Learning Monocular 3D Human Pose Estimation from Multi-view Images

TL;DR

This paper introduces a multi-view training approach for monocular 3D human pose estimation that reduces reliance on manual annotations by leveraging multi-view consistency and joint camera pose estimation, effective even with uncalibrated cameras.

Contribution

It presents a novel training method combining multi-view consistency, limited supervision, and camera pose estimation to improve 3D human pose accuracy without extensive labeled data.

Findings

Effective on standard benchmarks

Performs well on challenging Ski dataset

Reduces annotation effort significantly

Abstract

Accurate 3D human pose estimation from single images is possible with sophisticated deep-net architectures that have been trained on very large datasets. However, this still leaves open the problem of capturing motions for which no such database exists. Manual annotation is tedious, slow, and error-prone. In this paper, we propose to replace most of the annotations by the use of multiple views, at training time only. Specifically, we train the system to predict the same pose in all views. Such a consistency constraint is necessary but not sufficient to predict accurate poses. We therefore complement it with a supervised loss aiming to predict the correct pose in a small set of labeled images, and with a regularization term that penalizes drift from initial predictions. Furthermore, we propose a method to estimate camera pose jointly with human pose, which lets us utilize multi-view footage where calibration is difficult, e.g., for pan-tilt or moving handheld cameras. We demonstrate the effectiveness of our approach on established benchmarks, as well as on a new Ski dataset with rotating cameras and expert ski motion, for which annotations are truly hard to obtain.