The DOI matches its Crossref record; 119 of the 119 references it cites were found in a registry or an index; none of the 41 sources its claims cite is checked for support yet.
Medical image data augmentation: techniques, comparisons and interpretations
Open accessAbstract
Designing deep learning based methods with medical images has always been an attractive area of research to assist clinicians in rapid examination and accurate diagnosis. Those methods need a large number of datasets including all variations in their training stages. On the other hand, medical images are always scarce due to several reasons, such as not enough patients for some diseases, patients do not want to allow their images to be used, lack of medical equipment or equipment, inability to obtain images that meet the desired criteria. This issue leads to bias in datasets, overfitting, and inaccurate results. Data augmentation is a common solution to overcome this issue and various augmentation techniques have been applied to different types of images in the literature. However, it is not clear which data augmentation technique provides more efficient results for which image type since different diseases are handled, different network architectures are used, and these architectures are trained and tested with different numbers of data sets in the literature. Therefore, in this work, the augmentation techniques used to improve performances of deep learning based diagnosis of the diseases in different organs (brain, lung, breast, and eye) from different imaging modalities (MR, CT, mammography, and fundoscopy) have been examined. Also, the most commonly used augmentation methods have been implemented, and their effectiveness in classifications with a deep network has been discussed based on quantitative performance evaluations. Experiments indicated that augmentation techniques should be chosen carefully according to image types.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
40 Not checked yet · 26 No basis given · 25 Own interpretation · 13 Own result · 6 Own method
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
“However, the databases have imbalanced class distributions.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Materials and methods > Classification
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.
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
“Augmentation of patches or whole mammography images have usually been performed by scaling, noise addition, mirroring, shearing, and rotation techniques or using several combinations of them (Table 3).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of breast mammography 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.
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
“On the other hand, the combined technique consisting of salt and pepper noise addition and shearing is the least appropriate approach (Table 8).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Discussions
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.
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
“Experiments in this study indicated that the combined technique consisting of translation, shearing, and rotation is the most appropriate approach for the augmentation of breast mammography images in order to improve the classification of the images as normal, benign, and malignant.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Discussions
hedgeddirect- 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 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
“Experiments indicated that augmentation techniques should be chosen carefully according to image types.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Abstract
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.
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
“On the other hand, augmentation only by shearing is the least efficient approach (Table 7).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Discussions
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 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
“Therefore, there are a total of 888 images extracted from 1018 cases.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Materials and methods > Materials
plaindirectn = 1,018- 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
“It has been observed that usage of the augmented images obtained by the combination of rotation with shearing, and translation provides higher performance than the other augmentation techniques in the classifications of HGG and LGG cases from FLAIR types of brain images.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Discussions
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.
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 best augmentation method for different imaging modalities showing different organs has been determined according to quantitative values.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Introduction
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.
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
“It has been observed that usage of the augmented images obtained by the combination of translation and shearing techniques in the classification of images with nodules and without nodules provides the highest performance than the other augmentation methods.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Discussions
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. 29
- Probably not a substantial claim. 54
- The quote only partly states this claim. 3
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
“In those applications, different network architectures, datasets and number of images have been used, and also their performances have been evaluated with different metrics.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of breast mammography images
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
“The number of images in the categories is 974, 75, and 151, respectively.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Materials and methods > Materials
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.
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
“High performances have been obtained from those applications according to the reported results.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of breast mammography images
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. The quote only partly states this claim.
None of the sources cited for this claim have been checked yet.
In the paper
“The BRaTS database contains 285 images including HGG (210 images) and LGG (75 images) cases.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Materials and methods > Materials
plainn = 285- 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 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
“In the database, there are 1018 CT scans which belong to 1010 people.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Materials and methods > Materials
plainn = 1,010- 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
“It contains 410 images, of which only 107 images show breast masses, and the total number of mass lesions is 115 since an image may contain more than one lesion.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Materials and methods > Materials
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 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
“In those applications, different network structures, data sets and number of images have been used, and also their performances have been evaluated with different metrics.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of eye fundus images
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 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 number of benign and malignant masses is 52 and 63, respectively.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Materials and methods > Materials
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.
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
“High performances have been obtained from those applications according to the reported results.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of breast mammography images
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.
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
“Commonly used augmentation techniques with eye fundus images to boost the performance of deep learning based approaches are rotation, shearing, flipping, and translation (Table 4).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of eye fundus 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.
None of the sources cited for this claim have been checked yet.
In the paper
“Eye fundus images have been provided from the Messidor database (Decencière et al. 2014) including 1200 images categorized into three classes (0-risk, 1-risk, and 2-risk) according to the diabetic macular edema risk level.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Materials and methods > Materials
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.
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
“Thanks to the augmentation methods, deep network architectures’ learning and generalization ability can be improved, and the desired robustness property of the networks can be provided (Sect. 2).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Introduction
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.
- Rests on an assumption: Deep networks need training sets with many images covering all variations to achieve high generalization and robustness.
In the paper
“The chosen imaging modalities are widely used in medicine for many applications such as classification of brain tumors, lung nodules, and breast lesions (Meijering 2020).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Introduction
What the sources say
- Not checked yetcites Meijering 2020 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
“According to the segmentation or classification results presented in the literature, it is not possible to understand or compare the impacts or contributions of the used augmentation approaches on the results, and to decide the most appropriate augmentation method.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Introduction
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.
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
“Data augmentation is a common solution to overcome this issue”
§ Medical image data augmentation: techniques, comparisons and interpretations > Abstract
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.
- Rests on an assumption: Medical image datasets are usually limited in size.
- Rests on an assumption: Scarcity and imbalance in medical image data cause overfitting, bias, and inaccurate results.
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
“However, it is not clear which data augmentation technique provides more efficient results for which image type since different diseases are handled, different network architectures are used, and these architectures are trained and tested with different numbers of data sets in the literature.”
- Also stated in: § Introduction · plain
“It is not clear which data augmentation technique provides more efficient results for which image type since different diseases are handled, different network architectures are used, and these architectures are trained and tested with different numbers of data sets in the literature (Sect. 2).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Abstract
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.
- Also stated in: § Introduction · plain
None of the sources cited for this claim have been checked yet.
In the paper
“Therefore, translated images provide prevention of positional bias (Shorten and Khoshgoftaar 2019) and the models do not focus on properties in a single spatial location (Nalepa et al. 2019).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Commonly applied augmentation methods
What the sources say
- Not checked yetcites Shorten and Khoshgoftaar 2019 source
- Not checked yetcites Nalepa et al. 2019
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
“For instance, NeoVessel (NV)-like structures have been synthesized in a heuristic image augmentation (Ara´ujo et al. 2020) to improve detection of proliferative DR which is an advanced DR stage characterized by neovascularization.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of eye fundus images
What the sources say
- Not checked yetcites Ara´ujo et al. 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.
