3D Human Pose Regression using Graph Convolutional Network

TL;DR

This paper introduces PoseGraphNet, a graph convolutional network that improves 3D human pose estimation from 2D inputs by learning adaptive adjacency matrices, achieving near state-of-the-art results with fewer parameters.

Contribution

PoseGraphNet employs adaptive adjacency matrices and joint-specific kernels, capturing both physical and behavioral joint relations for improved 3D pose regression.

Findings

Achieves performance close to state-of-the-art on Human3.6M.

Uses fewer parameters than comparable models.

Learns meaningful non-physical joint relations.

Abstract

3D human pose estimation is a difficult task, due to challenges such as occluded body parts and ambiguous poses. Graph convolutional networks encode the structural information of the human skeleton in the form of an adjacency matrix, which is beneficial for better pose prediction. We propose one such graph convolutional network named PoseGraphNet for 3D human pose regression from 2D poses. Our network uses an adaptive adjacency matrix and kernels specific to neighbor groups. We evaluate our model on the Human3.6M dataset which is a standard dataset for 3D pose estimation. Our model's performance is close to the state-of-the-art, but with much fewer parameters. The model learns interesting adjacency relations between joints that have no physical connections, but are behaviorally similar.