Adversarial 3D Human Pose Estimation via Multimodal Depth Supervision

TL;DR

This paper introduces a deep learning framework that combines explicit and implicit depth information, along with adversarial training, to improve 3D human pose estimation from a single image, achieving competitive accuracy.

Contribution

It presents a novel two-phase approach with multimodal depth supervision and adversarial training for enhanced 3D pose estimation.

Findings

Achieves MPJPE of 58.68mm on ECCV2018 challenge

Utilizes a dual-branch generator for depth extraction

Employs adversarial scheme to boost performance

Abstract

In this paper, a novel deep-learning based framework is proposed to infer 3D human poses from a single image. Specifically, a two-phase approach is developed. We firstly utilize a generator with two branches for the extraction of explicit and implicit depth information respectively. During the training process, an adversarial scheme is also employed to further improve the performance. The implicit and explicit depth information with the estimated 2D joints generated by a widely used estimator, in the second step, are together fed into a deep 3D pose regressor for the final pose generation. Our method achieves MPJPE of 58.68mm on the ECCV2018 3D Human Pose Estimation Challenge.