HTNet: Human Topology Aware Network for 3D Human Pose Estimation

TL;DR

HTNet is a novel neural network that leverages human body topology through hierarchical graph convolutions and self-attention to significantly improve 3D human pose estimation accuracy, especially at limb extremities.

Contribution

The paper introduces HTNet, which incorporates topological constraints and hierarchical dependencies to enhance pose estimation accuracy over prior methods.

Findings

Improves limb end joint accuracy by 18.7%.

Achieves state-of-the-art results on Human3.6M and MPI-INF-3DHP datasets.

Utilizes a channel-split progressive strategy for multi-level structural learning.

Abstract

3D human pose estimation errors would propagate along the human body topology and accumulate at the end joints of limbs. Inspired by the backtracking mechanism in automatic control systems, we design an Intra-Part Constraint module that utilizes the parent nodes as the reference to build topological constraints for end joints at the part level. Further considering the hierarchy of the human topology, joint-level and body-level dependencies are captured via graph convolutional networks and self-attentions, respectively. Based on these designs, we propose a novel Human Topology aware Network (HTNet), which adopts a channel-split progressive strategy to sequentially learn the structural priors of the human topology from multiple semantic levels: joint, part, and body. Extensive experiments show that the proposed method improves the estimation accuracy by 18.7% on the end joints of limbs and achieves state-of-the-art results on Human3.6M and MPI-INF-3DHP datasets. Code is available at https://github.com/vefalun/HTNet.