Enhancing Contrastive Learning for Retinal Imaging via Adjusted Augmentation Scales

Zijie Cheng, Boxuan Li, Andre Altmann, Pearse Keane, Yukun Zhou
Proceedings of The 8th International Conference on Medical Imaging with Deep Learning, PMLR 301:214-225, 2026.

Abstract

Contrastive learning, a typical self-supervised learning strategy, operates on bringing similar data together while pushing dissimilar data apart in latent space. This approach extracts robust and discriminative representations, thus being widely used in natural computer vision tasks, such as object classification. However, unlike natural images, medical images (e.g., retinal images) tend to share substantial similarities in imaging area and anatomical tissues, leading to a denser distribution in latent space. As a result, the default use of strong augmentations in contrastive learning potentially exacerbates this intensive distribution in retinal images, making it difficult to distinguish between genuinely similar and dissimilar data, and therefore hindering model pre-training convergence. In this paper, we hypothesise that weaker augmentations are better suited to contrastive learning for medical image applications, and we investigate model performance under various augmentation strategies. Our study includes six publicly available retinal datasets covering multiple clinically relevant tasks. We assess the models\’{performance} and generalizability via extensive experiments. The model pre-trained with weak augmentation outperforms the one pre-trained with strong augmentation, achieving approximately a 6% increase in AUPR ($P$$<$0.001) and a 12.5% increase in sensitivity ($P$$<$0.001) on MESSIDOR-2. Similar improvements are observed across other datasets. Our findings suggest that optimizing the scale of augmentation is critical for enhancing the efficacy of contrastive learning in medical imaging. The model weights and relevant code are available at: https://github.com/ziijiecheng/Enhance-contrastive-SSL-for-Retinal-Imaging.

Cite this Paper

BibTeX
@InProceedings{pmlr-v301-cheng26a, title = {Enhancing Contrastive Learning for Retinal Imaging via Adjusted Augmentation Scales}, author = {Cheng, Zijie and Li, Boxuan and Altmann, Andre and Keane, Pearse and Zhou, Yukun}, booktitle = {Proceedings of The 8th International Conference on Medical Imaging with Deep Learning}, pages = {214--225}, year = {2026}, editor = {Tasdizen, Tolga and Elhabian, Shireen and Summers, Ronald and Chen, Chen and Koch, Lisa and Zhuang, Yan}, volume = {301}, series = {Proceedings of Machine Learning Research}, month = {09--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v301/main/assets/cheng26a/cheng26a.pdf}, url = {https://proceedings.mlr.press/v301/cheng26a.html}, abstract = {Contrastive learning, a typical self-supervised learning strategy, operates on bringing similar data together while pushing dissimilar data apart in latent space. This approach extracts robust and discriminative representations, thus being widely used in natural computer vision tasks, such as object classification. However, unlike natural images, medical images (e.g., retinal images) tend to share substantial similarities in imaging area and anatomical tissues, leading to a denser distribution in latent space. As a result, the default use of strong augmentations in contrastive learning potentially exacerbates this intensive distribution in retinal images, making it difficult to distinguish between genuinely similar and dissimilar data, and therefore hindering model pre-training convergence. In this paper, we hypothesise that weaker augmentations are better suited to contrastive learning for medical image applications, and we investigate model performance under various augmentation strategies. Our study includes six publicly available retinal datasets covering multiple clinically relevant tasks. We assess the models\’{performance} and generalizability via extensive experiments. The model pre-trained with weak augmentation outperforms the one pre-trained with strong augmentation, achieving approximately a 6% increase in AUPR ($P$$<$0.001) and a 12.5% increase in sensitivity ($P$$<$0.001) on MESSIDOR-2. Similar improvements are observed across other datasets. Our findings suggest that optimizing the scale of augmentation is critical for enhancing the efficacy of contrastive learning in medical imaging. The model weights and relevant code are available at: https://github.com/ziijiecheng/Enhance-contrastive-SSL-for-Retinal-Imaging.} }
Endnote
%0 Conference Paper %T Enhancing Contrastive Learning for Retinal Imaging via Adjusted Augmentation Scales %A Zijie Cheng %A Boxuan Li %A Andre Altmann %A Pearse Keane %A Yukun Zhou %B Proceedings of The 8th International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2026 %E Tolga Tasdizen %E Shireen Elhabian %E Ronald Summers %E Chen Chen %E Lisa Koch %E Yan Zhuang %F pmlr-v301-cheng26a %I PMLR %P 214--225 %U https://proceedings.mlr.press/v301/cheng26a.html %V 301 %X Contrastive learning, a typical self-supervised learning strategy, operates on bringing similar data together while pushing dissimilar data apart in latent space. This approach extracts robust and discriminative representations, thus being widely used in natural computer vision tasks, such as object classification. However, unlike natural images, medical images (e.g., retinal images) tend to share substantial similarities in imaging area and anatomical tissues, leading to a denser distribution in latent space. As a result, the default use of strong augmentations in contrastive learning potentially exacerbates this intensive distribution in retinal images, making it difficult to distinguish between genuinely similar and dissimilar data, and therefore hindering model pre-training convergence. In this paper, we hypothesise that weaker augmentations are better suited to contrastive learning for medical image applications, and we investigate model performance under various augmentation strategies. Our study includes six publicly available retinal datasets covering multiple clinically relevant tasks. We assess the models\’{performance} and generalizability via extensive experiments. The model pre-trained with weak augmentation outperforms the one pre-trained with strong augmentation, achieving approximately a 6% increase in AUPR ($P$$<$0.001) and a 12.5% increase in sensitivity ($P$$<$0.001) on MESSIDOR-2. Similar improvements are observed across other datasets. Our findings suggest that optimizing the scale of augmentation is critical for enhancing the efficacy of contrastive learning in medical imaging. The model weights and relevant code are available at: https://github.com/ziijiecheng/Enhance-contrastive-SSL-for-Retinal-Imaging.
APA
Cheng, Z., Li, B., Altmann, A., Keane, P. & Zhou, Y.. (2026). Enhancing Contrastive Learning for Retinal Imaging via Adjusted Augmentation Scales. Proceedings of The 8th International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 301:214-225 Available from https://proceedings.mlr.press/v301/cheng26a.html.

Related Material