IGen is a scalable data generation system for robot learning from open-world images. Instead of relying on labor-intensive teleoperation, IGen takes diverse images from the real world and automatically synthesizes robot trajectories for various tasks. With these generated data, we can train robot policies that transfer to real robots and perform effective manipulation in the real world — without any human teleoperation data.
- 🌍 Open-World Image Processing: Convert diverse 2D images into actionable 3D scenes
- 🤖 Visuomotor Data Generation: Generate realistic robot trajectories and observations
- 🧠 LLM-Powered Planning: Transform scene-specific instructions into executable plans
- 📊 Scalable Data Synthesis: Support for large-scale robotic policy training