Transfer Learning in Magnetic Resonance Brain Imaging: A Systematic Review
Open accessThe DOI matches its Crossref record; 195 of the 195 references it cites were found in a registry or an index.
Abstract
(1) Background: Transfer learning refers to machine learning techniques that focus on acquiring knowledge from related tasks to improve generalization in the tasks of interest. In magnetic resonance imaging (MRI), transfer learning is important for developing strategies that address the variation in MR images from different imaging protocols or scanners. Additionally, transfer learning is beneficial for reutilizing machine learning models that were trained to solve different (but related) tasks to the task of interest. The aim of this review is to identify research directions, gaps in knowledge, applications, and widely used strategies among the transfer learning approaches applied in MR brain imaging; (2) Methods: We performed a systematic literature search for articles that applied transfer learning to MR brain imaging tasks. We screened 433 studies for their relevance, and we categorized and extracted relevant information, including task type, application, availability of labels, and machine learning methods. Furthermore, we closely examined brain MRI-specific transfer learning approaches and other methods that tackled issues relevant to medical imaging, including privacy, unseen target domains, and unlabeled data; (3) Results: We found 129 articles that applied transfer learning to MR brain imaging tasks. The most frequent applications were dementia-related classification tasks and brain tumor segmentation. The majority of articles utilized transfer learning techniques based on convolutional neural networks (CNNs). Only a few approaches utilized clearly brain MRI-specific methodology, and considered privacy issues, unseen target domains, or unlabeled data. We proposed a new categorization to group specific, widely-used approaches such as pretraining and fine-tuning CNNs; (4) Discussion: There is increasing interest in transfer learning for brain MRI. Well-known public datasets have clearly contributed to the popularity of Alzheimer’s diagnostics/prognostics and tumor segmentation as applications. Likewise, the availability of pretrained CNNs has promoted their utilization. Finally, the majority of the surveyed studies did not examine in detail the interpretation of their strategies after applying transfer learning, and did not compare their approach with other transfer learning approaches.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Figures



Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
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
“though this approach demands substantial computational resources and curated pre-training data to avoid negative transfer (Valverde et al., 2021).”
From Multimodal neuroimaging and AI integration in cognitive disorders: advances, challenges, and future directions for precision medicine (2026) · cites this paper as Valverde et al., 2021
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
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 · cited 1×
“The most popular were VGG [156] (23), ResNet [157] (15), and Inception [158–160] (11).”
Very Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155 · cited 2×
“The demonstrations of outperformance of CNNs over other methods [168], and the availability of trained CNNs in ImageNet [124], such as AlexNet [169], VGG [156], and Inception [158], probably explains CNNs’ popularity.”
Preferred Reporting Items for Systematic Reviews and Meta-Analyses: the PRISMA statement
Moher, D., Liberati, A., Tetzlaff, J., Altman, D. G., & for the PRISMA Group (2009) · BMJ · doi:10.1136/bmj.b2535 · Cited by 83,823
ImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333 · cited 2×
“, ImageNet [124]) whereas 26 approaches pretrained on medical images.”
Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement
Moher, D., Liberati, A., Tetzlaff, J., Altman, D. G., & The PRISMA Group (2009) · PLoS Medicine · doi:10.1371/journal.pmed.1000097 · Cited by 64,042
Going deeper with convolutions
Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298594 · Cited by 48,562
Preferred reporting items for systematic reviews and meta-analyses: the PRISMA Statement.
(2009) · PubMed Central · Cited by 44,859
(2008) · Journal of Machine Learning Research · Cited by 33,051
Rethinking the Inception Architecture for Computer Vision
Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., & Wojna, Z. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.308 · Cited by 32,138
Research works printed but not linked: 1
- Jones, D.K. (2010). Diffusion Mri, Oxford University Press.
Cited by
Domain adaptation for medical image analysis: A survey
Guan, H., & Liu, M. (2022) · IEEE Transactions on Biomedical Engineering · doi:10.1109/tbme.2021.3117407 · Cited by 1,057
“Unlikeprevious review/surveypapersthatusuallyfocusonasinglemodalityor organ/object( e.g. , brain MRI [23]), this survey reviews studies that cover multiple imaging modalities and organs/objects ( e.g. , brain, lung and heart).”
Transfer learning for medical image classification: a literature review
Kim, H. E., Cosa-Linan, A., Santhanam, N., Jannesari, M., Maros, M. E., & Ganslandt, T. (2022) · BMC Medical Imaging · doi:10.1186/s12880-022-00793-7 · Cited by 965
“On the other hand, others surveyed articles focusing on TL with a specific case study such as microorganism counting [18], cervical cytopathology [19], neuroimaging biomarkers of Alzheimer’s disease [20] and magnetic resonance brain imaging in general [21].”
