Mining Latent Classes for Few-shot Segmentation

TL;DR

This paper introduces a novel joint-training framework for few-shot segmentation that mitigates feature undermining by mining latent classes and leveraging unlabeled data, significantly improving performance over previous methods.

Contribution

It proposes a transferable sub-cluster mining approach with a rectification technique and unlabeled data utilization, enhancing feature embedding for unseen classes in FSS.

Findings

Outperforms state-of-the-art by 3.7% mIOU on PASCAL-5i

Achieves 7.0% mIOU improvement on COCO-20i

Reduces parameters by 74% and doubles inference speed

Abstract

Few-shot segmentation (FSS) aims to segment unseen classes given only a few annotated samples. Existing methods suffer the problem of feature undermining, i.e. potential novel classes are treated as background during training phase. Our method aims to alleviate this problem and enhance the feature embedding on latent novel classes. In our work, we propose a novel joint-training framework. Based on conventional episodic training on support-query pairs, we add an additional mining branch that exploits latent novel classes via transferable sub-clusters, and a new rectification technique on both background and foreground categories to enforce more stable prototypes. Over and above that, our transferable sub-cluster has the ability to leverage extra unlabeled data for further feature enhancement. Extensive experiments on two FSS benchmarks demonstrate that our method outperforms previous state-of-the-art by a large margin of 3.7% mIOU on PASCAL-5i and 7.0% mIOU on COCO-20i at the cost of 74% fewer parameters and 2.5x faster inference speed. The source code is available at https://github.com/LiheYoung/MiningFSS.