AI
Morphometric Imitation: From Morphology and Contact Aware Hand Retargeting to Sim-to-Real Visuomotor Policy
Publication: Hugging Face Daily Papers
Author: Not available in collected metadata
Source published: 22/09/2026, 20:00:00 UTC
Briefing created: 28/09/2026, 02:22:16 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 28/09/2026, 02:18:25 UTC.
Briefings
Morphometric Imitation is a three-stage framework converting human hand-object interactions into zero-shot sim-to-real visuomotor policies. It uses morphometric optimization to kinematically retarget motions across hand morphologies while preserving contacts, followed by residual reinforcement learning to refine dynamics using object pose and contact data. The framework was tested across three robot hands and ten human-object interactions, achieving at least 8-point improvements in contact F1 over five baselines and up to 35-point gains in dynamic retargeting success. Zero-shot visuomotor policies reached 89.3% success in 300 real-world trials on 30 objects.
Key points
- The framework uses morphometric optimization to kinematically retarget human motions across different hand morphologies while preserving demonstrated contact points.
- Residual reinforcement learning refines the kinematic reference by incorporating object pose and contact information from the human motion to ensure dynamic feasibility.
Context and limitations
Results are limited to three robot hands and ten human-object interactions; zero-shot success rates are measured only on 30 objects across 300 real-world trials.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 28/09/2026, 02:22:16 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeifv5ctumn7543tipai7touzxmqmb3rqx7o3cd7vzaqxrt6xqaapr4