BackgroundMulti-Modal Brain Tumor Detection Using Deep Neural Network and Multiclass SVM
Maqsood, S., Damaševičius, R., & Maskeliūnas, R. (2022) · Medicina · doi:10.3390/medicina58081090 · Cited by 306
“The disadvantages compared to traditional ML methods are also that it requires a large training dataset, high complexity of time, low accuracy for applications where small datasets are available and expensive GPUs which eventually elevate the user’s cost, although these disadvantages can be alleviated by using transfe…”
BackgroundDeep neural networks learn general and clinically relevant representations of the ageing brain
Leonardsen, E. H., Peng, H., Kaufmann, T., Agartz, I., Andreassen, O. A., Celius, E. G., Espeseth, T., Harbo, H. F., Høgestøl, E. A., de Lange, A. G., et al. (2022) · NeuroImage · doi:10.1016/j.neuroimage.2022.119210 · Cited by 174
Atasever, S., Azgınoğlu, N., Terzi, D. S., & Terzi, R. (2022) · Clinical Imaging · doi:10.1016/j.clinimag.2022.11.003 · Cited by 149
Survey on deep learning based computer vision for sonar imagery
Steiniger, Y., Kraus, D., & Meisen, T. (2022) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2022.105157 · Cited by 143
An artificial intelligence framework and its bias for brain tumor segmentation: A narrative review
Das, S., Nayak, G. K., Saba, L., Kalra, M. K., Suri, J. S., & Saxena, S. (2022) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2022.105273 · Cited by 139
Explainable, Domain-Adaptive, and Federated Artificial Intelligence in Medicine
Chaddad, A., Lu, Q., Li, J., Katib, Y., Kateb, R., Tanougast, C., Bouridane, A., & Abdulkadir, A. (2023) · IEEE/CAA Journal of Automatica Sinica · doi:10.1109/jas.2023.123123 · Cited by 139
“With SHAP [120], it illustrates how the design parameters affect performance using DNN features.”
MethodInfluentialKeleş, E., & Bağcı, U. (2023) · npj Digital Medicine · doi:10.1038/s41746-023-00941-5 · Cited by 111
“The initial layers of ImageNet-pretrained networks, which detect low-level image characteristics, including corners and borders, may not be ef fi cient for medical images 25,26 .”
BackgroundMethodIllakiya, T., & Karthik, R. (2023) · Neuroinformatics · doi:10.1007/s12021-023-09625-7 · Cited by 85
“It could be noted that many research works were presented based on TL techniques in brain MRI research (Valverde et al., 2021).”
BackgroundAlshamrani, H. A., Rashid, M. U., Alshamrani, S. S., & Alshehri, A. H. D. (2023) · Healthcare · doi:10.3390/healthcare11091206 · Cited by 70
“This technique will only work if the feature of the dataset is general and suitable to both the base and target models, instead of being specific to the base model [29,30].”
MethodA Novel AI-Based System for Detection and Severity Prediction of Dementia Using MRI
Jain, V., Nankar, O., Jerrish, D. J., Gite, S., Patil, S., & Kotecha, K. (2021) · IEEE Access · doi:10.1109/access.2021.3127394 · Cited by 61
“[15] performed a systematic literature review for articles that applied transfer learning to MRI brain scans, categorized them, and extracted relevant information.”
BackgroundInfluentialMedical image identification methods: A review
Li, J., Jiang, P., An, Q., Wang, G., & Kong, H. (2023) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2023.107777 · Cited by 70
La Rosa, F., Wynen, M., Al-Louzi, O., Beck, E. S., Huelnhagen, T., Maggi, P., Thiran, J.-P., Kober, T., Shinohara, R. T., Sati, P., et al. (2022) · NeuroImage: Clinical · doi:10.1016/j.nicl.2022.103205 · Cited by 58
“extraction across different MRI device manufacturers or different pulse sequence acquisition parameters (Valverde et al., 2021).”
BackgroundGrade Classification of Tumors from Brain Magnetic Resonance Images Using a Deep Learning Technique
Srinivasan, S., Bai, P. S. M., Mathivanan, S. K., Muthukumaran, V., Babu, J. C., & Vilčeková, L. (2023) · Diagnostics · doi:10.3390/diagnostics13061153 · Cited by 57
“In addition, machine learning models that were trained to solve various tasks related to the task of interest can be used again thanks to transfer learning [7].”
Ghobadi, F., Yaseen, Z. M., & Kang, D. (2024) · Journal of Hydrology · doi:10.1016/j.jhydrol.2024.130772 · Cited by 50
Belt Tear Detection for Coal Mining Conveyors
Guo, X., Liu, X., Zhou, H. F., Stanisławski, R., Królczyk, G., & Li, Z. (2022) · Micromachines · doi:10.3390/mi13030449 · Cited by 48
“Based on the above advantages, TL methods have been widely employed in medical image analysis [79,80], engineering [81], text processing [82], natural language processing [83] etc.”
