Learning Human Motion from Monocular Videos via Cross-Modal Manifold Alignment

TL;DR

This paper introduces VTM, a novel method for reconstructing 3D human motion from monocular videos by aligning 3D motion and 2D inputs in a shared latent space, effectively handling ambiguities and variations.

Contribution

The paper proposes a cross-modal latent space alignment approach that models upper and lower body motions separately and uses a scale-invariant skeleton for robust 3D motion reconstruction.

Findings

Achieves state-of-the-art results on AIST++ dataset.

Generalizes well to unseen view angles.

Performs effectively on in-the-wild videos.

Abstract

Learning 3D human motion from 2D inputs is a fundamental task in the realms of computer vision and computer graphics. Many previous methods grapple with this inherently ambiguous task by introducing motion priors into the learning process. However, these approaches face difficulties in defining the complete configurations of such priors or training a robust model. In this paper, we present the Video-to-Motion Generator (VTM), which leverages motion priors through cross-modal latent feature space alignment between 3D human motion and 2D inputs, namely videos and 2D keypoints. To reduce the complexity of modeling motion priors, we model the motion data separately for the upper and lower body parts. Additionally, we align the motion data with a scale-invariant virtual skeleton to mitigate the interference of human skeleton variations to the motion priors. Evaluated on AIST++, the VTM showcases state-of-the-art performance in reconstructing 3D human motion from monocular videos. Notably, our VTM exhibits the capabilities for generalization to unseen view angles and in-the-wild videos.