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
“Also, refinement and segmentation of the lesion regions before the application of an augmentation method is the most preferred way to provide good performance from the detection and classification of the lesions.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of breast mammography images
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.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“Additionally, to evaluate the effectiveness of those augmentation methods in classifications from the four types of images, a classifier model has been designed and classifications with the augmented images for each augmentation method have been performed with the same classifier model.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Introduction
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. 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
“Also, pixel intensity values have been used in some augmentation techniques since fundus images are colored images.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of eye fundus images
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
“For example, in a study (Shyamalee and Meedeniya 2022), a combined method including rotation, shearing, zooming, flipping, and shifting has been applied for glaucoma identification.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of eye fundus images
What the sources say
- Not checked yetcites Shyamalee and Meedeniya 2022
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
“For instance, contextual GAN, which generates new images by synthesizing lesions according to the context of surrounding tissue, has been applied to improve binary classification (as normal and malignant) by Wu et al. (2020b).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of breast mammography images
What the sources say
- Not checked yetcites Wu et al. (2020b)
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
“For instance, a set of augmentation methods (namely mirroring, zooming, and resizing) has been applied to improve automated classification of images into three categories (as normal, malignant, and benign) (Zeiser et al. 2020).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of breast mammography images
What the sources say
- Not checked yetcites Zeiser et al. 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.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“Five evaluation metrics have been used, and the results provided from each classification have been evaluated with the same evaluation metrics.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Introduction
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. It may state two separate claims as one.
In the paper
“The reason to choose this architecture is to utilize its advantage in addressing the gradient vanishing issue and high learning ability (He et al. 2016).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Materials and methods > Classification
What the sources say
- Not checked yetcites He et al. 2016
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.
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
“Those features are used in deep learning based methods being developed for different purposes like glaucoma identification, classification of diabetic retinopathy (DR), and identification of DR severity.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of eye fundus images
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.
- Rests on an assumption: Eye fundus features such as neovascularization, hemorrhages/microaneurysms, and exudates are symptoms of eye diseases.
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 mostly those techniques have been preferred in the literature, augmentation has also been provided by producing synthetic images. To generate synthetic images, Mixup method, which combines two randomly selected images and their labels, or modified Mixup, or GAN structures have been used (Table 1).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of brain MR images
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
“However, converting a color image to its grayscale version should be performed carefully because this conversion reduces performance by up 3% (Chatfield et al. 2014).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Commonly applied augmentation methods
What the sources say
- Not checked yetcites Chatfield et al. 2014
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
“In another work (Shen et al. 2021), both contextual and deep convolutional GAN have been used to generate synthetic lesion images with margin and texture information and achieve improvement in the detection and segmentation of lesions.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of breast mammography images
What the sources say
- Not checked yetcites Shen et al. 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.
It may state two separate claims as one.
In the paper
“The authors have also applied a channel-wise vessel augmentation using morphological transformations.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of eye fundus images
What the sources say
- Not checked yetcites Sun et al. 2021
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.
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
“Also, several GAN models have been used to augment lung CT images by generating new synthetic images (Table 2).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of lung CT images
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
“In another work (Agustin et al. 2020), to improve the robustness of DR detection performance, image augmentation has been performed by zooming and a contrast enhancement process known as Contrast Limited Adaptive Histogram Equalization (CLAHE).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of eye fundus images
What the sources say
- Not checked yetcites Agustin et al. 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.
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
“Therefore, the safety of this augmentation technique depends on the rotation degree.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Commonly applied augmentation methods
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.
- Rests on an assumption: Image labels may not always be preserved if the degree of rotation increases.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“Also, in this study, the most commonly used augmentation methods have been implemented using the same datasets.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Introduction
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.
None of the sources cited for this claim have been checked yet.
In the paper
“Only GAN based augmentations have also been used.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of breast mammography images
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.
None of the sources cited for this claim have been checked yet.
In the paper
“For instance, conditional GAN (Zhou et al. 2020) and deep convolutional GAN (Balasubramanian et al. 2020) based augmentations have been used to increase the performance of the methods developed for automatic classifications of DR according to its severity level.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of eye fundus images
What the sources say
- Not checked yetcites Zhou et al. 2020
- Not checked yetcites Balasubramanian et al. 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.
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
“To overcome those issues, several GAN variants were developed such as conditional, deep convolutional, and cycle GAN.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Commonly applied augmentation methods
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.
- Rests on an assumption: The main challenging issues of GAN-based image augmentation are generating high-quality images and maintaining training stability.
In the paper
“For instance, to achieve robust segmentation of retinal vessels, augmentation has been provided by a channel-wise random gamma correction (Sun et al. 2021).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of eye fundus images
What the sources say
- Not checked yetcites Sun et al. 2021
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
“Also, those techniques have been used together with GANs to provide augmentations with new synthetic images.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of breast mammography images
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
“For instance, flipping and deep convolutional GAN has been combined for binary classifications of images (as normal and with a mass) by Alyafi et al. (2020).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of breast mammography images
What the sources say
- Not checked yetcites Alyafi et al. (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.
None of the sources cited for this claim have been checked yet.
In the paper
“In some works, retinal features (such as lesion and vascular features) have been developed and added to new images.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of eye fundus images
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.
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 computerized methods being developed to assist medical professionals in various tasks [e.g., nodule or lung parenchyma segmentation, nodule classification, prediction of malignancy level, disease diagnosis like coronavirus (covid-19), etc.], augmentations of CT lung images are commonly provided by traditional techniques like rotation, flipping, and scaling (Table 2).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of lung CT images
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
“Also, it is not possible to determine augmentation methods’ contributions to those performances according to the presented results in those publications.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of breast mammography images
plain- Asserted: fails — The authors may not 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 authors may not put it forward as true. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. The quote only partly states this claim.
In the paper
“Similarly, noise and artifacts have been eliminated before the segmentation of the images into non-overlapping parts and augmentation (Karthiga et al. 2022).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of breast mammography images
What the sources say
- Not checked yetcites Karthiga et al. 2022
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.
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
“While doing a GAN based augmentation of lung CT scans, researchers usually tend to generate the area showing lung nodules in the whole image. Because the lung nodules are small and therefore it is more cost-effective than generating a larger whole image or background area.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of lung CT images
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
“Although it can be applied with vertical or/and horizontal axes, vertical flipping is rarely preferred since the bottom and top regions of an image may not always be interchangeable (Nalepa et al. 2019).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Commonly applied augmentation methods
What the sources say
- Not checked yetcites Nalepa et al. 2019
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.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“Therefore, to obtain unbiased classification results, an equal number of images have been taken from the databases for the classes.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Materials and methods > Classification
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.
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 those applications, different network architectures, datasets, and numbers of images have been used, and also their performances have been evaluated with different metrics. High performances have been obtained from those applications according to the reported results.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of brain MR 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.
In the paper
“In a different study (Kurup et al. 2020), combination of rotation, translation, mirroring, and zooming has been used to achieve automated detection of malaria retinopathy.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of eye fundus images
What the sources say
- Not checked yetcites Kurup et al. 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.
None of the sources cited for this claim have been checked yet.