MethodJekel, L., Brim, W. R., von Reppert, M., Staib, L. H., Petersen, G. C., Merkaj, S., Subramanian, H. E., Zeevi, T., Payabvash, S., Bousabarah, K., et al. (2022) · Cancers · doi:10.3390/cancers14061369 · Cited by 48
“If data is scarce, measures, such as transfer learning [52] or data augmentation should be explored to increase model robustness.”
BackgroundHybrid Classical–Quantum Transfer Learning for Cardiomegaly Detection in Chest X-rays
Decoodt, P., Liang, T. J., Bopardikar, S., Santhanam, H., Eyembe, A., García-Zapirain, B., & Sierra-Sosa, D. (2023) · Journal of Imaging · doi:10.3390/jimaging9070128 · Cited by 48
“The models pre-trained on colored images work for grayscale medical images, such as those of CT, mammography, MRI or CXR [19,20], and they have been developed for several conditions, including Alzheimer’s disease, brain tumors, breast tumors and COVID-19.”
MethodSource Free Semi-Supervised Transfer Learning for Diagnosis of Mental Disorders on fMRI Scans
Hu, Y. F., Huang, Z.-A., Liu, R., Xue, X., Sun, X., Song, L., & Tan, K. C. (2023) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3298332 · Cited by 53
“…two separated institutions have their own private datasets for model training and their learning tasks are related to each other, the well-trained model with sufficient labeled samples (a.k.a. source domain) can help to train the other model lacking in labeled samples (a.k.a, the target domain) [9].”
BackgroundRafanan, J., Ghani, N., Kazemeini, S., Nadeem-Tariq, A., Shih, R. M., & Vida, T. A. (2025) · International Journal of Molecular Sciences · doi:10.3390/ijms26030917 · Cited by 40
“With the growing use of AI in neuro-oncology, TL utilizes pre-trained models to reduce extensive training while still providing high-quality decision-making support [201].”
Zhao, L., & Qin, P. (2022) · IEEE Transactions on Industrial Electronics · doi:10.1109/tie.2022.3146505 · Cited by 35
“Thetransfer learning methods can be divided into three groups [20].”
BackgroundAamir, M., Rahman, Z.-U., Choudhry, N., Bhutto, J. A., Abro, W. A., & Zhu, Z. (2025) · IEEE Access · doi:10.1109/access.2025.3625607 · Cited by 36
“This approach involves taking a model that has been pre-trained on a large, general-purpose dataset (e.g., ImageNet) or a large-scale medical dataset (e.g., RadIm-ageNet) and fine-tuning it on the specific task of brain tumor classification [68], [69], [70].”
Haller, S., Van Cauter, S., Federau, C., Hedderich, D. M., & Edjlali, M. (2022) · Neuroradiology · doi:10.1007/s00234-021-02890-w · Cited by 33
“Artificial intelligence (AI) and big data analyses are considered the driving forces of the fourth industrial revolution and many hopes have been raised regarding the application of AI in healthcare, particularly medical imaging [1].”
BackgroundIllakiya, T., & Karthik, R. (2022) · Traitement du signal · doi:10.18280/ts.390608 · Cited by 33
“The performance of AlexNet in 2012 ImageNet LSVRC-2012 Competition showed it to be one of the most popular deep CNN architectures [35, 36].”
BackgroundQin, P., & Zhao, L. (2023) · IEEE Transactions on Industrial Electronics · doi:10.1109/tie.2023.3250768 · Cited by 33
“is to find the similarity or lessen the difference in distribution between data domains [29].”
BackgroundDomainATM: Domain adaptation toolbox for medical data analysis
Guan, H., & Liu, M. (2023) · NeuroImage · doi:10.1016/j.neuroimage.2023.119863 · Cited by 36
“…of the “domain shift” phenomenon [7] (top row) and the fundamental of domain adaptation (distribution of source and target samples before and after adaptation). field of medical data analysis, domain adaptation has gained considerable attention and increasing interest recently [17], [18].”
Classification of Osteo-Arthritis with the Help of Deep Learning and Transfer Learning
Sharma, G., Anand, V., & Kumar, V. (2023) · International Conference on Inventive Research in Computing Applications (ICIRCA) · doi:10.1109/icirca57980.2023.10220816 · Cited by 31
Gkintoni, E., Panagioti, M., Vassilopoulos, S. P., Νικολάου, Γ., Boutsinas, B., & Vantarakis, A. (2025) · Healthcare · doi:10.3390/healthcare13151776 · Cited by 37
Concatenated CNN-Based Pneumonia Detection Using a Fuzzy-Enhanced Dataset
Buriboev, A. S., Muhamediyeva, D., Primova, H., Sultanov, D., Akhmatovich, T. K., & Jeon, H. S. (2024) · Sensors · doi:10.3390/s24206750 · Cited by 28
Multi-Class Brain Lesion Classification Using Deep Transfer Learning With MobileNetV3
Majeed, A. F., Salehpour, P., Farzinvash, L., & Pashazadeh, S. (2024) · IEEE Access · doi:10.1109/access.2024.3413008 · Cited by 25
“To address the need for streamlined computational processes with simpler trainable neural network models and lessen the dependency on vast amounts of data, the adoption of Transfer Learning (TL) methods becomes imperative [7].”
