Semantic Segmentation from Image Labels by Reconstruction from Structured Decomposition

TL;DR

This paper introduces a novel weakly supervised image segmentation method that reconstructs images from their decomposed masks, embedding regularization implicitly and demonstrating robustness against background ambiguity.

Contribution

It proposes framing weakly supervised segmentation as a reconstruction from decomposition problem, differing from traditional CAM-based approaches.

Findings

Promising initial experimental results

Robustness against background ambiguity

Implicit regularization within the reconstruction framework

Abstract

Weakly supervised image segmentation (WSSS) from image tags remains challenging due to its under-constraint nature. Most mainstream work focus on the extraction of class activation map (CAM) and imposing various additional regularization. Contrary to the mainstream, we propose to frame WSSS as a problem of reconstruction from decomposition of the image using its mask, under which most regularization are embedded implicitly within the framework of the new problem. Our approach has demonstrated promising results on initial experiments, and shown robustness against the problem of background ambiguity. Our code is available at \url{https://github.com/xuanrui-work/WSSSByRec}.