Perceptron AI launches EgoHome, an open benchmark for sub-second egocentric hand understanding
rohanpaul_ai · x · 2026-11-11
Perceptron AI introduced EgoHome, an open benchmark for dense, sub-second egocentric understanding of household-chore videos, where GPT-6 Astra came out on top. Every clip carries four annotation types — hand boxes, subtasks, hand actions, and hand captions — with 750+ segments under 500ms. The benchmark matters because robot training increasingly relies on human videos auto-labeled by AI ("left hand picks up sponge, right hand turns tap"); wrong labels mean robots learn wrong skills. Data, metrics, and judge prompts are all open.
More from Embodied
- CoRL results show sim-only training now delivers top real-world robot performance — carlosdponx · 2026-11-12
- Robot fails at folding a towel, reassuring roboticists frontier models won't eat robotics yet — chris_j_paxton · 2026-11-12
- Nuro podcast: manufacturing, charging, and regulation gate robotaxi mass adoption — R_D · 2026-11-12
- Rodeo Robotics demos robot that changes your car tire — ycombinator · 2026-11-12
- Cybercab rear-ended by human driver, researcher quips: the problem is the humans — chris_j_paxton · 2026-11-12
- HT-Policy Uses Heavy-Tailed Student-t Regression to Close Robot Policy Gap — chris_j_paxton · 2026-11-12