Dealing With Strong Imbalance to Classifying Glomerular Lesions via Self-Supervised Learning
Resumo
The histopathological analysis of renal glomeruli is fundamental for diagnosing kidney diseases, but it is hampered by challenges such as inter-observer variability, the high cost of expert annotation, and the inherent complexity of renal lesions, which often coexist. Self-Supervised Learning (SSL) offers a promising strategy to mitigate the reliance on large-scale annotated datasets. This work-in-progress paper presents a systematic comparison of SSL methods for classifying glomerular lesions on a challenging real-world dataset of 11,928 histopathology images. Our preliminary results demonstrate that the DINO method, significantly outperforms supervised baselines and other SSL techniques. An interesting finding from our data analysis is that several images depict glomeruli with multiple, co-occurring lesions, meaning that only normal glomeruli consistently represent a single, isolated class. This suggests that a standard multiclass classification approach may be suboptimal. Based on this finding, we propose the development of a novel two-stage, binarymultilabel pipeline as the main direction for our future work. This pipeline will first distinguish between normal and Lesion images and then apply a multi-label model to identify all specific lesions present, better reflecting clinical reality.Referências
M. Haas et al., “Consensus definitions for glomerular lesions by light and electron microscopy: recommendations from a working group of the renal pathology society,” Kidney international, vol. 98, no. 5, pp. 1120–1134, 2020.
A. Tiard et al., “Stain-invariant self supervised learning for histopathology image analysis,” arXiv preprint arXiv:2211.07590, 2022. [Online]. DOI: 10.48550/arXiv.2211.07590
D. Tellez et al., “Whole-slide mitosis detection in h&e breast histology using phh3 as a reference to train distilled stain-invariant convolutional networks,” IEEE Transactions on Medical Imaging, vol. 38, no. 2, pp. 636–647, 2019.
M. Gadermayr et al., “Stain-adaptive self-supervised learning for histopathology image analysis,” Medical Image Analysis, vol. 77, p. 102385, 2022.
N. Tajbakhsh, L. Jeyaseelan, Q. Li, J. N. Chiang, Z. Wu, and X. Ding, “Embracing imperfect datasets: A review of deep learning solutions for medical image segmentation,” Medical image analysis, vol. 63, p. 101693, 2020.
W. He, T. Liu, Y. Han, W. Ming, J. Du, Y. Liu, Y. Yang, L. Wang, Z. Jiang, Y. Wang et al., “A review: The detection of cancer cells in histopathology based on machine vision,” Computers in Biology and Medicine, vol. 146, p. 105636, 2022.
T. Chen, S. Kornblith, M. Norouzi, and G. Hinton, “A simple framework for contrastive learning of visual representations,” in International conference on machine learning. PMLR, 2020, pp. 1597–1607.
J.-B. Grill, F. Strub, F. Altché, C. Tallec, P. Richemond, E. Buchatskaya, C. Doersch, B. Avila Pires, Z. Guo, M. Gheshlaghi Azar et al., “Bootstrap your own latent-a new approach to self-supervised learning,” Advances in neural information processing systems, vol. 33, pp. 21 271–21 284, 2020.
M. Oquab, T. Darcet, T. Moutakanni, H. Vo, M. Szafraniec, V. Khalidov, P. Fernandez, D. Haziza, F. Massa, A. El-Nouby et al., “Dinov2: Learning robust visual features without supervision,” arXiv preprint arXiv:2304.07193, 2023.
A. Taleb, W. Loetzsch, N. Danz, J. Severin, T. Gaertner, B. Bergner, and C. Lippert, “3d self-supervised methods for medical imaging,” Advances in neural information processing systems, vol. 33, pp. 18 158–18 172, 2020.
S.-C. Huang, A. Pareek, M. Jensen, M. P. Lungren, S. Yeung, and A. S. Chaudhari, “Self-supervised learning for medical image classification: a systematic review and implementation guidelines,” NPJ Digital Medicine, vol. 6, no. 1, p. 74, 2023.
J. Wang, H. Quan, C. Wang, and G. Yang, “Pyramid-based self-supervised learning for histopathological image classification,” Computers in Biology and Medicine, vol. 165, p. 107336, 2023.
