An Adversarial Human Pose Estimation Network Injected with Graph Structure

TL;DR

This paper introduces a novel GAN-based framework with graph structures to improve human pose estimation, especially for invisible joints, by leveraging joint relationships and adversarial training.

Contribution

It proposes a new GAN architecture incorporating a Graph Structure Network to enhance localization of invisible joints in human pose estimation.

Findings

Improved accuracy on LSP, MPII, and COCO datasets.

Effective localization of invisible joints.

Demonstrated superiority over existing methods.

Abstract

Because of the invisible human keypoints in images caused by illumination, occlusion and overlap, it is likely to produce unreasonable human pose prediction for most of the current human pose estimation methods. In this paper, we design a novel generative adversarial network (GAN) to improve the localization accuracy of visible joints when some joints are invisible. The network consists of two simple but efficient modules, Cascade Feature Network (CFN) and Graph Structure Network (GSN). First, the CFN utilizes the prediction maps from the previous stages to guide the prediction maps in the next stage to produce accurate human pose. Second, the GSN is designed to contribute to the localization of invisible joints by passing message among different joints. According to GAN, if the prediction pose produced by the generator G cannot be distinguished by the discriminator D, the generator network G has successfully obtained the underlying dependence of human joints. We conduct experiments on three widely used human pose estimation benchmark datasets, LSP, MPII and COCO, whose results show the effectiveness of our proposed framework.