Weakly supervised segmentation with point annotations for histopathology images via contrast-based variational model

Zhang, Hongrun and Burrows, Liam and Meng, Yanda and Sculthorpe, Declan and Mukherjee, Abhik and Coupland, Sarah E and Chen, Ke and Zheng, Yalin; (2023) Weakly supervised segmentation with point annotations for histopathology images via contrast-based variational model. In: 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition . IEEE, CAN, pp. 15630-15640. ISBN 9798350301298 (https://doi.org/10.1109/CVPR52729.2023.01500)

[thumbnail of Zhang-etal-IEEE-CVPR-2023-Weakly-supervised-segmentation-with-point-annotations-for-histopathology-images]
Preview
 Text. Filename: Zhang-etal-IEEE-CVPR-2023-Weakly-supervised-segmentation-with-point-annotations-for-histopathology-images.pdf
Accepted Author Manuscript
License: Strathprints license 1.0
Download (3MB)| Preview

Abstract

Image segmentation is a fundamental task in the field of imaging and vision. Supervised deep learning for segmentation has achieved unparalleled success when sufficient training data with annotated labels are available. However, annotation is known to be expensive to obtain, especially for histopathology images where the target regions are usually with high morphology variations and irregular shapes. Thus, weakly supervised learning with sparse annotations of points is promising to reduce the annotation workload. In this work, we propose a contrast-based variational model to generate segmentation results, which serve as reliable complementary supervision to train a deep segmentation model for histopathology images. The proposed method considers the common characteristics of target regions in histopathology images and can be trained in an end-to-end manner. It can generate more regionally consistent and smoother boundary segmentation, and is more robust to unlabeled 'novel' regions. Experiments on two different histology datasets demonstrate its effectiveness and efficiency in comparison to previous models. Code is available at: https://github.com/hrzhang1123/CVM-WS-Segmentation.

ORCID iDs

Zhang, Hongrun, Burrows, Liam, Meng, Yanda, Sculthorpe, Declan, Mukherjee, Abhik, Coupland, Sarah E, Chen, Ke ORCID logoORCID: https://orcid.org/0000-0002-6093-6623 and Zheng, Yalin;
CORE (COnnecting REpositories)