Weakly-Supervised Semantic Segmentation with Image-Level Labels: from Traditional Models to Foundation Models

TL;DR

This paper surveys traditional weakly-supervised semantic segmentation methods using image-level labels and explores the potential of foundation models like SAM for advancing WSSS, highlighting challenges and future directions.

Contribution

It provides a comprehensive categorization of traditional WSSS methods and investigates the application of foundation models such as SAM in WSSS tasks.

Findings

Traditional methods are categorized into pixel-wise, image-wise, cross-image, and external data groups.

Foundation models like SAM show potential for WSSS through text prompting and zero-shot learning.

Insights into challenges and future opportunities for foundation models in WSSS.

Abstract

The rapid development of deep learning has driven significant progress in image semantic segmentation - a fundamental task in computer vision. Semantic segmentation algorithms often depend on the availability of pixel-level labels (i.e., masks of objects), which are expensive, time-consuming, and labor-intensive. Weakly-supervised semantic segmentation (WSSS) is an effective solution to avoid such labeling. It utilizes only partial or incomplete annotations and provides a cost-effective alternative to fully-supervised semantic segmentation. In this journal, our focus is on the WSSS with image-level labels, which is the most challenging form of WSSS. Our work has two parts. First, we conduct a comprehensive survey on traditional methods, primarily focusing on those presented at premier research conferences. We categorize them into four groups based on where their methods operate: pixel-wise, image-wise, cross-image, and external data. Second, we investigate the applicability of visual foundation models, such as the Segment Anything Model (SAM), in the context of WSSS. We scrutinize SAM in two intriguing scenarios: text prompting and zero-shot learning. We provide insights into the potential and challenges of deploying visual foundational models for WSSS, facilitating future developments in this exciting research area.