pith. sign in

arxiv: 2605.23995 · v4 · pith:LKNWGSQKnew · submitted 2026-05-18 · 💻 cs.CV · cs.AI

Task-Aligned Self-Supervised Learning for Medical Image Analysis: A Task-Oriented Review with Practical Design Guidelines

Chathura Wimalasiri , Kishor Nandakishor , Marimuthu Palaniswami This is my paper

Pith reviewed 2026-07-01 07:22 UTC · model grok-4.3

classification 💻 cs.CV cs.AI
keywords self-supervised learningmedical image analysiscontrastive learningreconstruction-based learningtask alignmentsegmentationclassificationfew-shot learning
0
0 comments X

The pith

No single SSL strategy is optimal for all medical imaging tasks; alignment between pretext and downstream objective determines performance.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This review synthesizes 75 studies on self-supervised learning for medical images. It shows that pretext task design must match the clinical goal because contrastive methods produce global features suited to classification while reconstruction and spatial prediction methods retain local structure needed for segmentation. The four-paradigm organization reveals that hybrid approaches trade training cost for balanced representations, and that SSL gains are largest in low-label settings when augmentations preserve pathology. Practical guidelines follow directly from these task-paradigm interactions.

Core claim

The evidence suggests that no SSL strategy is universally optimal. Contrastive objectives generally encourage global discriminative representations and are well aligned with classification, but may underrepresent subtle or localized pathology. Spatial prediction, masked modeling, and reconstruction-based objectives better preserve anatomical structure and are often more suitable for segmentation and dense prediction. Hybrid methods can provide balanced representations, although they increase training complexity.

What carries the argument

The four-paradigm categorization (contrastive, non-contrastive/predictive, generative/reconstruction-based, hybrid) that links each paradigm to specific downstream tasks, modalities, and label regimes.

If this is right

  • Contrastive SSL is the default choice when the downstream task is image-level classification.
  • Reconstruction or masked modeling SSL is preferred when the downstream task requires pixel-level or dense predictions.
  • Hybrid SSL is useful only when training compute can absorb the added complexity.
  • SSL yields the largest gains in few-shot and low-label regimes provided the augmentations avoid erasing pathology.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Pathology-aware corruption strategies could further improve detection of subtle lesions that current generic augmentations may destroy.
  • Standardized cross-task benchmarks would make it possible to quantify the claimed task-alignment effect rather than relying on study-by-study comparisons.
  • Resource-efficient variants of the hybrid paradigm are needed before high-resolution 3D volumes can be routinely pre-trained without prohibitive cost.

Load-bearing premise

The 75 selected studies are representative of the field and the four-paradigm split adequately captures how pretext design interacts with modality, label count, and task performance.

What would settle it

A controlled experiment that trains one model from each of the four paradigms on identical medical datasets for both a classification task and a segmentation task, then measures whether the performance ordering reverses or remains stable across tasks.

Figures

Figures reproduced from arXiv: 2605.23995 by Chathura Wimalasiri, Kishor Nandakishor, Marimuthu Palaniswami.

Figure 1
Figure 1. Figure 1: General workflow of SSL. The top panel illustrates the pretraining stage using unlabeled data, while the [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: This categorization captures recent methodological developments and provides a structured foundation for the [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 2
Figure 2. Figure 2: Literature-derived taxonomy of SSL pretext task strategies in medical image analysis. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: PRISMA flow diagram illustrating the study selection process for the systematic review. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
read the original abstract

Self-supervised learning (SSL) is increasingly used in medical image analysis to reduce dependence on costly expert annotations by learning transferable representations from unlabeled data. However, SSL performance depends not only on model architecture, but also on whether the pretext task preserves information required by the downstream clinical objective. This review presents a task-oriented synthesis of SSL methods for medical imaging, focusing on how pretext-task design interacts with imaging modality, label availability, and downstream performance. We analyze 75 studies published from 2017 to 2025 and organize them into four paradigms: contrastive learning, non-contrastive and predictive learning, generative and reconstruction-based learning, and hybrid learning. Rather than cataloging methods chronologically, we examine how these paradigms support classification, segmentation, detection, reconstruction, and regression. The evidence suggests that no SSL strategy is universally optimal. Contrastive objectives generally encourage global discriminative representations and are well aligned with classification, but may underrepresent subtle or localized pathology. Spatial prediction, masked modeling, and reconstruction-based objectives better preserve anatomical structure and are often more suitable for segmentation and dense prediction. Hybrid methods can provide balanced representations, although they increase training complexity. Across modalities, SSL is most beneficial in low-label and few-shot regimes, but its effectiveness depends on modality-aware augmentation, pathology-preserving corruption, and clinically meaningful evaluation. We conclude with practical design guidelines and identify open challenges, including pathology-aware pretext tasks, resource-efficient training for high-dimensional data, and standardized evaluation protocols.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. This paper is a task-oriented literature review synthesizing 75 SSL studies (2017–2025) in medical image analysis. It categorizes methods into four paradigms (contrastive, non-contrastive/predictive, generative/reconstruction-based, hybrid) and argues that pretext-task design must align with downstream clinical objectives, imaging modality, and label regime. No SSL approach is universally optimal: contrastive methods favor global classification but may miss localized pathology, while spatial/masked/reconstruction objectives better support segmentation and dense prediction; hybrids balance representations at higher complexity. The review emphasizes benefits in low-label regimes and ends with design guidelines plus open challenges.

