Wuji open-sources MINT, turning 1,000+ hours of egocentric video into robot training data
CyberRobooo · x · 2026-09-08
Wuji Technology released MINT, an open-source system that converts egocentric human videos (cooking, cleaning, repairs, tool use) into data robots can learn from.
- MINT jointly estimates 3D hand motion, camera motion (with FOV), and world-space trajectories from plain RGB egocentric video, then uses them for robot motion retargeting
- A companion EgoPipeline generates pseudo-labels at scale; 1,000+ hours and 110 million frames already processed
- 1.139B trainable parameters, 4 heads (camera motion, FOV, MANO, hand presence); built with ShanghaiTech, Tsinghua, HKU, Zhejiang University and Wuji
- Project page, paper, model, and open datasets are all public
The author argues Wuji is building dexterous-intelligence data infrastructure, not just selling dexterous hands.
Related event: WuJi Open-Sources MINT to Turn Egocentric Video Into Robot Training Data(3 posts)→
More from Embodied
- Shanghai AI Lab-led team releases 100-hour egocentric wearable dataset for humanoid training — ChuckDBrooks · 2026-10-26
- RealSense D585 Pro AI-Powered 3D Stereo Depth Camera Unboxed for Physical AI — chrismatthieu · 2026-10-26
- Physical world agent swarm demo shared on X — pduan · 2026-10-26
- Looking for a wearable mic with best battery life that feeds recordings to personal agents via MCP — altryne · 2026-10-26
- 'Lamp Swarm': a desk of lamp-shaped robots shows what physical AI work looks like — dee_hw · 2026-10-26
- Tencent's HiSTEVE 1.0: hierarchical VLA agent plays Minecraft with native keyboard and mouse, hits 85% SOTA — _akhaliq · 2026-10-26