MethodReal-Time Detection of UAV Small Target’s Infrared Imaging via OSTD-YOLOv8
Dahri, F. H., Laghari, A. A., Dahri, N. A., Abro, G. E. M., Estrela, V. V., de Oliveira, G. G., Iano, Y., & Chuma, E. L. (2024) · IEEE URUCON · doi:10.1109/urucon63440.2024.10850067 · Cited by 23
“Recently, "unmanned aerial vehicles" (UAVs, aka drones) have broadly taken aerial photos across various sectors [1] [2], [3], encompassing economics [4], commercial ventures [5], leisure, military activities [6], and academic research.”
Diao, Y., & Jelescu, I. (2022) · Magnetic Resonance in Medicine · doi:10.1002/mrm.29495 · Cited by 22
“Transfer learning algorithms may be better suited in such cases (Banerjee et al., 2018; Chen et al., 2020; Valverde et al., 2021).”
BackgroundAhamed, M. F., Islam, M. R., Hossain, T., Syfullah, K., & Sarkar, O. (2023) · Studies in autonomic, data-driven and industrial computing · doi:10.1007/978-981-19-7528-8_18 · Cited by 26
Valverde, J. M., Shatillo, A., & Tohka, J. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.127817 · Cited by 25
Self-citationPandimurugan, V., Rajasoundaran, S., Routray, S., Prabu, A. V., Alyami, H., Alharbi, A., & Ahmad, S. (2022) · Computational Intelligence and Neuroscience · doi:10.1155/2022/6671234 · Cited by 19
“Thirdly, Haralick model texture descriptors help in classifying diseased images from the normal image [35].”
BackgroundRetractedGuan, H., Yue, L., Yap, P., Xiao, S., Bozoki, A., & Liu, M. (2023) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2023.3257081 · Cited by 20
“Transfer learning [19], [20], [21] and domain adaptation [22], [23] have been used to tackle problems when the target domain has relatively fewer data and different distribution from the source domain.”
Sriraam, N., Chinta, B., Seshadri, S., & Suresh, S. (2025) · Artificial Intelligence Review · doi:10.1007/s10462-025-11160-7 · Cited by 24
“This motivates the model to extract features that are applicable to both the source and target domains (Valverde et al. 2021 ).”
Montin, E., Deniz, C. M., Kijowski, R., Youm, T. J., & Lattanzi, R. (2024) · Informatics in Medicine Unlocked · doi:10.1016/j.imu.2023.101444 · Cited by 17
“To maintain the anatomical shape of the hips as realistic as possible, no scaling was applied to the datasets [21].”
BackgroundMethodInfluentialSelf-Supervised Learning with Generative Adversarial Networks for Electron Microscopy
Kazimi, B., Ruzaeva, K., & Sandfeld, S. (2024) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw63382.2024.00012 · Cited by 26
“The robust representations acquired through self-supervised learning contribute to improved performance in real-world scenarios, establishing it as an essential strategy in computer vision tasks across diverse domains [16, 35, 42, 48].”
Qin, P., & Zhao, L. (2023) · IEEE Transactions on Transportation Electrification · doi:10.1109/tte.2023.3324822 · Cited by 16
“Transfer learning offers a solution way, the core of which is to find the similarity or lessen the distribution difference between data domains [17].”
BackgroundZaboski, B. A., & Bednarek, L. (2025) · Journal of Clinical Medicine · doi:10.3390/jcm14072442 · Cited by 17
Diagnosis of Alzheimer’s Disease and Frontotemporal Dementia From Electroencephalography Signals
Lee, D.-G., & Lee, S.-B. (2025) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2025.3575840 · Cited by 16
“While deep learning approaches have demonstrated improved performance [24], [25], these methodologies require extensive datasets—a significant constraint in medical artificial intelligence (AI) applications [26].”
Deep-Learning-Based Segmentation of Extraocular Muscles from Magnetic Resonance Images
Qureshi, A., Lim, S., Suh, S. Y., Mutawak, B., Chitnis, P. V., Demer, J. L., & Wei, Q. (2023) · Bioengineering · doi:10.3390/bioengineering10060699 · Cited by 18
Benchmarking the influence of pre-training on explanation performance in MR image classification
Oliveira, M., Wilming, R., Clark, B., Budding, C., Eitel, F., Ritter, K., & Haufe, S. (2024) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2024.1330919 · Cited by 18
“TL is, therefore, frequently employed for prediction tasks in medical imaging, where it is believed that TL techniques improves the generalization by identifying common features between the two tasks (Valverde et al., 2021).”