In the paper
“In the literature, augmentation of brain MR images have been performed with various methods such as rotation, noise addition, shearing, and translation to increase number of images and improve performances of several tasks (Table 1).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of brain MR images
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
“Another combined augmentation including shifting and blurring has been used by Tufail et al. (2021) before binary classification (as healthy and nonhealthy) and multi-class (5 classes) classification of images to differentiate different stages of DR.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of eye fundus images
What the sources say
- Not checked yetcites Tufail et al. (2021)
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
“For instance, image refinement has been provided by noise reduction and resizing steps, and then segmentation and augmentation steps have been applied (Zeiser et al. 2020).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of breast mammography images
What the sources say
- Not checked yetcites Zeiser et al. 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.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“For comparisons of the effectiveness of the augmentation methods in deep learning-based classifications, the same network architecture and expanded data sets with the augmented images from each augmentation methods have been used.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Introduction
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.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“The papers published between 2020 and 2022 and mainly collected by Springer, IEEE Xplore, and ELSEVIER, have been reviewed, and a systematic and comprehensive review and analyses of the augmentation methods have been carried out.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Introduction
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
“Also, their effectiveness depends on the characteristics (e.g., color, intensity, etc.) of the images, and whether all significant visual features according to medical conditions exist in the training sets.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Discussions
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.
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
“Also, it is not possible to determine augmentation methods’ contributions to those performances according to the presented results in those publications.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of brain MR images
plain- Asserted: fails — The authors may not 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 authors may not put it forward as true. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. The quote only partly states this claim.
In the paper
“In this augmentation, different NV kinds (trees, wheels, brooms) have been generated depending on the expected shape and location of NVs to synthesize new images.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of eye fundus images
What the sources say
- Not checked yetcites Ara´ujo et al. 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.
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 those applications, different network models, data sets and number of images have been used, and also their performances have been evaluated with different metrics. High performances have been obtained from those applications according to the reported results.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of lung CT images
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.
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
“their effectiveness depends on the characteristics (e.g., color, intensity, etc.) of the images, and whether all significant visual features according to medical conditions exist in the training sets.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Conclusion
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
“Manual evaluations are subjective as well as time-consuming (Xu et al. 2018) since they are based on the domain knowledge of the physician.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Discussions
What the sources say
- Not checked yetcites Xu et al. 2018
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 GAN based augmentation might lead to biased results because of matching the generated images to the distribution of the target domain”
§ Medical image data augmentation: techniques, comparisons and interpretations > Discussions
What the sources say
- Not checked yetcites Ju et al. (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
“GANs have vanishing gradient and mode collapsing problems.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Conclusion
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.
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
“Additionally, they require strong computer resources, a long processing time, and hamper the quality of the generated images in case of running on computers having low computational power with constrained resources.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Discussions
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
“Since the collapse of the mode restricts the capacity of GAN architecture to be varied, this interconnection is detrimental for augmentations of medical images (Zhang 2021).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Discussions
What the sources say
- Not checked yetcites Zhang 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.
It may state two separate claims as one.
None of the sources cited for this claim have been checked yet.
In the paper
“A synthetic image’s quality is characterized by a low level of distortions, fuzziness, noise (Thung and Raveendran 2009), and its feature distribution matching with the class label (Zhou et al. 2020; Costa et al. 2017; Yu et al. 2019).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Discussions
What the sources say
- Not checked yetcites Thung and Raveendran 2009
- Not checked yetcites Zhou et al. 2020
- Not checked yetcites Costa et al. 2017
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.
Probably not a substantial claim. 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
“obtaining satisfactory training results is not easy if the training procedure does not assure the symmetry and alignment of both generator and discriminator networks.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Conclusion
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.
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
“Artifacts can be added while synthesizing images, and therefore the quality of the produced images can be very low.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Discussions
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: unclear — It is unclear why the paper says 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
“GAN based augmentation methods can increase diversity.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Conclusion
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.
In the paper
“A significant problem is vanishing gradient and mode collapsing (Yi et al. 2019).”
§ Medical image data augmentation: techniques, comparisons and interpretations > Discussions
What the sources say
- Not checked yetcites Yi et al. 2019
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
“Also, the trained model can be unpredictable since GANs are complex structures and managing the coordination of the generator and discriminator is difficult.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Discussions
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
“GANs are complex structures and managing the coordination of the generator and discriminator is difficult.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Conclusion
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. It may state two separate claims as one.
The paper relies on this without support.
Medical image datasets are usually limited in size.
“However, the number of medical images is usually limited.”
The paper relies on this without support.
Scarcity and imbalance in medical image data cause overfitting, bias, and inaccurate results.
“Therefore, imbalanced data causing overfitting, biased and inaccurate results is a significant challenge that should be handled while developing deep network based methods.”
The paper relies on this without support.
Deep networks need training sets with many images covering all variations to achieve high generalization and robustness.
“Because training data sets should be constructed with a large number of images including all variations (i.e., heterogeneous images) to provide the high generalization ability and robustness of those architectures.”
The paper relies on this without support.
No earlier comprehensive review covers a wider range of medical imaging modalities and applications than this paper.
“According to our knowledge, there is no (i) comprehensive review that covers a greater range of medical imaging modalities and applications than articles in the literature”
The paper relies on this without support.
No earlier work quantitatively evaluates augmentation methods across different medical images to identify the most suitable method.
“and (ii) any work on quantitative evaluations of the augmentation methods applied with different medical images to determine the most appropriate augmentation method.”
The paper relies on this without support.
Eye fundus features such as neovascularization, hemorrhages/microaneurysms, and exudates are symptoms of eye diseases.
“Eye fundus images show various important features, which are symptoms of eye diseases, such as abnormal blood veins (neovascularization), red/reddish spots (hemorrhages, microaneurysms), and bright lesions (soft or hard exudates).”
The paper relies on this without support.
Image labels may not always be preserved if the degree of rotation increases.
“Image labels may not always be preserved if the degree of rotation increases.”
The paper relies on this without support.
The main challenging issues of GAN-based image augmentation are generating high-quality images and maintaining training stability.
“The main challenging issues of GAN based image augmentations are generating images with high quality (i.e., high-resolution, clear) and maintaining training stability.”
The paper relies on this without support.
Failures in detection or identification of breast lesions lead to unnecessary biopsies or inaccurate diagnoses.
“Because failures in the detection or identification of the lesions lead to unnecessary biopsies or inaccurate diagnoses.”
The paper relies on this without support.
Rotation transformation is generally safe on medical images.
“Although rotation transformation is not safe on images showing 6 and 9 in digit recognition applications, it is generally safe on medical images.”
The paper relies on this without support.
Quality and diversity are critical criteria for evaluating synthetic images from GANs.
“Quality and diversity are two critical criteria in the evaluation of synthetically generated images from GANs.”
The paper relies on this without support.
The validity of similarity measurements for medical image datasets is still under investigation and there is no accepted consensus.
“The validity of these measurements for medical image data sets is still under investigation and there is no accepted consensus.”
The paper relies on this without support.
There is no commonly accepted metric for evaluating the diversity and quality of synthesized images.
