Learning Tactile-Aware Hand Representations from Human Hand-Object Interaction Data
To this end, we propose HuTact, a framework that pretrains tactile-aware hand representations from human HOI data without requiring real tactile signals during pretraining.
ProofPaper ↗
Key points
- Tactile representation learning for dexterous manipulation often requires substantial robot-side data collection, limiting its scalability across different hands and sensors.
- We explore whether large-scale human hand-object interaction (HOI) data can provide a useful prior for robotic tactile learning.
- HuTact transfers the interaction structure learned from human hand motions to robotic tactile observations and can be further adapted with a small amount of unlabeled robot interaction data.
- We evaluate HuTact on tactile perception and manipulation tasks, including grasp prediction, slip detection, object classification, and tactile-conditioned policy learning.
Sources (1)
- [1]Learning Tactile-Aware Hand Representations from Human Hand-Object Interaction DataarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 20, 10:31 AM
To this end, we propose HuTact, a framework that pretrains tactile-aware hand representations from human HOI data without requiring real tactile signals during pretraining.
Tactile representation learning for dexterous manipulation often requires substantial robot-side data collection, limiting its scalability across different hands and sensors.
Extractive summary: sentences quoted from the sources.