Chen, J., Wee, L., Dekker, A. L. A. J., & Bermejo, I. (2023) · Medical Physics · doi:10.1002/mp.16430 · Cited by 16
Gorenshtein, A., Liba, T., & Goren, A. (2025) · Journal of Imaging Informatics in Medicine · doi:10.1007/s10278-025-01686-1 · Cited by 15
“TumorGANet [9] employs ResNet-50 in conjunction with a GAN-based data augmentation strategy, reporting a classification accuracy of 99.53%.”
InfluentialTalaat, M., Si, X. A., & Xi, J. (2023) · Applied Sciences · doi:10.3390/app131810270 · Cited by 14
Tian, S., Wang, C., Zhang, R., Dai, Z., Jia, L., Zhang, W., Wang, J., & Liu, Y. (2022) · Frontiers in Oncology · doi:10.3389/fonc.2022.856346 · Cited by 14
Feo, R. D., Hämäläinen, E., Manninen, E., Immonen, R. J., Valverde, J. M., Ndode‐Ekane, X. E., Gröhn, O. H. J., Pitkänen, A., & Tohka, J. (2022) · Frontiers in Neurology · doi:10.3389/fneur.2022.820267 · Cited by 16
“, transfer learning (60), it is still an open question for CNNs.”
BackgroundSelf-citationPseudo-Label Assisted nnU-Net enables automatic segmentation of 7T MRI from a single acquisition
Donnay, C. A., Dieckhaus, H., Tsagkas, C., Gaitán, M. I., Beck, E., Mullins, A., Reich, D. S., & Nair, G. (2023) · Frontiers in Neuroimaging · doi:10.3389/fnimg.2023.1252261 · Cited by 13
“Prior studies indicate that transfer learning is often advantageous when the training dataset in the target domain is small (Valverde et al., 2021).”
BackgroundClassification of Brain Tumors: A Comparative Approach of Shallow and Deep Neural Networks
Montoya, S., Rojas, A. E., & Vásquez, L. F. N. (2024) · SN Computer Science · doi:10.1007/s42979-023-02431-7 · Cited by 13
“This technique is commonly employed in transfer learning, where a model leverages its existing knowledge to excel in new tasks, resulting in significant time and resource savings compared to training a model from scratch [17].”
Gómez‐Ramírez, J., Fernández‐Blázquez, M. A., & González-Rosa, J. J. (2022) · Brain Sciences · doi:10.3390/brainsci12050579 · Cited by 13
“A promising approach for achieving acceptable harmonization and coherence prediction results is transfer learning [83].”
BackgroundWang, L., Fatemi, M., & Alizad, A. (2025) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2025.110312 · Cited by 13
Du, Y., Wang, G., Wang, C., Zhang, Y., Xi, X., Zhang, L., & Liu, M. (2023) · Frontiers in Aging Neuroscience · doi:10.3389/fnagi.2023.1101879 · Cited by 15
“To alleviate this problem, we will utilize transfer learning (Pan and Yang, 2010; Valverde et al., 2021) or meta-learning (Finn et al., 2017; Hospedales et al., 2021) strategy to model the modularity structure of FBN.”
MethodMRI-based model for MCI conversion using deep zero-shot transfer learning
Ren, F., Yang, C., & Nanehkaran, Y. A. (2022) · The Journal of Supercomputing · doi:10.1007/s11227-022-04668-0 · Cited by 13
“Transfer learning techniques are effective in overcoming the overfitting problem associated with relatively complex models when the number of available direct training samples is limited, and they are gradually being applied to solve medical image-related tasks [35].”
Dang, M., Liu, B., Chen, Y., & Zhang, Z. (2026) · Psychoradiology · doi:10.1093/psyrad/kkag007 · Cited by 10
Mounika, G., Kollem, S., & Samala, S. (2025) · Neuroradiology · doi:10.1007/s00234-025-03685-z · Cited by 13
Decoodt, P., Sierra-Sosa, D., Anghel, L., Cuminetti, G., Keyzer, E. D., & Morissens, M. (2024) · Diagnostics · doi:10.3390/diagnostics14131439 · Cited by 10
“A description of the video clip classification models can be found on GitHub at the official site of MoViNet https://github.com/tensorflow/ models/blob/master/official/projects/movinet/README.md (accessed on 12 December 2013).”