Significance. If the synthesis is representative, the review supplies actionable, task-specific guidance for choosing SSL pretext tasks in medical imaging, a domain where annotation costs are high. It usefully distinguishes paradigm-task alignments and flags the need for pathology-aware augmentations and standardized evaluation, which could reduce trial-and-error in future work.

major comments (2)
  1. [Abstract] Abstract and (presumed) Methods section: the synthesis rests on 75 studies yet supplies no search strategy, inclusion/exclusion criteria, database sources, or quantitative aggregation of performance deltas. Without these, the representativeness of the sample and the claimed paradigm-task alignments cannot be verified and risk selection bias.
  2. [Introduction] Introduction / Paradigm organization: the four-paradigm taxonomy is asserted without explicit justification of why these categories (rather than, e.g., a modality- or loss-based taxonomy) best capture interactions with label availability and downstream task performance; the central claim that alignments are evidence-based therefore depends on an unexamined organizational choice.
minor comments (2)
  1. [Abstract] Abstract: the date range ends in 2025; clarify whether this includes in-press or early-access papers and how the cutoff was applied.
  2. [Abstract] The abstract states that hybrid methods “increase training complexity” but offers no concrete metrics (parameters, epochs, GPU hours) from the reviewed studies to support the claim.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on our task-oriented review. We address each major comment below, indicating planned revisions to enhance transparency and justification while preserving the manuscript's focus on paradigm-task alignments.

read point-by-point responses

  1. Referee: [Abstract] Abstract and (presumed) Methods section: the synthesis rests on 75 studies yet supplies no search strategy, inclusion/exclusion criteria, database sources, or quantitative aggregation of performance deltas. Without these, the representativeness of the sample and the claimed paradigm-task alignments cannot be verified and risk selection bias.

    Authors: We agree that explicit documentation of the selection process would improve verifiability. In the revised manuscript we will add a dedicated 'Literature Search and Selection' subsection describing the databases consulted (PubMed, arXiv, IEEE Xplore, Google Scholar), search keywords and time window (2017–2025), inclusion criteria (peer-reviewed or preprint works applying SSL to medical images with downstream clinical-task evaluation), and exclusion criteria (purely theoretical works or non-medical applications). Because the synthesis is qualitative rather than a formal meta-analysis, we will include a supplementary table summarizing reported performance trends in low-label regimes, accompanied by explicit caveats on experimental heterogeneity that preclude direct delta aggregation. revision: yes

  2. Referee: [Introduction] Introduction / Paradigm organization: the four-paradigm taxonomy is asserted without explicit justification of why these categories (rather than, e.g., a modality- or loss-based taxonomy) best capture interactions with label availability and downstream task performance; the central claim that alignments are evidence-based therefore depends on an unexamined organizational choice.

    Authors: The four paradigms were chosen precisely because they map onto distinct representation-learning mechanisms (global discrimination, local prediction, structural reconstruction, and their combinations) whose differential effects on global versus localized features are directly observable across the collected studies and align with clinical task requirements. We will expand the Introduction with a new paragraph that (i) contrasts this objective-based taxonomy against modality- or loss-centric alternatives, (ii) explains why the latter obscure the label-regime and task-type interactions that constitute the review's central contribution, and (iii) anchors the justification in concrete examples drawn from the 75 papers. revision: yes

Circularity Check

0 steps flagged

No circularity: literature review synthesis of external studies

full rationale

This is a qualitative literature review that synthesizes findings from 75 external papers across four paradigms. The central claims (no universal SSL optimum; paradigm-task alignments) are presented as conclusions drawn from the cited body of work rather than derived via equations, parameter fitting, or self-referential prediction. No self-definitional steps, fitted inputs renamed as predictions, or load-bearing self-citations appear in the provided text. The selection of studies and taxonomy are acknowledged as assumptions in the skeptic note, but these are standard review limitations rather than circular reductions of the sort defined in the analysis criteria. The derivation chain is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

The review rests on the representativeness of the selected literature and the validity of the four-paradigm taxonomy; no free parameters, new axioms, or invented entities are introduced.

pith-pipeline@v0.9.1-grok · 5810 in / 1044 out tokens · 38070 ms · 2026-07-01T07:22:14.242437+00:00 · methodology

Review history (2 revisions) →

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

103 extracted references · 5 canonical work pages · 1 internal anchor

  1. [1]

    A guide to deep learning in healthcare.Nature medicine, 25(1):24–29, 2019

    Andre Esteva, Alexandre Robicquet, Bharath Ramsundar, V olodymyr Kuleshov, Mark DePristo, Katherine Chou, Claire Cui, Greg Corrado, Sebastian Thrun, and Jeff Dean. A guide to deep learning in healthcare.Nature medicine, 25(1):24–29, 2019

    2019

  2. [2]

    Xiaoxuan Liu, Livia Faes, Aditya U Kale, Siegfried K Wagner, Dun Jack Fu, Alice Bruynseels, Thushika Mahendiran, Gabriella Moraes, Mohith Shamdas, Christoph Kern, et al. A comparison of deep learning performance against health-care professionals in detecting diseases from medical imaging: a systematic review and meta-analysis.The lancet digital health, 1(...