J. Snell, K. Swersky, and R. S. Zemel, “Prototypical networks for few-shot learning,” in Advances in Neural Information Processing Systems (NeurIPS), vol. 30, 2017. [Online]. Available: [link]
C. Finn, P. Abbeel, and S. Levine, “Model-agnostic meta-learning for fast adaptation of deep networks,” in Proceedings of the 34th International Conference on Machine Learning (ICML), vol. 70. PMLR, 2017, pp. 1126–1135. [Online]. Available: [link]
E. Pachetti and S. Colantonio, “A systematic review of few-shot learning in medical imaging,” Artificial Intelligence in Medicine (preprint / arXiv), 2023.
N. Schiavone, J. Wang, S. Li, R. Zemp, and X. Li, “Myriadal: Active few shot learning for histopathology,” in Proceedings of IEEE Conference on Artificial Intelligence (or relevant workshop), 2024, combines contrastive encoder and pseudo-label active few-shot learning in histology.
K. Stacke et al., “A closer look at domain shift for deep learning in histopathology,” Medical Image Analysis, vol. 73, p. 102152, 2021.
Y. Jin et al., “Histossl: Unleashing the power of self-supervised learning for histopathology,” IEEE Transactions on Medical Imaging, vol. 41, no. 12, pp. 3652–3663, 2022.
O. Ciga, T. Xu, and A. L. Martel, “Self supervised contrastive learning for digital histopathology,” Machine Learning with Applications, vol. 7, p. 100198, 2022.
—, “Addressing class imbalance in renal amyloidosis classification: A comparative study of few-shot learning and conventional machine learning,” in IMPROVE 2025 – Proceedings of the 5th International Conference on Image Processing and Vision Engineering.
Z. Chen, J. Ge, H. Zhan, S. Huang, and D. Wang, “Pareto self-supervised training for few-shot learning,” arXiv preprint, 2021, arXiv:2104.07841.
Y. Lu, L. Wen, J. Liu, Y. Liu, and X. Tian, “Self-supervision can be a good few-shot learner,” arXiv preprint, 2022, arXiv:2207.09176.
M. Caron, H. Touvron, I. Misra, H. Jégou, J. Mairal, P. Bojanowski, and A. Joulin, “Emerging properties in self-supervised vision transformers,” in IEEE/CVF International Conference on Computer Vision (ICCV), 2021, pp. 9650–9660. [Online]. Available: [link]
A. Tiard et al., “Stain-invariant self supervised learning for histopathology image analysis,” arXiv preprint arXiv:2211.07590, 2022. [Online]. DOI: 10.48550/arXiv.2211.07590
D. Tellez et al., “Whole-slide mitosis detection in h&e breast histology using phh3 as a reference to train distilled stain-invariant convolutional networks,” IEEE Transactions on Medical Imaging, vol. 38, no. 2, pp. 636–647, 2019.
M. Gadermayr et al., “Stain-adaptive self-supervised learning for histopathology image analysis,” Medical Image Analysis, vol. 77, p. 102385, 2022.
N. Tajbakhsh, L. Jeyaseelan, Q. Li, J. N. Chiang, Z. Wu, and X. Ding, “Embracing imperfect datasets: A review of deep learning solutions for medical image segmentation,” Medical image analysis, vol. 63, p. 101693, 2020.
W. He, T. Liu, Y. Han, W. Ming, J. Du, Y. Liu, Y. Yang, L. Wang, Z. Jiang, Y. Wang et al., “A review: The detection of cancer cells in histopathology based on machine vision,” Computers in Biology and Medicine, vol. 146, p. 105636, 2022.
T. Chen, S. Kornblith, M. Norouzi, and G. Hinton, “A simple framework for contrastive learning of visual representations,” in International conference on machine learning. PMLR, 2020, pp. 1597–1607.
J.-B. Grill, F. Strub, F. Altché, C. Tallec, P. Richemond, E. Buchatskaya, C. Doersch, B. Avila Pires, Z. Guo, M. Gheshlaghi Azar et al., “Bootstrap your own latent-a new approach to self-supervised learning,” Advances in neural information processing systems, vol. 33, pp. 21 271–21 284, 2020.
