Boundary-aware Contrastive Learning for Semi-supervised Nuclei Instance Segmentation

Ye Zhang, Ziyue Wang, Yifeng Wang, Hao Bian, Linghan Cai, Hengrui Li, Lingbo Zhang, Yongbing Zhang
Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning, PMLR 250:1851-1861, 2024.

Abstract

Semi-supervised segmentation methods have demonstrated promising results in natural scenarios, providing a solution to reduce dependency on manual annotation. However, these methods face significant challenges when directly applied to pathological images due to the subtle color differences between nuclei and tissues, as well as the significant morphological variations among nuclei. Consequently, the generated pseudo-labels often contain much noise, especially at the nuclei boundaries. To address the above problem, this paper proposes a boundary-aware contrastive learning network to denoise the boundary noise in a semi-supervised nuclei segmentation task. The model has two key designs: a low-resolution denoising (LRD) module and a cross-RoI contrastive learning (CRC) module. The LRD improves the smoothness of the nuclei boundary by pseudo-labels denoising, and the CRC enhances the discrimination between foreground and background by boundary feature contrastive learning. We conduct extensive experiments to demonstrate the superiority of our proposed method over existing semi-supervised instance segmentation methods.

Cite this Paper

BibTeX
@InProceedings{pmlr-v250-zhang24a, title = {Boundary-aware Contrastive Learning for Semi-supervised Nuclei Instance Segmentation}, author = {Zhang, Ye and Wang, Ziyue and Wang, Yifeng and Bian, Hao and Cai, Linghan and Li, Hengrui and Zhang, Lingbo and Zhang, Yongbing}, booktitle = {Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning}, pages = {1851--1861}, year = {2024}, editor = {Burgos, Ninon and Petitjean, Caroline and Vakalopoulou, Maria and Christodoulidis, Stergios and Coupe, Pierrick and Delingette, Hervé and Lartizien, Carole and Mateus, Diana}, volume = {250}, series = {Proceedings of Machine Learning Research}, month = {03--05 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v250/main/assets/zhang24a/zhang24a.pdf}, url = {https://proceedings.mlr.press/v250/zhang24a.html}, abstract = {Semi-supervised segmentation methods have demonstrated promising results in natural scenarios, providing a solution to reduce dependency on manual annotation. However, these methods face significant challenges when directly applied to pathological images due to the subtle color differences between nuclei and tissues, as well as the significant morphological variations among nuclei. Consequently, the generated pseudo-labels often contain much noise, especially at the nuclei boundaries. To address the above problem, this paper proposes a boundary-aware contrastive learning network to denoise the boundary noise in a semi-supervised nuclei segmentation task. The model has two key designs: a low-resolution denoising (LRD) module and a cross-RoI contrastive learning (CRC) module. The LRD improves the smoothness of the nuclei boundary by pseudo-labels denoising, and the CRC enhances the discrimination between foreground and background by boundary feature contrastive learning. We conduct extensive experiments to demonstrate the superiority of our proposed method over existing semi-supervised instance segmentation methods.} }
Endnote
%0 Conference Paper %T Boundary-aware Contrastive Learning for Semi-supervised Nuclei Instance Segmentation %A Ye Zhang %A Ziyue Wang %A Yifeng Wang %A Hao Bian %A Linghan Cai %A Hengrui Li %A Lingbo Zhang %A Yongbing Zhang %B Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2024 %E Ninon Burgos %E Caroline Petitjean %E Maria Vakalopoulou %E Stergios Christodoulidis %E Pierrick Coupe %E Hervé Delingette %E Carole Lartizien %E Diana Mateus %F pmlr-v250-zhang24a %I PMLR %P 1851--1861 %U https://proceedings.mlr.press/v250/zhang24a.html %V 250 %X Semi-supervised segmentation methods have demonstrated promising results in natural scenarios, providing a solution to reduce dependency on manual annotation. However, these methods face significant challenges when directly applied to pathological images due to the subtle color differences between nuclei and tissues, as well as the significant morphological variations among nuclei. Consequently, the generated pseudo-labels often contain much noise, especially at the nuclei boundaries. To address the above problem, this paper proposes a boundary-aware contrastive learning network to denoise the boundary noise in a semi-supervised nuclei segmentation task. The model has two key designs: a low-resolution denoising (LRD) module and a cross-RoI contrastive learning (CRC) module. The LRD improves the smoothness of the nuclei boundary by pseudo-labels denoising, and the CRC enhances the discrimination between foreground and background by boundary feature contrastive learning. We conduct extensive experiments to demonstrate the superiority of our proposed method over existing semi-supervised instance segmentation methods.
APA
Zhang, Y., Wang, Z., Wang, Y., Bian, H., Cai, L., Li, H., Zhang, L. & Zhang, Y.. (2024). Boundary-aware Contrastive Learning for Semi-supervised Nuclei Instance Segmentation. Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 250:1851-1861 Available from https://proceedings.mlr.press/v250/zhang24a.html.

Related Material