REWIND: Real-Time Egocentric Whole-Body Motion Diffusion with Exemplar-Based Identity Conditioning

2025-04-07

AI Summary

Existing egocentric full-body motion estimation methods rely on non-causal diffusion models, compromising real-time inference and high-fidelity reconstruction. This paper introduces the first causal, real-time, high-fidelity full-body (including hands) motion estimation framework. We propose a cascaded body-hand denoising diffusion architecture to explicitly decouple and model their dynamic interdependencies; employ single-step diffusion distillation to accelerate inference; and design an identity-conditioning mechanism leveraging only a few pose exemplars for lightweight, identity-aware generation. An enhanced causal Transformer ensures temporal consistency. Our method achieves state-of-the-art performance across multiple benchmarks—delivering real-time inference (>30 FPS), superior accuracy, and robust generalization across varying sequence lengths.

Technology Category

• Computer Vision: Motion & Tracking
• Intelligent Robots: State Estimation
• Machine Learning: Causal Learning

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

We present REWIND (Real-Time Egocentric Whole-Body Motion Diffusion), a one-step diffusion model for real-time, high-fidelity human motion estimation from egocentric image inputs. While an existing method for egocentric whole-body (i.e., body and hands) motion estimation is non-real-time and acausal due to diffusion-based iterative motion refinement to capture correlations between body and hand poses, REWIND operates in a fully causal and real-time manner. To enable real-time inference, we introduce (1) cascaded body-hand denoising diffusion, which effectively models the correlation between egocentric body and hand motions in a fast, feed-forward manner, and (2) diffusion distillation, which enables high-quality motion estimation with a single denoising step. Our denoising diffusion model is based on a modified Transformer architecture, designed to causally model output motions while enhancing generalizability to unseen motion lengths. Additionally, REWIND optionally supports identity-conditioned motion estimation when identity prior is available. To this end, we propose a novel identity conditioning method based on a small set of pose exemplars of the target identity, which further enhances motion estimation quality. Through extensive experiments, we demonstrate that REWIND significantly outperforms the existing baselines both with and without exemplar-based identity conditioning.

Problem

• Real-time egocentric whole-body motion estimation from images
• Modeling body-hand correlations causally and efficiently
• Enhancing motion quality with exemplar-based identity conditioning

Innovation

• One-step diffusion model for real-time motion estimation
• Cascaded body-hand denoising diffusion for fast correlation modeling
• Exemplar-based identity conditioning for enhanced motion quality