Adaptive Prototype Learning and Allocation for Few-Shot Segmentation

TL;DR

This paper introduces ASGNet, a lightweight adaptive network for few-shot segmentation that uses superpixel-guided clustering and prototype allocation to improve accuracy and handle object variations without extra computational cost.

Contribution

It proposes novel modules SGC and GPA for multiple prototype extraction and allocation, enhancing few-shot segmentation performance.

Findings

ASGNet surpasses state-of-the-art by 5% in 5-shot segmentation on COCO.

The model adapts to object scale and shape variations effectively.

No additional computational cost for k-shot segmentation.

Abstract

Prototype learning is extensively used for few-shot segmentation. Typically, a single prototype is obtained from the support feature by averaging the global object information. However, using one prototype to represent all the information may lead to ambiguities. In this paper, we propose two novel modules, named superpixel-guided clustering (SGC) and guided prototype allocation (GPA), for multiple prototype extraction and allocation. Specifically, SGC is a parameter-free and training-free approach, which extracts more representative prototypes by aggregating similar feature vectors, while GPA is able to select matched prototypes to provide more accurate guidance. By integrating the SGC and GPA together, we propose the Adaptive Superpixel-guided Network (ASGNet), which is a lightweight model and adapts to object scale and shape variation. In addition, our network can easily generalize to k-shot segmentation with substantial improvement and no additional computational cost. In particular, our evaluations on COCO demonstrate that ASGNet surpasses the state-of-the-art method by 5% in 5-shot segmentation.