Semi-Supervised Skin Lesion Segmentation under Dual Mask Ensemble with Feature Discrepancy Co-Training

Thanh-Huy Nguyen, Thien Nguyen, Xuan Bach Nguyen, Nguyen Lan Vi Vu, Vinh Quang Dinh, Fabrice Meriaudeau
Proceedings of The 8th International Conference on Medical Imaging with Deep Learning, PMLR 301:1213-1226, 2026.

Abstract

Skin Lesion Segmentation with supportive Deep Learning has become essential in skin lesion analysis and skin cancer diagnosis. However, in the practical scenario of clinical implementation, there is a limitation in human-annotated labels for training data, which leads to poor performance in supervised training models. In this paper, we propose Dual Mask Ensemble (DME) based on a dual-branch co-training network, which aims to enforce two models to exploit information from different views. Specifically, we introduce a novel feature discrepancy loss trained with a cross-pseudo supervision strategy, which enhances model representation by encouraging the sub-networks to learn from distinct features, thereby mitigating feature collapse. Additionally, Dual Mask Ensemble training enables the sub-models to extract more meaningful information from unlabeled data by combining mask predictions. Experimental results demonstrate the effectiveness of our approach, achieving state-of-the-art performance across several metrics (Dice and Jaccard) on the ISIC2018 and HAM10000 datasets. Our code is available at https://github.com/antares0811/DME-FD.

Cite this Paper

BibTeX
@InProceedings{pmlr-v301-nguyen26a, title = {Semi-Supervised Skin Lesion Segmentation under Dual Mask Ensemble with Feature Discrepancy Co-Training}, author = {Nguyen, Thanh-Huy and Nguyen, Thien and Nguyen, Xuan Bach and Vu, Nguyen Lan Vi and Dinh, Vinh Quang and Meriaudeau, Fabrice}, booktitle = {Proceedings of The 8th International Conference on Medical Imaging with Deep Learning}, pages = {1213--1226}, 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/nguyen26a/nguyen26a.pdf}, url = {https://proceedings.mlr.press/v301/nguyen26a.html}, abstract = {Skin Lesion Segmentation with supportive Deep Learning has become essential in skin lesion analysis and skin cancer diagnosis. However, in the practical scenario of clinical implementation, there is a limitation in human-annotated labels for training data, which leads to poor performance in supervised training models. In this paper, we propose Dual Mask Ensemble (DME) based on a dual-branch co-training network, which aims to enforce two models to exploit information from different views. Specifically, we introduce a novel feature discrepancy loss trained with a cross-pseudo supervision strategy, which enhances model representation by encouraging the sub-networks to learn from distinct features, thereby mitigating feature collapse. Additionally, Dual Mask Ensemble training enables the sub-models to extract more meaningful information from unlabeled data by combining mask predictions. Experimental results demonstrate the effectiveness of our approach, achieving state-of-the-art performance across several metrics (Dice and Jaccard) on the ISIC2018 and HAM10000 datasets. Our code is available at https://github.com/antares0811/DME-FD.} }
Endnote
%0 Conference Paper %T Semi-Supervised Skin Lesion Segmentation under Dual Mask Ensemble with Feature Discrepancy Co-Training %A Thanh-Huy Nguyen %A Thien Nguyen %A Xuan Bach Nguyen %A Nguyen Lan Vi Vu %A Vinh Quang Dinh %A Fabrice Meriaudeau %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-nguyen26a %I PMLR %P 1213--1226 %U https://proceedings.mlr.press/v301/nguyen26a.html %V 301 %X Skin Lesion Segmentation with supportive Deep Learning has become essential in skin lesion analysis and skin cancer diagnosis. However, in the practical scenario of clinical implementation, there is a limitation in human-annotated labels for training data, which leads to poor performance in supervised training models. In this paper, we propose Dual Mask Ensemble (DME) based on a dual-branch co-training network, which aims to enforce two models to exploit information from different views. Specifically, we introduce a novel feature discrepancy loss trained with a cross-pseudo supervision strategy, which enhances model representation by encouraging the sub-networks to learn from distinct features, thereby mitigating feature collapse. Additionally, Dual Mask Ensemble training enables the sub-models to extract more meaningful information from unlabeled data by combining mask predictions. Experimental results demonstrate the effectiveness of our approach, achieving state-of-the-art performance across several metrics (Dice and Jaccard) on the ISIC2018 and HAM10000 datasets. Our code is available at https://github.com/antares0811/DME-FD.
APA
Nguyen, T., Nguyen, T., Nguyen, X.B., Vu, N.L.V., Dinh, V.Q. & Meriaudeau, F.. (2026). Semi-Supervised Skin Lesion Segmentation under Dual Mask Ensemble with Feature Discrepancy Co-Training. Proceedings of The 8th International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 301:1213-1226 Available from https://proceedings.mlr.press/v301/nguyen26a.html.

Related Material