An EM Framework for Online Incremental Learning of Semantic Segmentation

TL;DR

This paper introduces an EM-based incremental learning framework for semantic segmentation that effectively adapts models to new classes in an online setting while mitigating catastrophic forgetting.

Contribution

It proposes a unified EM-based strategy with relabeling, rehearsal, adaptive sampling, and class balancing for online incremental semantic segmentation.

Findings

Outperforms existing incremental methods on PASCAL VOC 2012.

Effectively balances stability and plasticity in model learning.

Handles partial annotations and evolving label spaces.

Abstract

Incremental learning of semantic segmentation has emerged as a promising strategy for visual scene interpretation in the open- world setting. However, it remains challenging to acquire novel classes in an online fashion for the segmentation task, mainly due to its continuously-evolving semantic label space, partial pixelwise ground-truth annotations, and constrained data availability. To ad- dress this, we propose an incremental learning strategy that can fast adapt deep segmentation models without catastrophic forgetting, using a streaming input data with pixel annotations on the novel classes only. To this end, we develop a uni ed learning strategy based on the Expectation-Maximization (EM) framework, which integrates an iterative relabeling strategy that lls in the missing labels and a rehearsal-based incremental learning step that balances the stability-plasticity of the model. Moreover, our EM algorithm adopts an adaptive sampling method to select informative train- ing data and a class-balancing training strategy in the incremental model updates, both improving the e cacy of model learning. We validate our approach on the PASCAL VOC 2012 and ADE20K datasets, and the results demonstrate its superior performance over the existing incremental methods.