M. Oquab, T. Darcet, T. Moutakanni, H. Vo, M. Szafraniec, V. Khalidov, P. Fernandez, D. Haziza, F. Massa, A. El-Nouby et al., “Dinov2: Learning robust visual features without supervision,” arXiv preprint arXiv:2304.07193, 2023.
A. Taleb, W. Loetzsch, N. Danz, J. Severin, T. Gaertner, B. Bergner, and C. Lippert, “3d self-supervised methods for medical imaging,” Advances in neural information processing systems, vol. 33, pp. 18 158–18 172, 2020.
S.-C. Huang, A. Pareek, M. Jensen, M. P. Lungren, S. Yeung, and A. S. Chaudhari, “Self-supervised learning for medical image classification: a systematic review and implementation guidelines,” NPJ Digital Medicine, vol. 6, no. 1, p. 74, 2023.
J. Wang, H. Quan, C. Wang, and G. Yang, “Pyramid-based self-supervised learning for histopathological image classification,” Computers in Biology and Medicine, vol. 165, p. 107336, 2023.
J. Snell, K. Swersky, and R. S. Zemel, “Prototypical networks for few-shot learning,” in Advances in Neural Information Processing Systems (NeurIPS), vol. 30, 2017. [Online]. Available: [link]
C. Finn, P. Abbeel, and S. Levine, “Model-agnostic meta-learning for fast adaptation of deep networks,” in Proceedings of the 34th International Conference on Machine Learning (ICML), vol. 70. PMLR, 2017, pp. 1126–1135. [Online]. Available: [link]
E. Pachetti and S. Colantonio, “A systematic review of few-shot learning in medical imaging,” Artificial Intelligence in Medicine (preprint / arXiv), 2023.
N. Schiavone, J. Wang, S. Li, R. Zemp, and X. Li, “Myriadal: Active few shot learning for histopathology,” in Proceedings of IEEE Conference on Artificial Intelligence (or relevant workshop), 2024, combines contrastive encoder and pseudo-label active few-shot learning in histology.
K. Stacke et al., “A closer look at domain shift for deep learning in histopathology,” Medical Image Analysis, vol. 73, p. 102152, 2021.
Y. Jin et al., “Histossl: Unleashing the power of self-supervised learning for histopathology,” IEEE Transactions on Medical Imaging, vol. 41, no. 12, pp. 3652–3663, 2022.
O. Ciga, T. Xu, and A. L. Martel, “Self supervised contrastive learning for digital histopathology,” Machine Learning with Applications, vol. 7, p. 100198, 2022.
—, “Addressing class imbalance in renal amyloidosis classification: A comparative study of few-shot learning and conventional machine learning,” in IMPROVE 2025 – Proceedings of the 5th International Conference on Image Processing and Vision Engineering.
Z. Chen, J. Ge, H. Zhan, S. Huang, and D. Wang, “Pareto self-supervised training for few-shot learning,” arXiv preprint, 2021, arXiv:2104.07841.
Y. Lu, L. Wen, J. Liu, Y. Liu, and X. Tian, “Self-supervision can be a good few-shot learner,” arXiv preprint, 2022, arXiv:2207.09176.
M. Caron, H. Touvron, I. Misra, H. Jégou, J. Mairal, P. Bojanowski, and A. Joulin, “Emerging properties in self-supervised vision transformers,” in IEEE/CVF International Conference on Computer Vision (ICCV), 2021, pp. 9650–9660. [Online]. Available: [link]
Publicado
30/09/2025
Como Citar
AZEVEDO, Gabriel de; SANTOS, Alexsandro Silva; OLIVEIRA, Luciano Rebouças de; SANTOS, Washington Luís Conrado dos; DUARTE, Angelo Amâncio.
Dealing With Strong Imbalance to Classifying Glomerular Lesions via Self-Supervised Learning. In: WORKSHOP DE TRABALHOS EM ANDAMENTO - CONFERENCE ON GRAPHICS, PATTERNS AND IMAGES (SIBGRAPI), 38. , 2025, Salvador/BA.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2025
.
p. 138-143.