“Also, there is no commonly accepted metric for evaluating the diversity and quality of synthesized images.”
The paper relies on this without support.
The images used in the literature have been acquired from different types of fundus cameras.
“the images used in the literature have been acquired from different types of fundus cameras”
The paper relies on this without support.
Fundus images can be affected by pathological alterations.
“Also, images can be affected by pathological alterations.”
The paper relies on this without support.
The contributions of applied augmentation techniques to reported performances in the current literature are not clear.
“Also, the contributions of the applied augmentation techniques to the presented performances are not clear.”
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?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Laboratory performance does not always translate directly to clinical practice [142,143,144,145,146].”
From Artificial Intelligence in Clinical Oncology: From Productivity Enhancement to Creative Discovery (2025) · cites this paper as [143]
Experimental — accuracy not yet measured
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
§ Medical image data augmentation: techniques, comparisons and interpretations > Results
“| 7th augmentation method | 85.24 |”
§ Medical image data augmentation: techniques, comparisons and interpretations > Results
“| 86.10 | 96.64 | 84.35 | 81.41 |”
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
§ Medical image data augmentation: techniques, comparisons and interpretations > Results
“| 2nd augmentation method | 60.95 |”
§ Medical image data augmentation: techniques, comparisons and interpretations > Results
“| 63.13 | 73.57 | 60.28 | 51.97 |”
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
§ Medical image data augmentation: techniques, comparisons and interpretations > Results
“| 10th augmentation method | 88.24 |”
§ Medical image data augmentation: techniques, comparisons and interpretations > Results
“| 90.48 | 97.13 | 88.52 | 82.79 |”
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
§ Medical image data augmentation: techniques, comparisons and interpretations > Results
“| 6th augmentation method | 62.38 |”
§ Medical image data augmentation: techniques, comparisons and interpretations > Results
“| 65.16 | 70.06 | 62.51 | 56.08 |”
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
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of lung CT images
“Accuracy, specificity, sensitivity, F1-score are 99.49%,”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of lung CT images
“99.25%, 99.25%, and 99.12%, respectively”
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
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of lung CT images
“All methods applied for augmentation of breast mammography images in the publications reviewed in this study and the details in those publications are presented in Table 2.”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of lung CT images
“Table: Table 2 Augmentation methods applied with lung CT images and details given in publications”
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
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of eye fundus images
“Sensitivity, specificity, accuracy values are 90%, 100%, and 81% respectively”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of eye fundus images
“Malarial retinopathy detection”
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
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of breast mammography images
“accuracy, and sensitivity are 88.0%,86.1%, 87.0%, and 85.6%, respectively”
§ Medical image data augmentation: techniques, comparisons and interpretations > Related works > Augmentation of breast mammography images
“Image classification (as normal, and cancer)”
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2017) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2577031 · Cited by 73,472
ImageNet Large Scale Visual Recognition Challenge
Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., Huang, Z., Karpathy, A., Khosla, A., Bernstein, M., et al. (2015) · International Journal of Computer Vision · doi:10.1007/s11263-015-0816-y · Cited by 39,380
Fully convolutional networks for semantic segmentation
Long, J., Shelhamer, E., & Darrell, T. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298965 · Cited by 38,834
Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation
Girshick, R., Donahue, J., Darrell, T., & Malik, J. (2014) · 2014 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2014.81 · Cited by 32,684
He, K., Gkioxari, G., Dollar, P., & Girshick, R. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.322 · Cited by 29,137
Girshick, R. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.169 · Cited by 28,105
A survey on Image Data Augmentation for Deep Learning
Shorten, C., & Khoshgoftaar, T. M. (2019) · Journal of Big Data · doi:10.1186/s40537-019-0197-0 · Cited by 13,203
Fully convolutional networks for semantic segmentation
Shelhamer, E., Long, J., & Darrell, T. (2017) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2572683 · Cited by 11,898
Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset
Carreira, J., & Zisserman, A. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.502 · Cited by 9,570
Research works printed but not linked: 5
- Asia AO, Zhu CZ, Althubiti SA, Al-Alimi D, Xiao YL, Ouyang PB, Al-Qaness MA (2020) Detection of diabetic retinopathy in retinal fundus images using cnn classification models. Electronics 11:1–20
- Bakas S, Reyes M, Jakab A, Bauer S, Rempfler M, Crimi A, Shinohara RT, Berger C, Ha SM, Rozycki M, Prastawa M (2018) Identifying the best machine learning algorithms for brain tumor segmentation, progression assessment, and overall survival prediction in the brats challenge. arXiv preprint, pp 1–49. arXiv:1811.02629 · arXiv
- Dorizza A (2021) Data augmentation approaches for polyp segmentation. Dissertation, Universita Degli Studi Di Padova
- Simonyan K, Zisserman A (2014) Two-stream convolutional networks for action recognition in videos. Adv Neural Inf Process Syst 27:1–9
- Tyler ME, Hubbard LD, Boydston K, Pugliese AJ (2009) Characteristics of digital fundus camera systems affecting tonal resolution in color retinal images. J Ophthalmic Photogr 31:1–9
Cited by
Sorted by citations
Islam, T., Hafiz, M. S., Jim, J. R., Kabir, M. M., & Mridha, M. F. (2024) · Healthcare Analytics · doi:10.1016/j.health.2024.100340 · Cited by 234
Gan-based data augmentation to improve breast ultrasound and mammography mass classification
Jiménez-Gaona, Y., Carrión-Figueroa, D., Lakshminarayanan, V., & Rodríguez-Álvarez, M. J. (2024) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2024.106255 · Cited by 79
Variational Autoencoders for Data Augmentation in Clinical Studies
Papadopoulos, D., & Karalis, V. D. (2023) · Applied Sciences · doi:10.3390/app13158793 · Cited by 67
Brain tumour segmentation framework with deep nuanced reasoning and Swin‐T
Yang, X., Yu, K., Qi, G., et al. (2024) · IET Image Processing · doi:10.1049/ipr2.13048 · Cited by 54
Advancements in AI-Driven detection and localisation of solar panel defects
Ghahremani, A., Adams, S. D., Norton, M., Khoo, S., & Kouzani, A. Z. (2025) · Advanced Engineering Informatics · doi:10.1016/j.aei.2024.103104 · Cited by 51
Singh, L. K., Khanna, M., Thawkar, S., & Singh, R. (2023) · Multimedia Tools and Applications · doi:10.1007/s11042-023-15175-6 · Cited by 45
“In order to segment and categorize retinal fundus pictures for the diagnosis of glaucoma, this research offers a computer approach [40].”
MethodA Recent Survey of Vision Transformers for Medical Image Segmentation
Khan, A., Rauf, Z., Khan, A. R., Rathore, S., Khan, S. H., Shah, N. S., Farooq, U., Asif, H., Asif, A., Zahoora, U., et al. (2025) · IEEE Access · doi:10.1109/access.2025.3618215 · Cited by 109
“Current trends to mitigate these include data augmentation, synthetic data generation, and advanced training techniques like self-supervised learning to make the models more robust [189], [190], [191], [192], [193].”