    2019

  3. [3]

    Preparing medical imaging data for machine learning.Radiology, 295(1):4–15, 2020

    Martin J Willemink, Wojciech A Koszek, Cailin Hardell, Jie Wu, Dominik Fleischmann, Hugh Harvey, Les R Folio, Ronald M Summers, Daniel L Rubin, and Matthew P Lungren. Preparing medical imaging data for machine learning.Radiology, 295(1):4–15, 2020

    2020

  4. [4]

    Embracing imperfect datasets: A review of deep learning solutions for medical image segmentation.Medical image analysis, 63:101693, 2020

    Nima Tajbakhsh, Laura Jeyaseelan, Qian Li, Jeffrey N Chiang, Zhihao Wu, and Xiaowei Ding. Embracing imperfect datasets: A review of deep learning solutions for medical image segmentation.Medical image analysis, 63:101693, 2020

    2020

  5. [5]

    Chexpert: A large chest radiograph dataset with uncertainty labels and expert comparison

    Jeremy Irvin, Pranav Rajpurkar, Michael Ko, Yifan Yu, Silviana Ciurea-Ilcus, Chris Chute, Henrik Marklund, Behzad Haghgoo, Robyn Ball, Katie Shpanskaya, et al. Chexpert: A large chest radiograph dataset with uncertainty labels and expert comparison. InProceedings of the AAAI conference on artificial intelligence, volume 33, pages 590–597, 2019

    2019

  6. [6]

    Transfusion: Understanding transfer learning for medical imaging.Advances in neural information processing systems, 32, 2019

    Maithra Raghu, Chiyuan Zhang, Jon Kleinberg, and Samy Bengio. Transfusion: Understanding transfer learning for medical imaging.Advances in neural information processing systems, 32, 2019

    2019

  7. [7]

    Big self-supervised models advance medical image classification

    Shekoofeh Azizi, Basil Mustafa, Fiona Ryan, Zachary Beaver, Jan Freyberg, Jonathan Deaton, Aaron Loh, Alan Karthikesalingam, Simon Kornblith, Ting Chen, et al. Big self-supervised models advance medical image classification. InProceedings of the IEEE/CVF international conference on computer vision, pages 3478–3488, 2021

    2021

  8. [8]

    Self-supervised learning in medicine and healthcare

    Rayan Krishnan, Pranav Rajpurkar, and Eric J Topol. Self-supervised learning in medicine and healthcare. Nature Biomedical Engineering, 6(12):1346–1352, 2022

    2022

  9. [9]

    Self-supervised visual feature learning with deep neural networks: A survey

    Longlong Jing and Yingli Tian. Self-supervised visual feature learning with deep neural networks: A survey. IEEE transactions on pattern analysis and machine intelligence, 43(11):4037–4058, 2020

    2020

  10. [10]

    S Kevin Zhou, Hayit Greenspan, Christos Davatzikos, James S Duncan, Bram Van Ginneken, Anant Madabhushi, Jerry L Prince, Daniel Rueckert, and Ronald M Summers. A review of deep learning in medical imaging: Imaging traits, technology trends, case studies with progress highlights, and future promises.Proceedings of the IEEE, 109(5):820–838, 2021

    2021

  11. [11]

    Masked autoencoders are scalable vision learners

    Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, and Ross Girshick. Masked autoencoders are scalable vision learners. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 16000–16009, 2022

    2022

  12. [12]

    A simple framework for contrastive learning of visual representations

    Ting Chen, Simon Kornblith, Mohammad Norouzi, and Geoffrey Hinton. A simple framework for contrastive learning of visual representations. InInternational conference on machine learning, pages 1597–1607. PmLR, 2020

    2020

  13. [13]

    Unsupervised Representation Learning by Predicting Image Rotations

    Spyros Gidaris, Praveer Singh, and Nikos Komodakis. Unsupervised representation learning by predicting image rotations.arXiv preprint arXiv:1803.07728, 2018

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  14. [14]

    Unsupervised learning of visual representations by solving jigsaw puzzles

    Mehdi Noroozi and Paolo Favaro. Unsupervised learning of visual representations by solving jigsaw puzzles. In European conference on computer vision, pages 69–84. Springer, 2016

    2016

  15. [15]

    Contrastive learning of global and local features for medical image segmentation with limited annotations.Advances in neural information processing systems, 33:12546–12558, 2020

    Krishna Chaitanya, Ertunc Erdil, Neerav Karani, and Ender Konukoglu. Contrastive learning of global and local features for medical image segmentation with limited annotations.Advances in neural information processing systems, 33:12546–12558, 2020

    2020

  16. [16]

    Demystifying contrastive self-supervised learning: Invariances, augmentations and dataset biases.Advances in Neural Information Processing Systems, 33:3407–3418, 2020

    Senthil Purushwalkam and Abhinav Gupta. Demystifying contrastive self-supervised learning: Invariances, augmentations and dataset biases.Advances in Neural Information Processing Systems, 33:3407–3418, 2020. 26

    2020

  17. [17]

    Multimodal mr synthesis via modality-invariant latent representation.IEEE transactions on medical imaging, 37(3):803–814, 2017

