Spatial Temporal Synergy: Balancing Change and Invariance in Text Driven 3D Human Motion Editing
2026-08-17 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors created a new method called CIME to edit human motion sequences based on text instructions while keeping the original motion's structure intact. Their approach separates the changes in body poses from the timing or rhythm, using special learning techniques to keep motions realistic and aligned with the text. They also developed a module to better capture natural movement rhythms with variable timing. Testing on standard datasets showed that CIME improves both how well edits match the instructions and how natural the motions look.
human motion editingdiffusion modelsspatial posetemporal rhythmfeature supervisionsemantic alignmentRiemannian manifoldnon-uniform timestampskinematicsnatural language instructions
Authors
Shaohui Lin, Zhenwu Shi, Jingyu Gong, Jiao Xie, Yu Zhou, Baochang Zhang, Lizhuang Ma, Chia-Wen Lin
Abstract
Text-driven human motion editing aims to modify existing motion sequences according to natural language instructions while maintaining the structural consistency of the original motion. Existing diffusion-based approaches struggle to balance text-responsive "change" and inertial "invariance". They often rely on coarse spatial constraints and rigid uniform time assumptions, leading to spatial motion distortions and the destruction of intrinsic physical rhythms during variable-length editing. To handle these challenges, we propose Change and Invariance Motion Editing (CIME), a unified framework that comprehensively decouples change and invariance into spatial pose and temporal rhythm dimensions. For spatial poses, our method integrates an omni-supervised positive-negative learning mechanism comprising hierarchical retrospective feature supervision, subtle motion preservation, and triplet-based semantic alignment. For temporal rhythms, we introduce the Riemannian Non-uniform Integral Manifold Mapping (RNIMM) module, which achieves high-fidelity reproduction of physical beats in the edited text via kinematics-aware non-uniform timestamps. Extensive experiments on the MotionFix and STANCE Adjustment datasets demonstrate that CIME achieves state-of-the-art performance in editing alignment and structural fidelity, validating the effectiveness of our unified architecture. Our source codes and models have been released at: github.com/ZhenwuShi/CIME.git