Unbiased Subclass Regularization for Semi-Supervised Semantic Segmentation

TL;DR

This paper introduces USRN, a novel semi-supervised segmentation method that addresses class imbalance by learning from balanced subclasses and uses an entropy-based gate to improve class-unbiased segmentation.

Contribution

The paper proposes an unbiased subclass regularization approach with a clustering-based subclass distribution and an entropy gate for better semi-supervised segmentation.

Findings

USRN outperforms state-of-the-art methods on multiple benchmarks.

Balanced subclass learning reduces class bias in segmentation.

Entropy gating improves the coordination between classes and subclasses.

Abstract

Semi-supervised semantic segmentation learns from small amounts of labelled images and large amounts of unlabelled images, which has witnessed impressive progress with the recent advance of deep neural networks. However, it often suffers from severe class-bias problem while exploring the unlabelled images, largely due to the clear pixel-wise class imbalance in the labelled images. This paper presents an unbiased subclass regularization network (USRN) that alleviates the class imbalance issue by learning class-unbiased segmentation from balanced subclass distributions. We build the balanced subclass distributions by clustering pixels of each original class into multiple subclasses of similar sizes, which provide class-balanced pseudo supervision to regularize the class-biased segmentation. In addition, we design an entropy-based gate mechanism to coordinate learning between the original classes and the clustered subclasses which facilitates subclass regularization effectively by suppressing unconfident subclass predictions. Extensive experiments over multiple public benchmarks show that USRN achieves superior performance as compared with the state-of-the-art.