    Agisilaos Chartsias, Thomas Joyce, Mario Valerio Giuffrida, and Sotirios A Tsaftaris. Multimodal mr synthesis via modality-invariant latent representation.IEEE transactions on medical imaging, 37(3):803–814, 2017

    2017

  18. [18]

    Medgan: Medical image translation using gans.Computerized medical imaging and graphics, 79:101684, 2020

    Karim Armanious, Chenming Jiang, Marc Fischer, Thomas Küstner, Tobias Hepp, Konstantin Nikolaou, Sergios Gatidis, and Bin Yang. Medgan: Medical image translation using gans.Computerized medical imaging and graphics, 79:101684, 2020

    2020

  19. [19]

    Self-supervised learning methods and applications in medical imaging analysis: A survey.PeerJ Computer Science, 8:e1045, 2022

    Saeed Shurrab and Rehab Duwairi. Self-supervised learning methods and applications in medical imaging analysis: A survey.PeerJ Computer Science, 8:e1045, 2022

    2022

  20. [20]

    Self-supervised learning for medical image classification: a systematic review and implementation guidelines

    Shih-Cheng Huang, Anuj Pareek, Malte Jensen, Matthew P Lungren, Serena Yeung, and Akshay S Chaudhari. Self-supervised learning for medical image classification: a systematic review and implementation guidelines. NPJ Digital Medicine, 6(1):74, 2023

    2023

  21. [21]

    A survey of the impact of self-supervised pretraining for diagnostic tasks in medical x-ray, ct, mri, and ultrasound.BMC medical imaging, 24(1):79, 2024

    Blake VanBerlo, Jesse Hoey, and Alexander Wong. A survey of the impact of self-supervised pretraining for diagnostic tasks in medical x-ray, ct, mri, and ultrasound.BMC medical imaging, 24(1):79, 2024

    2024

  22. [22]

    Self-supervised visual representation learning for medical image analysis: A comprehensive survey.Transactions on Machine Learning Research, 2024

    Siladittya Manna, Saumik Bhattacharya, and Umapada Pal. Self-supervised visual representation learning for medical image analysis: A comprehensive survey.Transactions on Machine Learning Research, 2024

    2024

  23. [23]

    3d self-supervised methods for medical imaging.Advances in neural information processing systems, 33:18158–18172, 2020

    Aiham Taleb, Winfried Loetzsch, Noel Danz, Julius Severin, Thomas Gaertner, Benjamin Bergner, and Christoph Lippert. 3d self-supervised methods for medical imaging.Advances in neural information processing systems, 33:18158–18172, 2020

    2020

  24. [24]

    Rubik’s cube+: A self-supervised feature learning framework for 3d medical image analysis.Medical image analysis, 64:101746, 2020

    Jiuwen Zhu, Yuexiang Li, Yifan Hu, Kai Ma, S Kevin Zhou, and Yefeng Zheng. Rubik’s cube+: A self-supervised feature learning framework for 3d medical image analysis.Medical image analysis, 64:101746, 2020

    2020

  25. [25]

    Extracting and composing robust features with denoising autoencoders

    Pascal Vincent, Hugo Larochelle, Yoshua Bengio, and Pierre-Antoine Manzagol. Extracting and composing robust features with denoising autoencoders. InProceedings of the 25th international conference on Machine learning, pages 1096–1103, 2008

    2008

  26. [26]

    Momentum contrast for unsupervised visual representation learning

    Kaiming He, Haoqi Fan, Yuxin Wu, Saining Xie, and Ross Girshick. Momentum contrast for unsupervised visual representation learning. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 9729–9738, 2020

    2020

  27. [27]

    Dense contrastive learning for self- supervised visual pre-training

    Xinlong Wang, Rufeng Zhang, Chunhua Shen, Tao Kong, and Lei Li. Dense contrastive learning for self- supervised visual pre-training. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 3024–3033, 2021

    2021

  28. [28]

    A survey on self-supervised learning: Recent advances and open problems.Neurocomputing, page 131409, 2025

    Jianping Zhang, Lei Yang, Seyed Mahmoud Sajjadi Mohammadabadi, and Feng Yan. A survey on self-supervised learning: Recent advances and open problems.Neurocomputing, page 131409, 2025

    2025

  29. [29]

    A review on discriminative self-supervised learning methods in computer vision.arXiv preprint arXiv:2405.04969, 2024

    Nikolaos Giakoumoglou, Tania Stathaki, and Athanasios Gkelias. A review on discriminative self-supervised learning methods in computer vision.arXiv preprint arXiv:2405.04969, 2024

    work page arXiv 2024

  30. [30]

    Maemc-net: a hybrid self-supervised learning method for predicting the malignancy of solitary pulmonary nodules from ct images.Frontiers in medicine, 12:1507258, 2025

    Tianhu Zhao, Yong Yue, Hang Sun, Jingxu Li, Yanhua Wen, Yudong Yao, Wei Qian, Yubao Guan, and Shouliang Qi. Maemc-net: a hybrid self-supervised learning method for predicting the malignancy of solitary pulmonary nodules from ct images.Frontiers in medicine, 12:1507258, 2025

    2025

  31. [31]

    The prisma 2020 statement: an updated guideline for reporting systematic reviews.bmj, 372, 2021