Colon cancer diagnosis by means of explainable deep learning
Giammarco, M. D., Martinelli, F., Santone, A., et al. (2024) · Scientific Reports · doi:10.1038/s41598-024-63659-8 · Cited by 43
A multi-stage deep learning approach for comprehensive lung disease classification from x-ray images
Deepak, G. D., & Bhat, S. K. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.127220 · Cited by 46
Detecting severity of Diabetic Retinopathy from fundus images: A transformer network-based review
Karkera, T. M., Adak, C., Chattopadhyay, S., & Saqib, M. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.127991 · Cited by 55
“Deep networks are data-hungry, and several augmentations have been applied to address diverse issues associated with deep networks [43], including model overfitting, and to enhance the robustness of the models.”
BackgroundHuman skin type classification using image processing and deep learning approaches
Saiwaeo, S., Arwatchananukul, S., Mungmai, L., et al. (2023) · Heliyon · doi:10.1016/j.heliyon.2023.e21176 · Cited by 39
Weng, Y., Zhang, Y., Wang, W., & Dening, T. (2024) · Information Fusion · doi:10.1016/j.inffus.2024.102263 · Cited by 41
Naz, H. U., Nijhawan, R., Ahuja, N. J., et al. (2023) · IEEE Access · doi:10.1109/access.2023.3327900 · Cited by 36
“However, the application of deep learning in this context is hindered by the availability of small or insufficiently annotated sample datasets [5].”
BackgroundMethodLou, M., Ying, H., Liu, X., Zhou, H.-Y., Zhang, Y., & Yu, Y. (2025) · Neural Networks · doi:10.1016/j.neunet.2025.107228 · Cited by 44
Deep Learning-Driven Automatic Segmentation of Weeds and Crops in UAV Imagery
Tao, J., Qiao, Q., Song, J., Sun, S., Chen, Y., Wu, Q., Liu, Y., Xue, F., Wu, H., & Zhao, F. (2025) · Sensors · doi:10.3390/s25216576 · Cited by 40
Fang, K., Zhang, Y., Du, S., & He, J. (2023) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2023.107372 · Cited by 35
Yurtsever, M. M. E., Atay, Y., Arslan, B., & Sağıroğlu, Ş. (2024) · BMC Medical Informatics and Decision Making · doi:10.1186/s12911-024-02699-6 · Cited by 37
“Additionally, data augmentation can help to address class imbalance issues by generating synthetic samples for underrepresented classes, thereby improving the overall performance of the models [24].”
BackgroundJeon, S., Kim, S. H., Heo, G., Heo, H. J., Chae, S. Y., Kwon, Y. W., Lee, S. K., Han, D., Kim, H., Kim, Y. H., et al. (2025) · Advanced Science · doi:10.1002/advs.202509658 · Cited by 29
Zhao, Q., Yang, L., & Lyu, N. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.122056 · Cited by 30
HI-MViT: A lightweight model for explainable skin disease classification based on modified MobileViT
Ding, Y., Yi, Z., Li, M., Long, J., Lei, S., Guo, Y., Fan, P., Zuo, C., & Wang, Y. (2023) · Digital Health · doi:10.1177/20552076231207197 · Cited by 31
Li, J., Tao, R., Chen, R., Chen, Y., Zhao, C., & Huang, X. (2024) · Computers & Industrial Engineering · doi:10.1016/j.cie.2024.110209 · Cited by 30
Attallah, O. (2024) · Image and Vision Computing · doi:10.1016/j.imavis.2024.105064 · Cited by 26
Gao, H., Wan, Y., Xu, H., et al. (2025) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2025.3527061 · Cited by 26
“3) Data Augmentation: Data augmentation [30] is a key approach for overcoming overfitting in small datasets.”
Zhu, A.-Q., Wang, Q., Shi, Y.-L., Ren, W.-W., Cao, X., Ren, T.-T., Wang, J., Zhang, Y.-Q., Sun, Y.-K., Chen, X.-W., et al. (2024) · EClinicalMedicine · doi:10.1016/j.eclinm.2023.102391 · Cited by 25
Zhou, C. R., Liu, C., Liao, Z., Pang, Y., & Sun, W. (2024) · Biofabrication · doi:10.1088/1758-5090/ad8966 · Cited by 30
“…of medical images by altering their contrast, brightness, and sharpness Currently, image enhancement techniques are widely applied to different body parts such as the brain, lungs, chest, and eyes, as well as various imaging diagnostic technologies including MRI, CT, and mammogram [ 28, 30].”
LCAMix: Local-and-contour aware grid mixing based data augmentation for medical image segmentation
Sun, D., Dornaika, F., & Charafeddine, J. (2024) · Information Fusion · doi:10.1016/j.inffus.2024.102484 · Cited by 30
Stroke Lesion Segmentation and Deep Learning: A Comprehensive Review
Malik, M., Chong, B., Fernandez, J., Shim, V., Kasabov, N. K., & Wang, A. (2024) · Bioengineering · doi:10.3390/bioengineering11010086 · Cited by 23
“Deep learning models require a lot of labelled data to learn real-world patterns accurately; however, labelled medical imaging data are scarce for various reasons, including but not limited to privacy issues and domain expertise required to label the data [87].”
BackgroundKnoedler, L., Knoedler, S. A., Allam, O., Remy, K., Miragall, M. F., Safi, A., Alfertshofer, M. G., Pomahač, B., & Kauke‐Navarro, M. (2023) · Frontiers in Surgery · doi:10.3389/fsurg.2023.1266399 · Cited by 21
“…augmentation techniques such as fl ipping (i.e., horizontally or vertically mirroring an image or data point), shearing (i.e., distorting the image by tilting it in a speci fi c direction), or color channel shifting (i.e., altering the intensity of individual color channels in the image) (56, 57).”
MethodMUCM-Net: a Mamba powered UCM-Net for skin lesion segmentation
Yuan, C., Zhao, D., & Agaian, S. С. (2024) · Exploration of Medicine · doi:10.37349/emed.2024.00250 · Cited by 21
Kojok, Z. E., Khansa, H. A., Trad, F. S., & Chehab, A. (2024) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2024.109446 · Cited by 21
Yixin, Z., Yang, L., Guofan, J., et al. (2025) · Information Sciences · doi:10.1016/j.ins.2025.122074 · Cited by 20
Ayana, G., Lee, E., & Choe, S. (2023) · American Journal Of Pathology · doi:10.1016/j.ajpath.2023.11.015 · Cited by 21
Deep learning-based automatic segmentation of cerebral infarcts on diffusion MRI
Ryu, W., Schellingerhout, D., Park, J.-H., Chung, J., Jeong, S., Gwak, D., Kim, B. J., Kim, J., Hong, K., Lee, K. B., et al. (2025) · Scientific Reports · doi:10.1038/s41598-025-91032-w · Cited by 19
“However, recent studies have demonstrated that segmentation algorithms built on datasets with a small number of original images, augmented to thousands of diverse variations, can achieve strong performance 36,37 .”