MethodDeep-Stacked Convolutional Neural Networks for Brain Abnormality Classification Based on MRI Images
Rumala, D. J., van Ooijen, P. M. A., Rachmadi, R. F., Sensusiati, A. D., & Purnama, I. K. E. (2023) · Journal of Imaging Informatics in Medicine · doi:10.1007/s10278-023-00828-7 · Cited by 10
“Since the features of medical images with the ImageNet dataset may differ immensely [45], fine-tuning is executed to address the differences in image features.”
BackgroundMethodTalaat, M., Xi, J. S., Tan, K., Si, X. A., & Xi, J. (2023) · Journal of Nanotheranostics · doi:10.3390/jnt4030011 · Cited by 9
“Transfer learning has become increasingly popular in medical image-based diagnostic systems based on existing CNN models such as AlexNet, GoogleNet, ResNet, DenseNet, MobileNet, etc. [22–24].”
MethodIsmaila, L., Rasti, P., Bernard, F., Labriffe, M., Meneï, P., Minassian, A. T., Rousseau, D., & Lemée, J.-M. (2022) · Applied Sciences · doi:10.3390/app12146925 · Cited by 11
Hu, T., Mousavirad, S. J., Afsharizadeh, M., & O’Nils, M. (2026) · International Journal of Computational Intelligence Systems · doi:10.1007/s44196-025-01115-9 · Cited by 8
“Similarly, Valverde et al. (2021) highlighted that TL not only improves performance but also enhances training stability across various imaging modalities [81].”
Shou, Q., Zhao, C., Shao, X., Herting, M. M., & Wang, D. J. J. (2025) · NeuroImage · doi:10.1016/j.neuroimage.2025.121070 · Cited by 7
“However, in recent years, pretrained models have shown to be powerful especially in the case where there are not sufficient training data(Valverde et al. 2021; Salehi et al. 2023; H. E. Kim et al. 2022).”
Guo, Y., Lin, Q., Wang, Y., Cao, X., Cao, Y., Man, Z., Zeng, X., & Huang, X. (2022) · Journal of Medical and Biological Engineering · doi:10.1007/s40846-022-00770-z · Cited by 7
Mukherjee, T., Gour, S., Abbasi, S. F., Pournik, O., & Arvanitis, T. N. (2024) · Studies in health technology and informatics · doi:10.3233/shti240745 · Cited by 8
A Multi-Label Deep Learning Model for Detailed Classification of Alzheimer's Disease
Yang, M., Zhao, Y., Yu, H., Chen, S., Gao, G., Li, D., Wu, X., Huang, L., & Ye, S. (2025) · Actas Españolas de Psiquiatría · doi:10.62641/aep.v53i1.1728 · Cited by 7
Niroomand, B., Mohammadzadeh, I., Hajikarimloo, B., Habibi, M. A., Mohammadzadeh, S., Bahri, A., Bagheri, M. H., Albakr, A. A., Karmur, B. S., & Borghei-Razavi, H. (2025) · Neurosurgical Review · doi:10.1007/s10143-025-03744-2 · Cited by 6
Transferring Models Trained on Natural Images to 3D MRI via Position Encoded Slice Models
Gupta, U., Chattopadhyay, T., Dhinagar, N. J., Thompson, P. M., Steeg, G. V., & Initiative, T. A. D. N. (2023) · IEEE International Symposium on Biomedical Imaging (ISBI) · doi:10.1109/isbi53787.2023.10230660 · Cited by 9
“This has led to an active debate in the radiology field on how to pretrain deep learning methods for MRI-based tasks, such as disease clas-sification and staging, identifying pathology, and anatomical segmentation (e.g., [1, 2, 3]).”
BackgroundDeep Learning Techniques in Classification of Stages in Dementia: An Ensemble Approach
Pandiyarajan, M., & Valarmathi, R. S. (2023) · Journal of Advanced Research in Applied Sciences and Engineering Technology · doi:10.37934/araset.36.1.131146 · Cited by 6
Bidirectional brain image translation using transfer learning from generic pre-trained models
Haimour, F., Al-Sayyed, R. M. H., Mahafza, W. S., & Al-Kadi, O. S. (2024) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2024.104100 · Cited by 6
Kim, S., Kim, M.-J., Kim, J.-S., Jang, J., Choi, W., Lee, H., Song, J.-H., & Hoe, H. (2026) · Alzheimer s Research & Therapy · doi:10.1186/s13195-026-02013-8 · Cited by 6
“Federated learning and transfer learning offer potential solutions by enabling model generalization while preserving patient privacy [111, 112].”
Petiton, S., Grigis, A., Dufumier, B., & Duchesnay, É. (2024) · IEEE International Symposium on Biomedical Imaging (ISBI) · doi:10.1109/isbi56570.2024.10635777 · Cited by 5
“On the other hand, TL can improve predictions by using a pre-trained model before fine-tuning previously acquired knowledge to the desired classification task [6].”