    Matthew J Page, Joanne E McKenzie, Patrick M Bossuyt, Isabelle Boutron, Tammy C Hoffmann, Cynthia D Mulrow, Larissa Shamseer, Jennifer M Tetzlaff, Elie A Akl, Sue E Brennan, et al. The prisma 2020 statement: an updated guideline for reporting systematic reviews.bmj, 372, 2021

    2020

  32. [32]

    Self supervised contrastive learning for digital histopathology

    Ozan Ciga, Tony Xu, and Anne Louise Martel. Self supervised contrastive learning for digital histopathology. Machine learning with applications, 7:100198, 2022

    2022

  33. [33]

    Self-supervised learning for accurate liver view classifica- tion in ultrasound images with minimal labeled data

    Abder-Rahman Ali, Anthony E Samir, and Peng Guo. Self-supervised learning for accurate liver view classifica- tion in ultrasound images with minimal labeled data. Inproceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3087–3093, 2023

    2023

  34. [34]

    Self-supervised learning leads to improved performance in biparametric prostate mri classification.Computers in Biology and Medicine, 198:111262, 2025

    José Guilherme de Almeida, Ana Sofia Castro Verde, Ana Mascarenhas Gaivão, Carlos Bilreiro, Inês Santiago, Joana Ip, Sara Belião, Celso Matos, Manolis Tsiknakis, Kostas Marias, et al. Self-supervised learning leads to improved performance in biparametric prostate mri classification.Computers in Biology and Medicine, 198:111262, 2025

    2025

  35. [35]

    Lingmiao Meng, Liwei Zhao, Xin Yi, and Qingming Yu. Self-supervised contrastive learning for automated segmentation of brain tumor mri images in schizophrenia.International Journal of Computational Intelligence Systems, 17(1):196, 2024

    2024

  36. [36]

    Covid-19 prognosis via self-supervised representation learning and multi-image prediction.arXiv preprint arXiv:2101.04909, 2021

    Anuroop Sriram, Matthew Muckley, Koustuv Sinha, Farah Shamout, Joelle Pineau, Krzysztof J Geras, Lea Azour, Yindalon Aphinyanaphongs, Nafissa Yakubova, and William Moore. Covid-19 prognosis via self-supervised representation learning and multi-image prediction.arXiv preprint arXiv:2101.04909, 2021. 27

    work page arXiv 2021

  37. [37]

    Momentum contrastive learning for few-shot covid-19 diagnosis from chest ct images.Pattern recognition, 113:107826, 2021

    Xiaocong Chen, Lina Yao, Tao Zhou, Jinming Dong, and Yu Zhang. Momentum contrastive learning for few-shot covid-19 diagnosis from chest ct images.Pattern recognition, 113:107826, 2021

    2021

  38. [38]

    Chess: Chest x-ray pre-trained model via self-supervised contrastive learning.Journal of Digital Imaging, 36(3):902–910, 2023

    Kyungjin Cho, Ki Duk Kim, Yujin Nam, Jiheon Jeong, Jeeyoung Kim, Changyong Choi, Soyoung Lee, Jun Soo Lee, Seoyeon Woo, Gil-Sun Hong, et al. Chess: Chest x-ray pre-trained model via self-supervised contrastive learning.Journal of Digital Imaging, 36(3):902–910, 2023

    2023

  39. [39]

    Contrastive self-supervised learning for neurodegenerative disorder classification.Frontiers in Neuroinformatics, 19:1527582, 2025

    Vadym Gryshchuk, Devesh Singh, Stefan Teipel, Martin Dyrba, and FTLDNI study groups ADNI, AIBL. Contrastive self-supervised learning for neurodegenerative disorder classification.Frontiers in Neuroinformatics, 19:1527582, 2025

    2025

  40. [40]

    Parcl: Pathology-aware representation contrastive learning for glaucoma classification on fundus images

    Junyan Yi, Ying Zheng, Dayong Ding, Jianchun Zhao, and Gang Yang. Parcl: Pathology-aware representation contrastive learning for glaucoma classification on fundus images. In2024 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), pages 3912–3917. IEEE, 2024

    2024

  41. [41]

    Self-fi: self-supervised learning for disease diagnosis in fundus images.Bioengineering, 10(9):1089, 2023

    Toan Duc Nguyen, Duc-Tai Le, Junghyun Bum, Seongho Kim, Su Jeong Song, and Hyunseung Choo. Self-fi: self-supervised learning for disease diagnosis in fundus images.Bioengineering, 10(9):1089, 2023

    2023

  42. [42]

    Self-supervised learning for spinal mris

    Amir Jamaludin, Timor Kadir, and Andrew Zisserman. Self-supervised learning for spinal mris. InInternational Workshop on Deep Learning in Medical Image Analysis, pages 294–302. Springer, 2017

    2017

  43. [43]

    Self-supervised feature learning via exploiting multi-modal data for retinal disease diagnosis.IEEE Transactions on Medical Imaging, 39(12):4023– 4033, 2020

    Xiaomeng Li, Mengyu Jia, Md Tauhidul Islam, Lequan Yu, and Lei Xing. Self-supervised feature learning via exploiting multi-modal data for retinal disease diagnosis.IEEE Transactions on Medical Imaging, 39(12):4023– 4033, 2020

