Spatial Temporal Synergy: Balancing Change and Invariance in Text Driven 3D Human Motion Editing

2026-08-16

AI Summary

This study addresses the inherent trade-off between semantic responsiveness and structural fidelity in text-driven 3D motion editing by proposing the CIME framework. The method introduces a novel spatiotemporal collaborative decoupling mechanism that disentangles variations and invariances into pose and rhythm dimensions. By integrating fully supervised positive-negative learning with Riemannian Non-uniform Integral Manifold Mapping (RNIMM), CIME achieves a refined balance between semantic alignment and physical rhythmic consistency. Experimental results on benchmarks such as MotionFix demonstrate that CIME attains state-of-the-art performance, significantly enhancing both editing alignment and structural preservation. The source code and pretrained models have been made publicly available to facilitate future research.

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

Problem

• Text-driven 3D Human Motion Editing
• Change and Invariance Balance
• Spatial Motion Distortion
• Physical Rhythm Preservation

Innovation

• Change and Invariance Decoupling
• Omni-supervised Positive-Negative Learning
• Riemannian Non-uniform Integral Manifold Mapping
• Text-driven 3D Human Motion Editing
• Kinematics-aware Timestamps