TAME:Topology-Aware Text-Driven Motion Editing across Heterogeneous Humanoid Skeletons

2026-10-03

AI Summary

This study addresses the limitation of existing text-driven motion editing methods that are constrained to a single skeletal topology and struggle to adapt to heterogeneous humanoid skeletons. To this end, we propose TAME, a Topology-Aware Motion Editor built upon a flow matching Transformer architecture. Specifically, TAME introduces a topology-constrained skeletal propagation mechanism to respect hierarchical structures and designs an edit-focused representation alignment strategy to enhance instruction fidelity. Furthermore, we construct TopoMotionFix, a multi-topology benchmark dataset. Experimental results demonstrate that TAME surpasses existing methods in both editing alignment and source motion preservation on the MotionFix dataset, while achieving reliable generalization to unseen skeletal topologies on TopoMotionFix.

Abstract

Text-driven motion editing modifies an existing motion sequence according to a text instruction while preserving the content of the source motion. Existing methods are typically built for a single, fixed skeletal topology, which limits their use in animation pipelines where characters differ in joint count and skeletal hierarchy. We present Topology-Aware Motion Editor (TAME), a flow-matching transformer that edits motions on humanoid skeletons of varying topology. TAME represents motion as per-joint, per-frame tokens and models interactions among joints, across frames, and with the text instruction through skeletal, temporal, and text cross-attention layers. To make the skeletal attention follow each character's hierarchy, TAME replaces full joint attention with Topology-Constrained Skeletal Propagation (TCSP), which restricts attention to one-hop kinematic neighbors in the skeleton's adjacency matrix. We further introduce Edit-Focused Representation Alignment (EFRA), a self-distilled representation alignment strategy that aligns student features with cleaner EMA-teacher features exclusively on edit-relevant joint-time tokens, making edits faithful to the instruction. To make this setting trainable and comparable, we construct TopoMotionFix, a multi-topology extension of MotionFix with seen- and unseen-topology evaluation protocols. TAME outperforms previous methods in edit alignment and source preservation on MotionFix and reliably edits motions on unseen skeletons in TopoMotionFix.

Problem

• text-driven motion editing
• heterogeneous humanoid skeletons
• skeletal topology
• motion representation

Innovation

• Topology-Aware Motion Editing
• Flow-Matching Transformer
• Topology-Constrained Skeletal Propagation
• Edit-Focused Representation Alignment
• Heterogeneous Skeletons