BackgroundCavicchioli, M., Moglia, A., Pierelli, L., Pugliese, G., & Cerveri, P. (2024) · Computerized Medical Imaging and Graphics · doi:10.1016/j.compmedimag.2024.102434 · Cited by 5
Autonomous Drone-Based Border Surveillance Using Real-Time Object Detection with Yolo
Alqahtani, S. K., & Abro, G. E. M. (2025) · IEEE Symposium on Computer Applications & Industrial Electronics (ISCAIE) · doi:10.1109/iscaie64985.2025.11080940 · Cited by 5
“Unmanned aerial vehicles (UAVs) have emerged as a disruptive solution due to their flexibility, cost-efficiency, and capacity to visit remote places [1–3].”
Lam, L., Lam, A., Bacchi, S., & Abou‐Hamden, A. (2022) · British Journal of Neurosurgery · doi:10.1080/02688697.2022.2151565 · Cited by 5
Sharma, R., & Acharya, V. (2025) · Scientific Reports · doi:10.1038/s41598-025-28243-8 · Cited by 4
“The adaptation leverages labelled data from the original training ensuring efficient knowledge transfer and reducing computational overhead 29,30 .”
Kaswar, A. B., Ilham, M. R., Tahir, R. A., Mappalotteng, A. M., Djawad, Y. A., & Andayani, D. D. (2023) · International Workshop on Artificial Intelligence and Image Processing (IWAIIP) · doi:10.1109/iwaiip58158.2023.10462730 · Cited by 4
“Transfer learning allows models to utilize knowledge from related tasks to improve performance on different tasks [22].”
BackgroundMethodInfluentialZhu, X. (2024) · International Conference on Communications, Information System and Computer Engineering (CISCE) · doi:10.1109/cisce62493.2024.10653207 · Cited by 7
“Valverde et al [19] conducted research on the application of transfer learning in magnetic resonance (MR) brain imaging.”
BackgroundYang, X., Xu, X., & Ying, J. (2024) · IEEE International Conference on Big Data and Artificial Intelligence (BDAI) · doi:10.1109/bdai62182.2024.10692826 · Cited by 4
Pafitis, M., Constantinou, M., & Christodoulou, C. C. (2024) · IEEE Access · doi:10.1109/access.2024.3487114 · Cited by 4
“Previous work has utilized transfer learning techniques to reduce the training time and improve performance of CNN models for MRI classification (reviewed in [16], [54]).”
Feng, J., Li, Y., Huang, Z., Chen, Y., Lu, S., Hu, R. L., Hu, Q., Chen, Y., Wang, X., Fan, Y., et al. (2025) · Frontiers in Human Neuroscience · doi:10.3389/fnhum.2025.1555690 · Cited by 4
Talaat, M., Si, X. A., & Xi, J. (2023) · Data · doi:10.3390/data8080126 · Cited by 4
Jiao, M., Liu, H., Yang, Z., Tian, S., Ouyang, H., Li, Y., Yuan, Y., Liu, J., Wang, C., Lang, N., et al. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-17266-3_6 · Cited by 4
Liu, S., Shang, X., Peng, W., Wan, W., & Zhang, J.-B. (2023) · IEEE Journal of Electromagnetics RF and Microwaves in Medicine and Biology · doi:10.1109/jerm.2023.3281057 · Cited by 4
“Current imaging techniques for detecting cerebrovascular diseases include computed tomography (CT), magnetic resonance imaging (MRI), positron emission tomography (PET) and transcranial doppler (TCD), etc. [3], [4], [5], [6].”
MethodAdvancing the Reliability of Ultra-Low Field MRI Brain Volume Analysis Using CycleGAN
Hsu, P., Marchetto, E., Sodickson, D. K., Johnson, P., & Veraart, J. (2025) · Communications in computer and information science · doi:10.1007/978-3-031-79103-1_6 · Cited by 4
A Review of Technical Coherence between Brain Tumors and their Diagnostic Imaging Spectra
Nekkantti, C., Ratna, P. K. V., Korabathina, A., Guddanti, S. S., Vallabhaneni, L. T., & Ramesh, P. (2022) · International Conference on Electronics and Renewable Systems (ICEARS) · doi:10.1109/icears53579.2022.9752458 · Cited by 3
L, D. P., Narain, P., & Sreena, V. G. (2023) · Intelligent Data Analysis · doi:10.3233/ida-227321 · Cited by 3
“Retraining of an existing model is known as transfer learning [16] and the preprocessing method that needs to be applied while transfer learning is image resizing as the input layer may be different in size when compared to the input data.”
