.03044 · Cited by 8,248 · cited 5×
“The proposed technique exploited learned attention for better interpretability, well-aligned to human intuition on the datasets with METEOR and BLEU metrics [92].”
BackgroundMethodInfluentialPeeking Inside the Black-Box: A Survey on Explainable Artificial Intelligence (XAI)
Adadi, A., & Berrada, M. (2018) · IEEE Access · doi:10.1109/access.2018.2870052 · Cited by 6,294 · cited 4×
“A study provided a taxonomy and evaluation techniques for XAI [21].”
BackgroundMethodInfluentialProceedings of the 25th international conference on Machine learning
Cohen, W. W., McCallum, A., & Roweis, S. T. (2008) · Cited by 5,691
Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps
Simonyan, K., Vedaldi, A., & Zisserman, A. (2013) · arXiv · doi:10.48550/arxiv.1312.6034 · Cited by 5,370 · cited 2×
“It used saliency maps for both convolution and deconvolution, and a gradient-based visualization technique [77].”
Method
Research works printed but not linked: 10
- Durán, 2021, vol. 297
- A. Preece, D. Harborne, D. Braines, R. Tomsett and S. Chakraborty, “Stakeholders in Explainable AI,” [Online]. Available: https://doi.org/10.48550/arXiv.1810.00184. [Accessed 15 September 2022]. · DOI · Paper page
- V. Petsiuk, A. Das and K. Saenko, “RISE: Randomized Input Sampling for Explanation of Black-box Models,” [Online]. Available: https://doi.org/10.48550/arXiv.1806.07421. [Accessed 15 September 2022]. · DOI
- Adebayo, 2018, Sanity checks for saliency maps, Adv. Neural Inf. Process. Syst., 31
- Alber, 2019, iNNvestigate neural networks, J. Mach. Learn. Res., 20, 1
- Spinner, 2020, IEEE Trans. Visual. Comput. Graph., 26, 1064
- Panwar, 2020, vol. 140
- Liz, 2021, Ensembles of Convolutional Neural Network models for pediatric pneumonia diagnosis, Future Generat. Comput. Syst., 122, 220, 10.1016/j.future.2021.04.007 · DOI
- Kim, 2018, ICADx: interpretable computer aided diagnosis of breast masses, Medical Imaging 2018: Computer-Aided Diagnosis, 10575, 450
- Hampe, 2019, vol. 6
Cited by
Sorted by citations
Muhammad, D., & Bendechache, M. (2024) · Computational and Structural Biotechnology Journal · doi:10.1016/j.csbj.2024.08.005 · Cited by 325
Dixon, D., Sattar, H., Moros, N., Kesireddy, S. R., Ahsan, H., Lakkimsetti, M., Fatima, M., Doshi, D., Sadhu, K., & Hassan, M. J. (2024) · Cureus · doi:10.7759/cureus.59954 · Cited by 249
“Reproduced under the terms of the Creative Commons CC-BY license from Nazir et al. [19].”
BackgroundMethodAmiri, Z., Heidari, A., Navimipour, N. J., Esmaeilpour, M., & Yazdani, Y. (2024) · Neural Computing and Applications · doi:10.1007/s00521-023-09366-3 · Cited by 191
Wani, N. A., Kumar, R., Mamta, ., Bedi, J., & Rida, I. (2024) · Information Fusion · doi:10.1016/j.inffus.2024.102472 · Cited by 172
Federated Learning for Medical Image Analysis with Deep Neural Networks
Nazir, S., & Kaleem, M. (2023) · Diagnostics · doi:10.3390/diagnostics13091532 · Cited by 163
“Medical image modalities, such as ultrasound, X-rays, Magnetic Resonance Imaging (MRI), play a crucial role in disease diagnosis, and are used for diagnosing human body ailments, diseases, and various forms of cancers [1,2].”
BackgroundSelf-citation
Metadata
- Authors
- Nazir, S., Dickson, D. M., & Akram, M. U.
- Affiliations
- Glasgow Caledonian University (GB)
- National University of Sciences and Technology (PK)
- Corresponding author
- Nazir, Sajid
- Venue
- Computers in Biology and Medicine
- Publisher
- Elsevier BV
- Published
- 2023-04
- Volume
- 156
- Pages
- 106668
- ISSN
- 0010-4825 (ISSN-L)
- OpenAlex
- W4321350922
- Semantic Scholar
- 77d942cfa2d4b9266080258943048825e7449ad2
- Cited by
- 283
- References
- 315
- Citations by year
- 2026: 70 · 2025: 104 · 2024: 70 · 2023: 37
- FWCI
- 5.83
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- Locations
- Computers in Biology and Medicine (publishedVersion, cc-by)
- PubMed (publishedVersion)
- ResearchOnline (acceptedVersion, cc-by)
- OSTI OAI (U.S. Department of Energy Office of Scientific and Technical Information) (submittedVersion)
- ResearchOnline (Glasgow Caledonian University) (submittedVersion, cc-by)
- Funders
- Glasgow Caledonian University (10386)
- Defense Advanced Research Projects Agency
- Global Challenges Research Fund
- Office of Defense Nuclear Nonproliferation
- Alzheimer's Disease Neuroimaging Initiative
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › COVID-19 diagnosis using AI
- Keywords
- explainable AI, biomedical imaging, deep neural network, medical image classification, diagnosis, model interpretability, regulatory compliance, black-box models
- MeSH terms
- Machine Learning, Algorithms, Artificial Intelligence, Diagnostic Imaging, Humans, Neural Networks, Computer
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-10-07, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-07 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