Deep Learning in Digital Breast Tomosynthesis: Current Status, Challenges, and Future Trends
Wang, R., Chen, F., Chen, H., Lin, C. X., Shuai, J., Wu, Y., Ma, L., Hu, X., Wu, M. H., Wang, J., et al. (2025) · MedComm · doi:10.1002/mco2.70247 · Cited by 22
“To this end, data augmentation techniques can be used to improve the learning and generalization capabilities of deep networks [294–297].”
ODFormer: Semantic fundus image segmentation using Transformer for optic nerve head detection
Wang, J., Mao, Y.-A., Ma, X., et al. (2024) · Information Fusion · doi:10.1016/j.inffus.2024.102533 · Cited by 18
“It should be noted here that although there are various augmentation methods, they should be chosen according to image characteristics and applied care-fully with appropriate parameters to increase the reliability and robustness of the methods [45, 46, 47].”
MethodZhang, J., Zhang, F., & Wei, H. (2024) · Pattern Recognition · doi:10.1016/j.patcog.2024.111005 · Cited by 20
Deep Learning Approaches for Chest Radiograph Interpretation: A Systematic Review
Iqbal, H., Khan, A., Nepal, N., Khan, F. N., & Moon, Y.-K. (2024) · Electronics · doi:10.3390/electronics13234688 · Cited by 21
Das, I., Sheakh, M. A., Abdulla, S., Tahosin, M. S., Hassan, M. M., Zaman, S., & Shukla, A. (2025) · IEEE Access · doi:10.1109/access.2025.3607639 · Cited by 20
“Many prior studies have successfully employed similar augmentation ranges, validating their efficacy in medical imaging tasks [51].”
Chamarthi, S., Fogelberg, K., Maron, R. C., Brinker, T. J., & Niebling, J. (2023) · Informatics in Medicine Unlocked · doi:10.1016/j.imu.2023.101430 · Cited by 29
HSMix: Hard and soft mixing data augmentation for medical image segmentation
Sun, D.-Y., Dornaika, F., & Barrena, N. (2024) · Information Fusion · doi:10.1016/j.inffus.2024.102741 · Cited by 20
“Data augmentation [9] can address data scarcity and overfitting problems by generating more training samples [10].”
Ilyas, T., Ahmad, K., Arsa, D. M. S., Jeong, Y. C., & Kim, H. (2024) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2024.108055 · Cited by 18
Novel hybrid transfer neural network for wheat crop growth stages recognition using field images
Naseer, A., Amjad, M., Raza, A., et al. (2025) · Scientific Reports · doi:10.1038/s41598-025-96332-9 · Cited by 17
Haider, S. A., Prabha, S., Gomez-Cabello, C. A., Borna, S., Pressman, S. M., Genovese, A., Trabilsy, M., Galvao, A., Aziz, K. T., Murray, P. M., et al. (2025) · Journal of Clinical Medicine · doi:10.3390/jcm14072136 · Cited by 17
Prezja, F., Annala, L., Kiiskinen, S., & Ojala, T. (2023) · Algorithms · doi:10.3390/a17010008 · Cited by 18
“Often, the choice of augmentation techniques relies heavily on informal recommendations or generic best practices that aren’t always tailored to medical images’ unique challenges and characteristics 51 .”
InfluentialAutomatic Lung Segmentation in Chest X-Ray Images Using SAM With Prompts From YOLO
Khalili, E., Priego, B., León‐Jiménez, A., & Morillo, D. (2024) · IEEE Access · doi:10.1109/access.2024.3454188 · Cited by 20
“Although the performance of the proposed approach was slightly lower than the InvertedNet CNN (97.6% Dice Score and 95.5% IoU) [39], it was still competitive.”
MethodInfluentialComposto, A., Privitera, L., Riva, M., Ardini, B., Manzoni, C., Riva, M., Aquilina, K., Valentini, G., & Giuliani, S. (2025) · Journal of Biomedical Optics · doi:10.1117/1.jbo.30.s2.s23909 · Cited by 16
“107,108 Although research in remote sensing has demonstrated the ability of such algorithms to generate accurate hyperspectral images, 91,109 – 111 their direct applicability to the medical domain and capacity to generate reliably annotated HSI datasets remains to be fully determined.”
GR-m6A: Prediction of N6-methyladenosine sites in mammals with molecular graph and residual network
Qiu, S., Liu, R., & Liang, Y. (2023) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2023.107202 · Cited by 15
Marullo, G., Ulrich, L., Antonaci, F. G., Audisio, A., Aprato, A., Massè, A., & Vezzetti, E. (2024) · Bone Reports · doi:10.1016/j.bonr.2024.101801 · Cited by 20
“…obtained through rotation, vertical and horizontal flipping and thus did not transform in contrast and brightness, and from the images obtained through the other transformations whose new luminosity and contrast was between the 25th and 75th percentile of the actual training images (Goceri, 2023).”
BackgroundPeter, O. O. E., Adeniran, O., John-Otumu, A. M., Khalifa, F., & Rahman, M. M. (2025) · Algorithms · doi:10.3390/a18030155 · Cited by 17
Berne, J. V., Saadi, S. B., Oliveira‐Santos, N., Marinho-Vieira, L. E., & Jacobs, R. (2025) · Journal of Dentistry · doi:10.1016/j.jdent.2025.105688 · Cited by 21
Salem, H., Elbaz, M., Soliman, R. S., & Elkhatib, A. A. (2025) · Scientific Reports · doi:10.1038/s41598-025-11955-2 · Cited by 15
Dataset of chilli and onion plant leaf images for classification and detection
Aishwarya, M. P., & Reddy, A. P. (2024) · Data in Brief · doi:10.1016/j.dib.2024.110524 · Cited by 15
“Image data has been increased using augmentation technique such as rotating, flipping, zooming and shearing to artificially increase the diversity of the dataset, thereby improving the robustness, generalization, and applicability of models and algorithms trained on these datasets [ 8 ].”
MethodHybrid augmentation for multi-channel deep learning in guava leaf disease detection
Güler, O., Etem, T., & Teke, M. (2025) · Ain Shams Engineering Journal · doi:10.1016/j.asej.2025.103716 · Cited by 17
Waqas, M., Hasan, S., Ghori, A. F., et al. (2025) · International Dental Journal · doi:10.1016/j.identj.2025.103878 · Cited by 14
“7 These models require large high quality datasets for effective training, but the scarcity of medical data, caused by factors such as privacy concerns, limited data sharing between organizations, and high storage costs, poses a major barrier to the adoption of modern machine learning tools in…”
Wang, Z., Chen, L., Xie, X., et al. (2025) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2025.3605162 · Cited by 14
“Unlike some SDG techniques [4], [14] that focus on complex model structures, data augmentation techniques [15] are more versatile, as they can be applied to a wider range of scenarios due to their orthogonality to model architectures.”
