MoSAIC: Aligned Intervention Supervision for Part-Local Motion Style Transfer

2026-07-28

AI Summary

This work addresses the challenge of localized style transfer in character motion editing by proposing MoSAIC, a fine-grained controllable framework based on latent diffusion models. The approach achieves high-fidelity local editing through disentangling content and reference features, independently modeling root trajectories, routing reference signals by body part, and incorporating an alignment intervention supervision mechanism that constructs synchronized reference and counterfactual targets via controlled local transformations. Experiments demonstrate that, across 128 motions and 896 routing configurations, local mask routing reduces error in preserved regions from 70.64 to 66.45 mm compared to global routing, decreases non-target leakage by 45.3% (from 18.08 to 9.88 mm), and improves target responsiveness by 8.8% and routing focus by 2.0 percentage points under budget-consistent settings.

Abstract

Editing character motion often requires transferring a gesture or gait from one or more reference motions while preserving the source action, timing, root trajectory, and unselected body regions. Existing motion datasets, however, rarely provide paired targets for arbitrary part-local content--reference combinations, and self-reconstruction training may allow a diffusion model to reproduce the content motion while underusing the routed reference. We present MoSAIC, a latent diffusion framework for part-local reference-conditioned motion style transfer. MoSAIC factorizes content and reference features by anatomical region, preserves the root trajectory through a separate conditioning pathway, and routes user-selected references to individual body parts. Its central contribution is aligned intervention supervision, which constructs synchronized references and counterfactual targets through controlled local transformations, making both the requested regional response and the motion to be preserved directly observable during training. In a frozen evaluation comprising 128 motions and 896 routed conditions, part-masked routing reduces preserved-region error from 70.64 to 66.45~mm and matched-noise off-target leakage from 18.08 to 9.88~mm relative to whole-body routing, while retaining a positive selected-region response. A matched-budget continuation study further shows that retaining aligned intervention supervision produces an 8.8\% relative increase in selected-target response and a 2.0-percentage-point increase in requested-route influence concentration. These results demonstrate that MoSAIC improves the response--preservation trade-off required for selective and controllable part-local motion editing.

Problem

• motion style transfer
• part-local editing
• reference-conditioned generation
• motion preservation
• aligned supervision

Innovation

• aligned intervention supervision
• part-local motion style transfer
• latent diffusion model
• reference-conditioned motion editing
• anatomical region factorization