Weakly Supervised Histopathology Image Segmentation With Sparse Point Annotations
- 15 September 2020
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 25 (5) , 1673-1685
Abstract
Digital histopathology image segmentation can facilitate computer-assisted cancer diagnostics. Given the difficulty of obtaining manual annotations, weak supervision is more suitable for the task than full supervision is. However, most weakly supervised models are not ideal for handling severe intra-class heterogeneity and inter-class homogeneity in histopathology images. Therefore, we propose a novel end-to-end weakly supervised learning framework named WESUP. With only sparse point annotations, it performs accurate segmentation and exhibits good generalizability. The training phase comprises two major parts, hierarchical feature representation and deep dynamic label propagation. The former uses superpixels to capture local details and global context from the convolutional feature maps obtained via transfer learning. The latter recognizes the manifold structure of the hierarchical features and identifies potential targets with the sparse annotations. Moreover, these two parts are trained jointly to improve the performance of the whole framework. To further boost test performance, pixel-wise inference is adopted for finer prediction. As demonstrated by experimental results, WESUP is able to largely resolve the confusion between histological foreground and background. It outperforms several state-of-the-art weakly supervised methods on a variety of histopathology datasets with minimal annotation efforts. Trained by very sparse point annotations, WESUP can even beat an advanced fully supervised segmentation network.Keywords
Funding Information
- National Natural Science Foundation of China (61702094)
- Shanghai Municipal Education Commission (18CG38)
- Science and Technology Commission of Shanghai Municipality (17YF1427400)
This publication has 33 references indexed in Scilit:
- Co-Saliency Detection via a Self-Paced Multiple-Instance Learning FrameworkIEEE Transactions on Pattern Analysis and Machine Intelligence, 2016
- Multiscale Combinatorial Grouping for Image Segmentation and Object Proposal GenerationIEEE Transactions on Pattern Analysis and Machine Intelligence, 2016
- BoxSup: Exploiting Bounding Boxes to Supervise Convolutional Networks for Semantic Segmentation2015
- U-Net: Convolutional Networks for Biomedical Image Segmentation2015
- Feedforward semantic segmentation with zoom-out features2015
- Deep learningNature, 2015
- A Stochastic Polygons Model for Glandular Structures in Colon Histology ImagesIEEE Transactions on Medical Imaging, 2015
- Object Detection in Optical Remote Sensing Images Based on Weakly Supervised Learning and High-Level Feature LearningIEEE Transactions on Geoscience and Remote Sensing, 2014
- Weakly supervised histopathology cancer image segmentation and classificationMedical Image Analysis, 2014
- Dynamic Label Propagation for Semi-supervised Multi-class Multi-label Classification2013