World2HOSI: A Multi-Expert Harness for Long-Horizon Human-Object-Scene Interaction Generation

Key points

• Existing human-object interaction (HOI) and human-scene interaction (HSI) models excel in their respective domains, but none alone covers tasks requiring multiple interaction skills.1
• We introduce World2HOSI, a multi-expert orchestration harness that generates continuous HOSI data from 3D Gaussian Splatting (3DGS) scenes and natural language instructions.1
• World2HOSI realizes this insight as a closed loop around a persistent scene state built from decoupled background and object assets.1
• Experiments and a perceptual user study show that World2HOSI achieves the best task completion and interaction quality on the evaluated HOSI tasks, while generating more diverse data.1

Sources (1)

• [1]World2HOSI: A Multi-Expert Harness for Long-Horizon Human-Object-Scene Interaction Generation
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 20, 06:43 AM

  Long-horizon human-object-scene interaction (HOSI) is central to embodied AI, animation, and game development.

  Existing human-object interaction (HOI) and human-scene interaction (HSI) models excel in their respective domains, but none alone covers tasks requiring multiple interaction skills.

Extractive summary: sentences quoted from the sources.