Pairwise Relation Learning for Semi-supervised Gland Segmentation

TL;DR

This paper introduces a semi-supervised gland segmentation model that leverages pairwise relations and shared encoders to improve segmentation accuracy on histology images, reducing the need for extensive labeled data.

Contribution

The proposed PRS^2 model combines a segmentation network with a pairwise relation network sharing encoders, enabling semi-supervised learning and improved gland segmentation performance.

Findings

Achieves state-of-the-art results on GlaS and CRAG datasets.

Effectively utilizes unlabeled data through pairwise relations.

Improves segmentation of touching glands with object-level Dice loss.

Abstract

Accurate and automated gland segmentation on histology tissue images is an essential but challenging task in the computer-aided diagnosis of adenocarcinoma. Despite their prevalence, deep learning models always require a myriad number of densely annotated training images, which are difficult to obtain due to extensive labor and associated expert costs related to histology image annotations. In this paper, we propose the pairwise relation-based semi-supervised (PRS^2) model for gland segmentation on histology images. This model consists of a segmentation network (S-Net) and a pairwise relation network (PR-Net). The S-Net is trained on labeled data for segmentation, and PR-Net is trained on both labeled and unlabeled data in an unsupervised way to enhance its image representation ability via exploiting the semantic consistency between each pair of images in the feature space. Since both networks share their encoders, the image representation ability learned by PR-Net can be transferred to S-Net to improve its segmentation performance. We also design the object-level Dice loss to address the issues caused by touching glands and combine it with other two loss functions for S-Net. We evaluated our model against five recent methods on the GlaS dataset and three recent methods on the CRAG dataset. Our results not only demonstrate the effectiveness of the proposed PR-Net and object-level Dice loss, but also indicate that our PRS^2 model achieves the state-of-the-art gland segmentation performance on both benchmarks.