Deng, K., Xing, L., Wu, H., Ma, H., Ling, Y., & Gao, J. (2025) · Journal of King Saud University - Computer and Information Sciences · doi:10.1007/s44443-025-00302-0 · Cited by 20
“Currently, the mainstream application of grayscale images is medical imaging, particularly in the field of radiology (Kelkar et al. 2023; Goceri 2023), such as X-rays, CT scans, and MRIs, which are widely used for diagnosing various medical conditions and assessing the internal structures of the…”
Lu, J., Tong, X., Wu, H., Liu, Y., Ouyang, H., & Zeng, Q. (2023) · Heliyon · doi:10.1016/j.heliyon.2023.e20186 · Cited by 15
A CRNN‐based method for Chinese ship license plate recognition
Xu, F., Chen, C., Shang, Z., et al. (2023) · IET Image Processing · doi:10.1049/ipr2.12949 · Cited by 14
Bougourzi, F., Dornaika, F., Nakib, A., & Ahmed, A. T. (2024) · Artificial Intelligence Review · doi:10.1007/s10462-024-10717-2 · Cited by 15
“In the last decade, Deep Leaning (DL) approaches have become the dominant trend in many machine learning and computer vision tasks (Van der Velden et al. 2022; Bougourzi et al. Apr. 2022; Goceri 2023).”
A Novel Data Augmentation Method for Radiomics Analysis Using Image Perturbations
Iacono, F. L., Maragna, R., Pontone, G., & Corino, V. (2024) · Journal of Imaging Informatics in Medicine · doi:10.1007/s10278-024-01013-0 · Cited by 15
“…by class unbalance or lack of data: the reasons of these issues lie in the low prevalence of some diseases, poor medical equipment or resources, difficulties in obtaining images that meet specific, or issues related to the complex and time-consuming tasks of image labeling and segmentation [15].”
BackgroundNikpasand, M., Middendorf, J. M., Ella, V. A., Jones, K. E., Ladd, B. M., Takahashi, T., Barocas, V. H., & Ellingson, A. M. (2024) · JOR Spine · doi:10.1002/jsp2.1353 · Cited by 12
Huang, J., Wu, Y., Wang, F., Fang, Y., Yang, N., Alkan, C., Abraham, D. P., Liao, C., Xu, L., Gao, Z., et al. (2024) · IEEE Reviews in Biomedical Engineering · doi:10.1109/rbme.2024.3485022 · Cited by 29
“(a) Model development: One effective strategy is the use of (a.i) data augmentation [199] methods for model training.”
BackgroundYue, G., Ma, X., Li, W., An, Z., & Yang, C. (2024) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2024.107140 · Cited by 15
BSDA: Bayesian Random Semantic Data Augmentation for Medical Image Classification
Zhu, Y., Cai, X., Wang, X., et al. (2024) · Sensors · doi:10.3390/s24237511 · Cited by 16
“Deep learning methods can assist clinicians in rapid examination and accurate diagnosis [1].”
Shen, Y., Chen, L., Liu, J., Chen, H., Wang, C., Ding, H., & Zhang, Q. (2025) · Computerized Medical Imaging and Graphics · doi:10.1016/j.compmedimag.2024.102490 · Cited by 14
Kababji, S. E., Mitsakakis, N., Jonker, E., Beltran-Bless, A.-A., Pond, G. R., Vandermeer, L., Radhakrishnan, D., Mosquera, L., Paterson, A. H. G., Shepherd, L. E., et al. (2025) · Journal of Medical Internet Research · doi:10.2196/66821 · Cited by 15
“Augmentation through simulation can be a potential solution when there are insufficient data and is a common practice for imaging data [23-25] and time series data [26,27].”
Hybrid CNN-Based Transfer Learning Enhances Brain Tumor Classification on MRI Images
Prayogo, R. D., Hamid, N., & Nambo, H. (2025) · IEEE Access · doi:10.1109/access.2025.3584376 · Cited by 16
“A range of medical imaging technologies, including positron emission tomography (PET), computed tomography (CT), and magnetic resonance imaging (MRI), have been utilized to identify brain abnormalities [4].”
Gidado, M. J., Gunny, A. A. N., Gopinath, S. C. B., Devi, M., & Ibrahim, A. M. (2026) · Journal of Stored Products Research · doi:10.1016/j.jspr.2026.102959 · Cited by 18
Deep learning-based efficient diagnosis of periapical diseases with dental X-rays
Wang, K., Zhang, S., Wei, Z., Fang, X., Liu, F., Han, M., & Du, M. (2024) · Image and Vision Computing · doi:10.1016/j.imavis.2024.105061 · Cited by 13
RADAP: A Robust and Adaptive Defense Against Diverse Adversarial Patches on face recognition
Liu, X., Shen, F., Zhao, J. Y., et al. (2024) · Pattern Recognition · doi:10.1016/j.patcog.2024.110915 · Cited by 13
Dai, G., Wang, C., Dai, D., et al. (2025) · Information Fusion · doi:10.1016/j.inffus.2025.103446 · Cited by 13
Qian, X., Shao, H., Li, Y., Lu, W., & Zhang, Y. (2025) · Medical Physics · doi:10.1002/mp.17757 · Cited by 12
Nimmy, S. F., Hussain, O. K., Chakrabortty, R. K., & Saha, S. (2025) · Knowledge-Based Systems · doi:10.1016/j.knosys.2025.113333 · Cited by 15
Qiang, Y.-R., Zhou, Q.-Y., Li, J.-N., Xie, M.-Y., Cui, X., & Zhang, S. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.127720 · Cited by 15
A Lightweight CNN for Multiclass Retinal Disease Screening with Explainable AI
Arnob, A. K. B., Chayon, M. H. R., Farid, F. A., et al. (2025) · Journal of Imaging · doi:10.3390/jimaging11080275 · Cited by 12
“Goceri et al. [28] emphasized that the performance of the augmentation techniques is dependent on the disease and imaging method and, thus, needs to be selected accordingly in order to provide excellent performance.”
Hybrid CNN-transformer network for interactive learning of challenging musculoskeletal images
Bi, L., Buehner, U., Fu, X., Williamson, T., Choong, P., & Kim, J. (2023) · Computer Methods and Programs in Biomedicine · doi:10.1016/j.cmpb.2023.107875 · Cited by 13
Aktas, B., Ates, D. D., Düzyel, O., & Gümüş, A. (2024) · International Journal of Machine Learning and Cybernetics · doi:10.1007/s13042-024-02485-w · Cited by 13
“Data scarcity often leads to a negative result in model performance, at which point classical data augmentation methods step in such as rotation, flipping, cropping, brightness, and contrast adjustment [40].”
Deep Learning for Detecting Dental Plaque and Gingivitis From Oral Photographs: A Systematic Review
Moharrami, M., Vahab, E., Bagherianlemraski, M., Hemmati, G., Singhal, S., Quiñonez, C., Schwendicke, F., & Glogauer, M. (2025) · Community Dentistry And Oral Epidemiology · doi:10.1111/cdoe.70001 · Cited by 20
“Furthermore, if a model is underfitting, demonstrated by consistently high bias on the training set, then increasing the sample size is unlikely to improve performance [90, 91].”