    2020

  44. [44]

    vox2vec: A framework for self-supervised contrastive learning of voxel-level representations in medical images

    Mikhail Goncharov, Vera Soboleva, Anvar Kurmukov, Maxim Pisov, and Mikhail Belyaev. vox2vec: A framework for self-supervised contrastive learning of voxel-level representations in medical images. InInternational Conference on Medical Image Computing and Computer-Assisted Intervention, pages 605–614. Springer, 2023

    2023

  45. [45]

    Multi-encoding contrastive learning for dual-stream self-supervised 3d dental segmentation network.Technologies, 13(9):419, 2025

    Tian Ma, Xiaoyuan Wei, Jiechen Zhai, Ziang Zhang, Yawen Li, and Yuancheng Li. Multi-encoding contrastive learning for dual-stream self-supervised 3d dental segmentation network.Technologies, 13(9):419, 2025

    2025

  46. [46]

    3d segmentation of necrotic lung lesions in ct images using self-supervised contrastive learning.IEEE Access, 12:32859–32869, 2024

    Yiqiao Liu, Sarah Halek, Randolph Crawford, Keith Persson, Michal Tomaszewski, Shubing Wang, Richard Baumgartner, Jianda Yuan, Gregory Goldmacher, and Antong Chen. 3d segmentation of necrotic lung lesions in ct images using self-supervised contrastive learning.IEEE Access, 12:32859–32869, 2024

    2024

  47. [47]

    Deep contrastive learning based tissue clustering for annotation-free histopathology image analysis.Computerized Medical Imaging and Graphics, 97:102053, 2022

    Jiangpeng Yan, Hanbo Chen, Xiu Li, and Jianhua Yao. Deep contrastive learning based tissue clustering for annotation-free histopathology image analysis.Computerized Medical Imaging and Graphics, 97:102053, 2022

    2022

  48. [48]

    Multi-modal representation learning in retinal imaging using self-supervised learning for enhanced clinical predictions.Scientific Reports, 14(1):26802, 2024

    Emese Suekei, Elisabeth Rumetshofer, Niklas Schmidinger, Andreas Mayr, Ursula Schmidt-Erfurth, Guenter Klambauer, and Hrvoje Bogunovi´c. Multi-modal representation learning in retinal imaging using self-supervised learning for enhanced clinical predictions.Scientific Reports, 14(1):26802, 2024

    2024

  49. [49]

    Exploring the utility of self- supervised pretraining strategies for the detection of absent lung sliding in m-mode lung ultrasound

    Blake VanBerlo, Brian Li, Alexander Wong, Jesse Hoey, and Robert Arntfield. Exploring the utility of self- supervised pretraining strategies for the detection of absent lung sliding in m-mode lung ultrasound. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3077–3086, 2023

    2023

  50. [50]

    Duy MH Nguyen, Hoang Nguyen, Nghiem Diep, Tan Ngoc Pham, Tri Cao, Binh Nguyen, Paul Swoboda, Nhat Ho, Shadi Albarqouni, Pengtao Xie, et al. Lvm-med: Learning large-scale self-supervised vision models for medical imaging via second-order graph matching.Advances in Neural Information Processing Systems, 36:27922–27950, 2023

    2023

  51. [51]

    Cl-mri: Self-supervised contrastive learning to improve the accuracy of undersampled mri reconstruction.Biomedical Signal Processing and Control, 100:107185, 2025

    Mevan Ekanayake, Zhifeng Chen, Mehrtash Harandi, Gary Egan, and Zhaolin Chen. Cl-mri: Self-supervised contrastive learning to improve the accuracy of undersampled mri reconstruction.Biomedical Signal Processing and Control, 100:107185, 2025

    2025

  52. [52]

    Artificial intelligence for detection of external cervical resorption using label-efficient self-supervised learning method.Journal of Endodontics, 50(2):144–153, 2024

    Hossein Mohammad-Rahimi, Omid Dianat, Reza Abbasi, Samira Zahedrozegar, Ali Ashkan, Saeed Reza Motamedian, Mohammad Hossein Rohban, and Ali Nosrat. Artificial intelligence for detection of external cervical resorption using label-efficient self-supervised learning method.Journal of Endodontics, 50(2):144–153, 2024

    2024

  53. [53]

    Self-supervised learning enhances periapical films segmentation with limited labeled data.Journal of Dentistry, page 106150, 2025

    Meiyu Hu, Qianli Zhang, Zhenyang Wei, Pingyi Jia, Mu Yuan, Huajie Yu, Xu-Cheng Yin, and Junran Peng. Self-supervised learning enhances periapical films segmentation with limited labeled data.Journal of Dentistry, page 106150, 2025

    2025

  54. [54]

    Pgl: Prior-guided local self-supervised learning for 3d medical image segmentation.arXiv preprint arXiv:2011.12640, 2020

    Yutong Xie, Jianpeng Zhang, Zehui Liao, Yong Xia, and Chunhua Shen. Pgl: Prior-guided local self-supervised learning for 3d medical image segmentation.arXiv preprint arXiv:2011.12640, 2020

    work page arXiv 2011

  55. [55]

