Human Motion Prediction Using Manifold-Aware Wasserstein GAN

TL;DR

This paper introduces a novel manifold-aware Wasserstein GAN that models human motion trajectories on a sphere manifold, improving prediction smoothness and long-term accuracy over existing methods.

Contribution

It proposes a compact manifold-valued representation of human motion and a Wasserstein GAN tailored for non-Euclidean data, enhancing motion prediction quality.

Findings

Outperforms state-of-the-art on CMU MoCap and Human 3.6M datasets.

Produces smoother and more accurate long-term human motion predictions.

Effectively captures temporal and spatial dependencies in human motion data.

Abstract

Human motion prediction aims to forecast future human poses given a prior pose sequence. The discontinuity of the predicted motion and the performance deterioration in long-term horizons are still the main challenges encountered in current literature. In this work, we tackle these issues by using a compact manifold-valued representation of human motion. Specifically, we model the temporal evolution of the 3D human poses as trajectory, what allows us to map human motions to single points on a sphere manifold. To learn these non-Euclidean representations, we build a manifold-aware Wasserstein generative adversarial model that captures the temporal and spatial dependencies of human motion through different losses. Extensive experiments show that our approach outperforms the state-of-the-art on CMU MoCap and Human 3.6M datasets. Our qualitative results show the smoothness of the predicted motions.