Mitigating Background Shift in Class-Incremental Semantic Segmentation

TL;DR

This paper introduces a background-class separation framework for class-incremental semantic segmentation, effectively mitigating background shift issues during incremental learning through selective pseudo-labeling, adaptive distillation, and orthogonal objectives.

Contribution

It proposes a novel background-class separation framework that addresses background shift in CISS by combining selective pseudo-labeling, adaptive distillation, and orthogonal objectives.

Findings

Achieves state-of-the-art results on benchmark datasets.

Effectively reduces background shift towards old and new classes.

Improves segmentation accuracy in incremental learning scenarios.

Abstract

Class-Incremental Semantic Segmentation(CISS) aims to learn new classes without forgetting the old ones, using only the labels of the new classes. To achieve this, two popular strategies are employed: 1) pseudo-labeling and knowledge distillation to preserve prior knowledge; and 2) background weight transfer, which leverages the broad coverage of background in learning new classes by transferring background weight to the new class classifier. However, the first strategy heavily relies on the old model in detecting old classes while undetected pixels are regarded as the background, thereby leading to the background shift towards the old classes(i.e., misclassification of old class as background). Additionally, in the case of the second approach, initializing the new class classifier with background knowledge triggers a similar background shift issue, but towards the new classes. To address these issues, we propose a background-class separation framework for CISS. To begin with, selective pseudo-labeling and adaptive feature distillation are to distill only trustworthy past knowledge. On the other hand, we encourage the separation between the background and new classes with a novel orthogonal objective along with label-guided output distillation. Our state-of-the-art results validate the effectiveness of these proposed methods.