    Self-supervised learning for endoscopic video analysis

    Roy Hirsch, Mathilde Caron, Regev Cohen, Amir Livne, Ron Shapiro, Tomer Golany, Roman Goldenberg, Daniel Freedman, and Ehud Rivlin. Self-supervised learning for endoscopic video analysis. InInternational conference on medical image Computing and computer-assisted intervention, pages 569–578. Springer, 2023. 28

    2023

  56. [56]

    Vim4path: Self-supervised vision mamba for histopathology images

    Ali Nasiri-Sarvi, Vincent Quoc-Huy Trinh, Hassan Rivaz, and Mahdi S Hosseini. Vim4path: Self-supervised vision mamba for histopathology images. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 6894–6903, 2024

    2024

  57. [57]

    Cads: A self-supervised learner via cross-modal alignment and deep self-distillation for ct volume segmentation.IEEE Transactions on Medical Imaging, 44(1):118–129, 2024

    Yiwen Ye, Jianpeng Zhang, Ziyang Chen, and Yong Xia. Cads: A self-supervised learner via cross-modal alignment and deep self-distillation for ct volume segmentation.IEEE Transactions on Medical Imaging, 44(1):118–129, 2024

    2024

  58. [58]

    Self-supervised learning methods for label-efficient dental caries classification.Diagnostics, 12(5):1237, 2022

    Aiham Taleb, Csaba Rohrer, Benjamin Bergner, Guilherme De Leon, Jonas Almeida Rodrigues, Falk Schwen- dicke, Christoph Lippert, and Joachim Krois. Self-supervised learning methods for label-efficient dental caries classification.Diagnostics, 12(5):1237, 2022

    2022

  59. [59]

    Swssl: Sliding window-based self-supervised learning for anomaly detection in high-resolution images.IEEE Transactions on Medical Imaging, 42(12):3860–3870, 2023

    Haoyu Dong, Yifan Zhang, Hanxue Gu, Nicholas Konz, Yixin Zhang, and Maciej A Mazurowski. Swssl: Sliding window-based self-supervised learning for anomaly detection in high-resolution images.IEEE Transactions on Medical Imaging, 42(12):3860–3870, 2023

    2023

  60. [60]

    Self-supervised feature learning for 3d medical images by playing a rubik’s cube

    Xinrui Zhuang, Yuexiang Li, Yifan Hu, Kai Ma, Yujiu Yang, and Yefeng Zheng. Self-supervised feature learning for 3d medical images by playing a rubik’s cube. InInternational conference on medical image computing and computer-assisted intervention, pages 420–428. Springer, 2019

    2019

  61. [61]

    Revisiting rubik’s cube: Self-supervised learning with volume-wise transformation for 3d medical image segmentation

    Xing Tao, Yuexiang Li, Wenhui Zhou, Kai Ma, and Yefeng Zheng. Revisiting rubik’s cube: Self-supervised learning with volume-wise transformation for 3d medical image segmentation. InInternational conference on medical image computing and computer-assisted intervention, pages 238–248. Springer, 2020

    2020

  62. [62]

    Self supervised deep representation learning for fine-grained body part recognition

    Pengyue Zhang, Fusheng Wang, and Yefeng Zheng. Self supervised deep representation learning for fine-grained body part recognition. In2017 IEEE 14th international symposium on biomedical imaging (ISBI 2017), pages 578–582. IEEE, 2017

    2017

  63. [63]

    Self-supervised learning based on spatial awareness for medical image analysis.IEEE access, 8:162973–162981, 2020

    Xuan-Bac Nguyen, Guee Sang Lee, Soo Hyung Kim, and Hyung Jeong Yang. Self-supervised learning based on spatial awareness for medical image analysis.IEEE access, 8:162973–162981, 2020

    2020

  64. [64]

    Self-supervised learning for cardiac mr image segmentation by anatomical position prediction

    Wenjia Bai, Chen Chen, Giacomo Tarroni, Jinming Duan, Florian Guitton, Steffen E Petersen, Yike Guo, Paul M Matthews, and Daniel Rueckert. Self-supervised learning for cardiac mr image segmentation by anatomical position prediction. InInternational conference on medical image computing and computer-assisted intervention, pages 541–549. Springer, 2019

    2019

  65. [65]

    Improving cytoarchitec- tonic segmentation of human brain areas with self-supervised siamese networks

    Hannah Spitzer, Kai Kiwitz, Katrin Amunts, Stefan Harmeling, and Timo Dickscheid. Improving cytoarchitec- tonic segmentation of human brain areas with self-supervised siamese networks. InInternational conference on medical image computing and computer-assisted intervention, pages 663–671. Springer, 2018

    2018

  66. [66]

    Med-seam: medical context aware self-supervised learning framework for anomaly classification in knee mri

    Akshay Daydar, Ajay Kumar Reddy, Sonal Kumar, Arijit Sur, and Hanif Laskar. Med-seam: medical context aware self-supervised learning framework for anomaly classification in knee mri. InProceedings of the Fifteenth Indian Conference on Computer Vision Graphics and Image Processing, pages 1–8, 2024

    2024

  67. [67]

    Self-supervised learning for stroke lesion segmentation on ct: a new pretext task for neuroimaging

    Juliette Moreau, Laura Mechtouff, David Rousseau, Tae-Hee Cho, Omer Eker, Yves Berthezéne, and Carole Frindel. Self-supervised learning for stroke lesion segmentation on ct: a new pretext task for neuroimaging. In 2025 47th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), pages 1–6. IEEE, 2025

