Beyond the Prototype: Divide-and-conquer Proxies for Few-shot Segmentation

TL;DR

This paper introduces a divide-and-conquer proxy framework for few-shot segmentation that improves object boundary accuracy and reduces segmentation errors by leveraging support image regions and a novel decoding structure, achieving state-of-the-art results.

Contribution

The paper proposes a novel divide-and-conquer proxy framework with a self-reasoning scheme and parallel decoder, enhancing few-shot segmentation beyond prototype-based methods.

Findings

Outperforms conventional prototype-based approaches by 5-10% on PASCAL-5i and COCO-20i datasets.

Achieves new state-of-the-art performance in few-shot segmentation.

Demonstrates robustness in segmenting incomplete objects and ambiguous boundaries.

Abstract

Few-shot segmentation, which aims to segment unseen-class objects given only a handful of densely labeled samples, has received widespread attention from the community. Existing approaches typically follow the prototype learning paradigm to perform meta-inference, which fails to fully exploit the underlying information from support image-mask pairs, resulting in various segmentation failures, e.g., incomplete objects, ambiguous boundaries, and distractor activation. To this end, we propose a simple yet versatile framework in the spirit of divide-and-conquer. Specifically, a novel self-reasoning scheme is first implemented on the annotated support image, and then the coarse segmentation mask is divided into multiple regions with different properties. Leveraging effective masked average pooling operations, a series of support-induced proxies are thus derived, each playing a specific role in conquering the above challenges. Moreover, we devise a unique parallel decoder structure that integrates proxies with similar attributes to boost the discrimination power. Our proposed approach, named divide-and-conquer proxies (DCP), allows for the development of appropriate and reliable information as a guide at the "episode" level, not just about the object cues themselves. Extensive experiments on PASCAL-5i and COCO-20i demonstrate the superiority of DCP over conventional prototype-based approaches (up to 5~10% on average), which also establishes a new state-of-the-art. Code is available at github.com/chunbolang/DCP.