Rumala, D. J., Rachmadi, R. F., Sensusiati, A. D., & Purnama, I. K. E. (2024) · EMITTER International Journal of Engineering Technology · doi:10.24003/emitter.v12i2 · Cited by 3
From Lesion to Decision: AI for ARIA Detection and Predictive Imaging in Alzheimer’s Disease
Christodoulou, R. C., Papageorgiou, P. S., Sarquis, M. D., Rivera, L., Gonzalez, C. M., Eller, D. W., Rivera, G., Petrou, V., Vamvouras, G., Vassiliou, E. K., et al. (2025) · Biomedicines · doi:10.3390/biomedicines13112739 · Cited by 3
“These findings support using pre-trained, general architectures as effective starting points for accelerating ARIA tool development by leveraging pre-trained weights and augmentation strategies to adapt to amyloid-treated populations [53,54].”
Deep Neural Networks with Transfer Learning for Medical Image Classification
Joshi, G. (2025) · EPJ Web of Conferences · doi:10.1051/epjconf/202534101040 · Cited by 2
Valverde, J. M., Imani, V., & Tohka, J. (2026) · Pattern Recognition Letters · doi:10.1016/j.patrec.2026.01.017 · Cited by 3
Self-citationAl-Sarairah, A. H., Al-Lasasmeh, O. A. A. M., Tawil, A. A., Alkasasbeh, M. S., & Gharaibeh, N. M. (2026) · The Egyptian Journal of Radiology and Nuclear Medicine · doi:10.1186/s43055-026-01782-4 · Cited by 2
“Transfer learning, which uses features learned on large natural-image datasets such as ImageNet, frequently improves the performance on medical tasks with limited labeled data [8].”
Brzakala, C., Moultos, O. A., van der Hoek, J. P., & Taormina, R. (2026) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.5c03121 · Cited by 2
Yanishevskaya, N., & Bolodurina, I. P. (2021) · Bulletin of the South Ural State University Ser Computer Technologies Automatic Control & Radioelectronics · doi:10.14529/ctcr210301 · Cited by 2
APPLICATION OF COMPUTER VISION FOR DIAGNOSTICS OF NOSOLOGICAL UNITS ON MEDICAL IMAGES
Teplyakova, A. R., & Старков, С. О. (2022) · Южно-Сибирский научный вестник · doi:10.25699/sssb.2022.44.4.004 · Cited by 2
Deep Learning-based Neural Network for Automated Brain Tumor Diagnosis from MRI Images
Uthradevi, G., Mohanbabu, G., Kumar, B. R. S., Tamilselvi, A., Ashok, A., & Sangeetha, R. (2023) · International Conference on Data Science, Agents & Artificial Intelligence (ICDSAAI) · doi:10.1109/icdsaai59313.2023.10452581 · Cited by 2
“In medical imaging and diagnostics, new paths have been made possible by the application of neural network-based object detection frameworks and algorithms in brain tumor diagnosis [9].”
MethodWang, Y., Shi, V., Hsu, W.-C., Dai, Y., Yao, S., Zhong, Z., Zhang, Z., Wu, J., Maxwell, A. W. P., Collins, S., et al. (2025) · Physics in Medicine and Biology · doi:10.1088/1361-6560/adfc20 · Cited by 5
“Transfer learning techniques, particularly those optimized for medical imaging tasks, offer a compelling path forward in this regard (Valverde et al 2021).”
Learning diverse and generic representations of the brain with large-scale multi-task pretrainingPreprint
Leonardsen, E. H., Dahl, A., Holm, M. C., de Lange, A. G., Grødem, E., Marquand, A. F., Sørensen, Ø., Schwarz, E., Andreassen, O. A., Wolfers, T., et al. (2025) · medRxiv · doi:10.64898/2025.12.19.25342659 · Cited by 2
Metadata
- Authors
- Valverde, J. M., Imani, V., Abdollahzadeh, A., De Feo, R., Prakash, M., Ciszek, R., & Tohka, J.
- Affiliations
- University of Eastern Finland (FI)
- A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, 70150 Kuopio, Finland
- Corresponding author
- Tohka, Jussi
- ORCID
- Venue
- Journal of Imaging
- Publisher
- MDPI AG
- Published
- 2021-04-01
- Volume
- 7
- Issue
- 4
- Pages
- 66
- ISSN
- 2313-433X (ISSN-L)
- PMC
- PMC8321322
- OpenAlex
- W3126990165
- Cited by
- 137
- References
- 186
- Citations by year
- 2026: 23 · 2025: 24 · 2024: 24 · 2023: 32 · 2022: 29 · 2021: 5
- FWCI
- 1.63
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Journal of Imaging (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- IRIS Research product catalog (Sapienza University of Rome) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Academy of Finland (316258)
- H2020 Marie Skłodowska-Curie Actions (740264)
- European Social Fund (S21770)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Medical Image Segmentation Techniques
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Brain Tumor Detection and Classification
- Keywords
- transfer learning, brain MRI, convolutional neural networks, pre-training and fine-tuning, dementia classification, brain tumor segmentation, Alzheimer's disease, unseen target domains, unlabeled data
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.
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
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.