    2025

  68. [68]

    Self-supervised representation learning for ultrasound video

    Jianbo Jiao, Richard Droste, Lior Drukker, Aris T Papageorghiou, and J Alison Noble. Self-supervised representation learning for ultrasound video. In2020 IEEE 17th international symposium on biomedical imaging (ISBI), pages 1847–1850. IEEE, 2020

    2020

  69. [69]

    Self-supervised representation learning for knee injury diagnosis from magnetic resonance data.IEEE Transactions on Artificial Intelligence, 5(4):1613–1623, 2023

    Siladittya Manna, Saumik Bhattacharya, and Umapada Pal. Self-supervised representation learning for knee injury diagnosis from magnetic resonance data.IEEE Transactions on Artificial Intelligence, 5(4):1613–1623, 2023

    2023

  70. [70]

    Multimodal self-supervised learning for medical image analysis

    Aiham Taleb, Christoph Lippert, Tassilo Klein, and Moin Nabi. Multimodal self-supervised learning for medical image analysis. InInternational conference on information processing in medical imaging, pages 661–673. Springer, 2021

    2021

  71. [71]

    Self-supervised learning application on covid-19 chest x-ray image classification using masked autoencoder.Bioengineering, 10(8):901, 2023

    Xin Xing, Gongbo Liang, Chris Wang, Nathan Jacobs, and Ai-Ling Lin. Self-supervised learning application on covid-19 chest x-ray image classification using masked autoencoder.Bioengineering, 10(8):901, 2023

    2023

  72. [72]

    Self-supervised learning with masked image modeling for teeth numbering, detection of dental restorations, and instance segmentation in dental panoramic radiographs

    Amani Almalki and Longin Jan Latecki. Self-supervised learning with masked image modeling for teeth numbering, detection of dental restorations, and instance segmentation in dental panoramic radiographs. In Proceedings of the IEEE/CVF winter conference on applications of computer vision, pages 5594–5603, 2023

    2023

  73. [73]

    Pyramid-based self-supervised learning for histopathological image classification.Computers in Biology and Medicine, 165:107336, 2023

    Junjie Wang, Hao Quan, Chengguang Wang, and Genke Yang. Pyramid-based self-supervised learning for histopathological image classification.Computers in Biology and Medicine, 165:107336, 2023. 29

    2023

  74. [74]

    Daniel Wolf, Tristan Payer, Catharina Silvia Lisson, Christoph Gerhard Lisson, Meinrad Beer, Michael Götz, and Timo Ropinski. Self-supervised pre-training with contrastive and masked autoencoder methods for dealing with small datasets in deep learning for medical imaging.Scientific Reports, 13(1):20260, 2023

    2023

  75. [75]

    Self-supervised learning with masked autoencoders for teeth segmenta- tion from intra-oral 3d scans

    Amani Almalki and Longin Jan Latecki. Self-supervised learning with masked autoencoders for teeth segmenta- tion from intra-oral 3d scans. InProceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 7820–7830, 2024

    2024

  76. [76]

    Kidney tumor classification on ct images using self-supervised learning.Computers in Biology and Medicine, 176:108554, 2024

    Erdal Özbay, Feyza Altunbey Özbay, and Farhad Soleimanian Gharehchopogh. Kidney tumor classification on ct images using self-supervised learning.Computers in Biology and Medicine, 176:108554, 2024

    2024

  77. [77]

    Models genesis: Generic autodidactic models for 3d medical image analysis

    Zongwei Zhou, Vatsal Sodha, Md Mahfuzur Rahman Siddiquee, Ruibin Feng, Nima Tajbakhsh, Michael B Gotway, and Jianming Liang. Models genesis: Generic autodidactic models for 3d medical image analysis. InInternational conference on medical image computing and computer-assisted intervention, pages 384–393. Springer, 2019

    2019

  78. [78]

    Yi-Kuan Liu, Jorge Cisneros, Girish Nair, Craig Stevens, Richard Castillo, Yevgeniy Vinogradskiy, and Edward Castillo. Perfusion estimation from dynamic non-contrast computed tomography using self-supervised learning and a physics-inspired u-net transformer architecture.International Journal of Computer Assisted Radiology and Surgery, 20(5):959–970, 2025

    2025

  79. [79]

    Self- supervised learning for medical image analysis using image context restoration.Medical image analysis, 58:101539, 2019

    Liang Chen, Paul Bentley, Kensaku Mori, Kazunari Misawa, Michitaka Fujiwara, and Daniel Rueckert. Self- supervised learning for medical image analysis using image context restoration.Medical image analysis, 58:101539, 2019

    2019

  80. [80]

    Improving data-efficiency and robustness of medical imaging segmentation using inpainting-based self-supervised learning.Bioengineering, 10(2):207, 2023

    Jeffrey Dominic, Nandita Bhaskhar, Arjun D Desai, Andrew Schmidt, Elka Rubin, Beliz Gunel, Garry E Gold, Brian A Hargreaves, Leon Lenchik, Robert Boutin, et al. Improving data-efficiency and robustness of medical imaging segmentation using inpainting-based self-supervised learning.Bioengineering, 10(2):207, 2023

    2023

Showing first 80 references.