Paper Overview
Field: Computer Graphics Authors: Kaifeng Zhao, Mathis Petrovich, Haotian Zhang Published: 2025-07-12 arXiv: 2507.08713
Abstract
Generating realistic 3D human motions in real-time within interactive applications is key for animation, simulation, and humanoid robotics. While recent offline motion generation approaches offer precise control via text and kinematic constraints, they lack the inference speed required for interactive settings. Conversely, existing online methods enable real-time synthesis but often sacrifice controllability. In this work, we introduce ARDY, a streaming generation framework that bridges this gap by enabling high-fidelity motion generation controllable via online text prompts and flexible kinematic constraints. ARDY employs a hybrid representation that combines explicit root features with a latent body embedding. Extensive evaluations on the HumanML3D benchmark and the large-scale, high-fidelity Bones Rigplay dataset demonstrate ARDY's high motion quality and constraint adherence.
Code and models: https://research.nvidia.com/labs/sil/projects/ardy/
Key Points
- Problem: Offline motion generation offers precise control but is too slow for interactive applications; online methods are fast but hard to control.
- Solution: ARDY, a streaming (autoregressive diffusion) generation framework that combines real-time synthesis with controllability.
- Control: Supports online text prompts and flexible kinematic constraints during generation.
- Architecture: Hybrid representation mixing explicit root features with a latent body embedding.
- Evaluation: Strong motion quality and constraint adherence on HumanML3D and the large-scale Bones Rigplay dataset.
*Auto-collected on 2025-07-13*