Event
Release of the EgoTac visual-to-tactile prediction model
1 reports · 1 sources
Researchers have released the EgoTac model, which can directly predict tactile information from first-person human videos; the work has been accepted to NeurIPS 2026. The model was trained on more than 5.7 million image-tactile pairs, covering continuous force measurements and binary contact. Experimental results reported in the paper show that in-domain average force prediction error is below 0.06N, that it consistently outperforms the current state-of-the-art contact estimator on an out-of-domain contact prediction benchmark, and that it can perform zero-shot prediction on unconstrained real-world videos; scaling analysis indicates that both data diversity and data volume can stably improve performance.
Report timeline
- EgoTac: Predicting In-the-Wild Touch from Egocentric Vision
EgoTac is a generalizable model that predicts rich tactile information directly from egocentric human videos. It was trained on more than 5.7 million image-tactile pairs covering both continuous force measurements and binary contact. Experiments show an average force error below 0.06N for in-domain prediction, and on out-of-domain contact prediction benchmarks it consistently outperforms the current state-of-the-art contact estimator. It also enables zero-shot prediction on unconstrained real-world videos; scaling analysis indicates that both data diversity and data volume steadily improve performance. The work was accepted to NeurIPS 2026.
arXiv cs.RO ·