Cho, J., Woo, S., Hwang, S. H., Kim, S., Lee, H., Hwang, J.-Y., Kim, J.-W., Kim, M. S., Smith, L., Lee, S., et al. (2025) · Diabetes Care · doi:10.2337/dc25-0355 · Cited by 16
Understanding the robustness of vision-language models to medical image artefacts
Cheng, Z., Ong, A. Y., Wagner, S. K., Merle, D. A., Ju, L., Zhang, H., Chen, R., Pang, L., Li, B., He, T., et al. (2025) · npj Digital Medicine · doi:10.1038/s41746-025-02108-w · Cited by 18
“In certain cases, MedGemma ’ s fi nal output diverges from the reasoning it presents, with an example of this discrepancy shown in Supplementary Data 13.”
InfluentialChang, J., & Hatfield, B. S. (2024) · Advances in cancer research · doi:10.1016/bs.acr.2024.05.006 · Cited by 13
EffSVMNet: An efficient hybrid neural network for improved skin disease classification
Sharma, Y., Tiwari, N. K., & Upaddhyay, V. K. (2024) · Smart Health · doi:10.1016/j.smhl.2024.100520 · Cited by 12
Wang, S., Zhao, Z.-A., Chen, Y., et al. (2025) · Technologies · doi:10.3390/technologies13010028 · Cited by 11
Dual-Stream CoAtNet models for accurate breast ultrasound image segmentation
Zaidkilani, N., García, M. Á., & Puig, D. (2024) · Neural Computing and Applications · doi:10.1007/s00521-024-09963-w · Cited by 11
“Although different augmentation techniques have been applied to overcome different issues caused by deep networks [39–41], efficient augmentation methods must be properly chosen according to the characteristics of the images.”
SkinHealthMate app: An AI-powered digital platform for skin disease diagnosis
Aboulmira, A., Lachgar, M., Hrimech, H., Camara, A., Elbahja, C., Elmansouri, A., & Hassini, Y. (2024) · Systems and Soft Computing · doi:10.1016/j.sasc.2024.200166 · Cited by 12
Moon, J., Jadhav, P., & Choi, S. T. (2025) · Journal of Rheumatic Diseases · doi:10.4078/jrd.2024.0128 · Cited by 11
“While DL models traditionally required large datasets, advancements such as transfer learning and data augmentation have made it possible to train models with smaller amounts of data [139,140].”
Segmentation of lungs from chest X-ray images based on Deep Atrous Attention UNet (DAA-UNet) model
Yadav, V. K., & Singhai, J. (2025) · Medical & Biological Engineering & Computing · doi:10.1007/s11517-025-03344-8 · Cited by 12
“To increase the number of images in the resultant datasets, image augmentation techniques such as random rotation by 45°, 90°, and flipping in the horizontal and vertical directions were applied to each original image, as suggested in the literature [34].”
A hybrid parallel convolutional spiking neural network for enhanced skin cancer detection
Kumar, K. A., & Vanmathi, C. (2025) · Scientific Reports · doi:10.1038/s41598-025-85627-6 · Cited by 10
Liu, X., Ono, K., & Bise, R. (2024) · Image and Vision Computing · doi:10.1016/j.imavis.2024.105056 · Cited by 14
Alasmari, H., Amoudi, G., & Alghamdi, H. S. (2025) · Diagnostics · doi:10.3390/diagnostics15182301 · Cited by 12
“The techniques we used are as follows [35]:”
Jia, J., Cao, X., & Wei, Z. (2023) · Current Genomics · doi:10.2174/0113892029270191231013111911 · Cited by 9
“Second, different data enhancement techniques [69-71] used in deep network models can be referenced and tried to be used to study ac4C site prediction.”
Li, P., Wang, H., Wang, Z., Wang, K., & Wang, C. (2024) · IEEE Access · doi:10.1109/access.2024.3387541 · Cited by 9
“These algorithms must be carefully matched to the specific characteristics of the images and the classification task at hand, as highlighted in references [29] and [30].”
BackgroundLiu, C., Xiu, C., Zou, Y., et al. (2024) · Spectrochimica Acta Part A Molecular and Biomolecular Spectroscopy · doi:10.1016/j.saa.2024.125353 · Cited by 9
He, K., Zhu, J., Li, L., Gou, F., & Wu, J. (2024) · Heliyon · doi:10.1016/j.heliyon.2024.e40350 · Cited by 10
Shoushtari, F. K., Dehkordi, A. N. V., & Sina, S. (2024) · Annals of Biomedical Engineering · doi:10.1007/s10439-024-03461-9 · Cited by 9
Gomaa, M. M., elabdeen, A. G. Z., Elnashar, A., et al. (2024) · International Journal of Information Technology · doi:10.1007/s41870-024-01830-0 · Cited by 9
“Data augmentation (DA) is a typical solution to this problem, and numerous augmentation techniques have been applied to various types of photos [4].”
Zhao, J., Jiang, T., Lin, Y., Chan, L.-C., Ping Keung, C. H. A. N., Wen, C., & Chen, H. (2022) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2024.3372576 · Cited by 8
“It is known that deep networks are data-hungry and several augmentation methods [54], [55], [56] have been applied to improve the performance and increase the reliability and robustness of the neural networks.”
MethodDiagnosing Cataracts in the Digital Age: A Survey on AI, Metaverse, and Digital Twin Applications
Jones, A., Vijayan, T. B., & John, S. (2024) · Seminars in Ophthalmology · doi:10.1080/08820538.2024.2403436 · Cited by 8
“12”
Mennemann, K., Ebert, N., Reichardt, L., et al. (2025) · Applied Sciences · doi:10.3390/app15031260 · Cited by 8
Chong, L., Cao, T., Tang, M., Pu, W., & Lei, P. (2025) · European Journal of Radiology · doi:10.1016/j.ejrad.2025.112060 · Cited by 9
Metadata
- Authors
- Goceri, E.
- Venue
- Artificial Intelligence Review
- Published
- 2023-03-20
- Volume
- 56
- Issue
- 11
- Pages
- 12561-12605
- ISSN
- 0269-2821 (ISSN-L)
- 1573-7462
- PMC
- PMC10027281
- OpenAlex
- W4327909703
- Semantic Scholar
- da92fa7949632062a3eb57bdfe6244227cba28be
- Cited by
- 459
- References
- 116
- Citations by year
- 2026: 116 · 2025: 200 · 2024: 115 · 2023: 28
- FWCI
- 18.28
- Type
- journal article
- Language
- en
- Open access
- bronze
- doi.org
- Licence
- Locations
- Artificial Intelligence Review (publishedVersion)
- PubMed Central (publishedVersion)
- PubMed (publishedVersion)
- Topics
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Medical Image Segmentation Techniques
- Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › COVID-19 diagnosis using AI
- Keywords
- medical image classification, data scarcity, overfitting, data bias, neuroimaging, lung imaging, breast imaging, ocular imaging, magnetic resonance imaging, CT, mammography, fundus examination, imaging modalities
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).
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
