Generalized Few-shot Semantic Segmentation

TL;DR

This paper introduces GFS-Seg, a new benchmark for evaluating the ability of models to segment both novel and base categories with few examples, revealing limitations of existing FS-Seg methods and proposing a context-aware approach to improve performance.

Contribution

The paper presents the GFS-Seg benchmark, highlights the shortcomings of current FS-Seg methods in this setting, and proposes CAPL, a context-aware prototype learning method that enhances segmentation accuracy.

Findings

GFS-Seg benchmark exposes performance gaps in existing FS-Seg methods.

CAPL significantly improves segmentation performance by leveraging contextual information.

CAPL achieves competitive results on Pascal-VOC and COCO datasets.

Abstract

Training semantic segmentation models requires a large amount of finely annotated data, making it hard to quickly adapt to novel classes not satisfying this condition. Few-Shot Segmentation (FS-Seg) tackles this problem with many constraints. In this paper, we introduce a new benchmark, called Generalized Few-Shot Semantic Segmentation (GFS-Seg), to analyze the generalization ability of simultaneously segmenting the novel categories with very few examples and the base categories with sufficient examples. It is the first study showing that previous representative state-of-the-art FS-Seg methods fall short in GFS-Seg and the performance discrepancy mainly comes from the constrained setting of FS-Seg. To make GFS-Seg tractable, we set up a GFS-Seg baseline that achieves decent performance without structural change on the original model. Then, since context is essential for semantic segmentation, we propose the Context-Aware Prototype Learning (CAPL) that significantly improves performance by 1) leveraging the co-occurrence prior knowledge from support samples, and 2) dynamically enriching contextual information to the classifier, conditioned on the content of each query image. Both two contributions are experimentally shown to have substantial practical merit. Extensive experiments on Pascal-VOC and COCO manifest the effectiveness of CAPL, and CAPL generalizes well to FS-Seg by achieving competitive performance. Code is available at https://github.com/dvlab-research/GFS-Seg.