Weakly-Supervised Image Semantic Segmentation Using Graph Convolutional Networks

TL;DR

This paper introduces a GCN-based framework for weakly-supervised image segmentation that improves pseudo label quality by regularizing feature propagation, outperforming existing methods on PASCAL VOC 2012.

Contribution

It proposes a novel GCN-based feature propagation method formulated as semi-supervised learning for better pseudo label generation in weakly-supervised segmentation.

Findings

Outperforms state-of-the-art baselines on PASCAL VOC 2012

Uses a 2-layer GCN with regularization for pseudo label refinement

Demonstrates the effectiveness of GCN in weakly-supervised segmentation

Abstract

This work addresses weakly-supervised image semantic segmentation based on image-level class labels. One common approach to this task is to propagate the activation scores of Class Activation Maps (CAMs) using a random-walk mechanism in order to arrive at complete pseudo labels for training a semantic segmentation network in a fully-supervised manner. However, the feed-forward nature of the random walk imposes no regularization on the quality of the resulting complete pseudo labels. To overcome this issue, we propose a Graph Convolutional Network (GCN)-based feature propagation framework. We formulate the generation of complete pseudo labels as a semi-supervised learning task and learn a 2-layer GCN separately for every training image by back-propagating a Laplacian and an entropy regularization loss. Experimental results on the PASCAL VOC 2012 dataset confirm the superiority of our scheme to several state-of-the-art baselines. Our code is available at https://github.com/Xavier-Pan/WSGCN.