Distribution-aware Fairness Learning in Medical Image Segmentation From A Control-Theoretic Perspective

Yujin Oh, Pengfei Jin, Sangjoon Park, Sekeun Kim, Siyeop Yoon, Jin Sung Kim, Kyungsang Kim, Xiang Li, Quanzheng Li
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:47003-47016, 2025.

Abstract

Ensuring fairness in medical image segmentation is critical due to biases in imbalanced clinical data acquisition caused by demographic attributes (e.g., age, sex, race) and clinical factors (e.g., disease severity). To address these challenges, we introduce Distribution-aware Mixture of Experts (dMoE), inspired by optimal control theory. We provide a comprehensive analysis of its underlying mechanisms and clarify dMoE’s role in adapting to heterogeneous distributions in medical image segmentation. Furthermore, we integrate dMoE into multiple network architectures, demonstrating its broad applicability across diverse medical image analysis tasks. By incorporating demographic and clinical factors, dMoE achieves state-of-the-art performance on two 2D benchmark datasets and a 3D in-house dataset. Our results highlight the effectiveness of dMoE in mitigating biases from imbalanced distributions, offering a promising approach to bridging control theory and medical image segmentation within fairness learning paradigms. The source code is available at https://github.com/tvseg/dMoE.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-oh25c, title = {Distribution-aware Fairness Learning in Medical Image Segmentation From A Control-Theoretic Perspective}, author = {Oh, Yujin and Jin, Pengfei and Park, Sangjoon and Kim, Sekeun and Yoon, Siyeop and Kim, Jin Sung and Kim, Kyungsang and Li, Xiang and Li, Quanzheng}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {47003--47016}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/oh25c/oh25c.pdf}, url = {https://proceedings.mlr.press/v267/oh25c.html}, abstract = {Ensuring fairness in medical image segmentation is critical due to biases in imbalanced clinical data acquisition caused by demographic attributes (e.g., age, sex, race) and clinical factors (e.g., disease severity). To address these challenges, we introduce Distribution-aware Mixture of Experts (dMoE), inspired by optimal control theory. We provide a comprehensive analysis of its underlying mechanisms and clarify dMoE’s role in adapting to heterogeneous distributions in medical image segmentation. Furthermore, we integrate dMoE into multiple network architectures, demonstrating its broad applicability across diverse medical image analysis tasks. By incorporating demographic and clinical factors, dMoE achieves state-of-the-art performance on two 2D benchmark datasets and a 3D in-house dataset. Our results highlight the effectiveness of dMoE in mitigating biases from imbalanced distributions, offering a promising approach to bridging control theory and medical image segmentation within fairness learning paradigms. The source code is available at https://github.com/tvseg/dMoE.} }
Endnote
%0 Conference Paper %T Distribution-aware Fairness Learning in Medical Image Segmentation From A Control-Theoretic Perspective %A Yujin Oh %A Pengfei Jin %A Sangjoon Park %A Sekeun Kim %A Siyeop Yoon %A Jin Sung Kim %A Kyungsang Kim %A Xiang Li %A Quanzheng Li %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-oh25c %I PMLR %P 47003--47016 %U https://proceedings.mlr.press/v267/oh25c.html %V 267 %X Ensuring fairness in medical image segmentation is critical due to biases in imbalanced clinical data acquisition caused by demographic attributes (e.g., age, sex, race) and clinical factors (e.g., disease severity). To address these challenges, we introduce Distribution-aware Mixture of Experts (dMoE), inspired by optimal control theory. We provide a comprehensive analysis of its underlying mechanisms and clarify dMoE’s role in adapting to heterogeneous distributions in medical image segmentation. Furthermore, we integrate dMoE into multiple network architectures, demonstrating its broad applicability across diverse medical image analysis tasks. By incorporating demographic and clinical factors, dMoE achieves state-of-the-art performance on two 2D benchmark datasets and a 3D in-house dataset. Our results highlight the effectiveness of dMoE in mitigating biases from imbalanced distributions, offering a promising approach to bridging control theory and medical image segmentation within fairness learning paradigms. The source code is available at https://github.com/tvseg/dMoE.
APA
Oh, Y., Jin, P., Park, S., Kim, S., Yoon, S., Kim, J.S., Kim, K., Li, X. & Li, Q.. (2025). Distribution-aware Fairness Learning in Medical Image Segmentation From A Control-Theoretic Perspective. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:47003-47016 Available from https://proceedings.mlr.press/v267/oh25c.html.

Related Material