Weakly-supervised 3D Human Pose Estimation with Cross-view U-shaped Graph Convolutional Network

TL;DR

This paper introduces a weakly-supervised cross-view 3D human pose estimation method that leverages two camera views and a novel graph convolutional network to achieve state-of-the-art accuracy without requiring 3D ground truth annotations.

Contribution

It proposes a new pipeline combining triangulation and a cross-view U-shaped graph convolutional network for weakly-supervised 3D pose estimation from two views.

Findings

Achieves 27.4 mm MPJPE on Human3.6M, outperforming previous methods.

Uses only 2D annotations and no 3D ground truth for training.

Demonstrates effectiveness of cross-view GCN in pose refinement.

Abstract

Although monocular 3D human pose estimation methods have made significant progress, it is far from being solved due to the inherent depth ambiguity. Instead, exploiting multi-view information is a practical way to achieve absolute 3D human pose estimation. In this paper, we propose a simple yet effective pipeline for weakly-supervised cross-view 3D human pose estimation. By only using two camera views, our method can achieve state-of-the-art performance in a weakly-supervised manner, requiring no 3D ground truth but only 2D annotations. Specifically, our method contains two steps: triangulation and refinement. First, given the 2D keypoints that can be obtained through any classic 2D detection methods, triangulation is performed across two views to lift the 2D keypoints into coarse 3D poses. Then, a novel cross-view U-shaped graph convolutional network (CV-UGCN), which can explore the spatial configurations and cross-view correlations, is designed to refine the coarse 3D poses. In particular, the refinement progress is achieved through weakly-supervised learning, in which geometric and structure-aware consistency checks are performed. We evaluate our method on the standard benchmark dataset, Human3.6M. The Mean Per Joint Position Error on the benchmark dataset is 27.4 mm, which outperforms existing state-of-the-art methods remarkably (27.4 mm vs 30.2 mm).