BioHuman: Learning Biomechanical Human Representations from Video

TL;DR

BioHuman introduces a novel framework and large-scale dataset for estimating internal biomechanical states, like muscle activations, directly from monocular video, advancing human motion understanding.

Contribution

The paper presents BioHuman, an end-to-end model that predicts human motion and muscle activations from video, supported by the new BioHuman10M dataset, bridging visual data and internal biomechanics.

Findings

BioHuman accurately reconstructs muscle activations from video.

The model generalizes well across subjects and motions.

BioHuman sets a new benchmark for video-based biomechanical analysis.

Abstract

Understanding human motion beyond surface kinematics is crucial for motion analysis, rehabilitation, and injury risk assessment. However, progress in this domain is limited by the lack of large-scale datasets with biomechanical annotations, and by existing approaches that cannot directly infer internal biomechanical states from visual observations. In this paper, we introduce a simulation-based framework for estimating muscle activations from existing motion capture datasets, resulting in BioHuman10M, a large-scale dataset with synchronized video, motion, and activations. Building on BioHuman10M, we propose BioHuman, an end-to-end model that takes monocular video as input and jointly predicts human motion and muscle activations, effectively bridging visual observations and internal biomechanical states. Extensive experiments demonstrate that BioHuman enables accurate reconstruction of both kinematic motion and muscle activity, and generalizes across diverse subjects and motions. We believe our approach establishes a new benchmark for video-based biomechanical understanding and opens up new possibilities for physically grounded human modeling.