SpaceFlow: Locally Controllable 3D Generation

Key points

• Current 3D generation methods lack explicit local control: geometric adherence is often defined by a global control strength, and appearance cannot be specified locally.1
• During structure generation, we enforce these spatial constraints within the generative flow process.1
• Regional geometry metrics demonstrate that SpaceFlow preserves the specified geometry in high-control regions and enables plausible shape variation in low-control areas.1
• The project page is available at SpaceFlow3D.github.io.1

Sources (2)

• [1]SpaceFlow: Locally Controllable 3D Generation
  Hugging Face Daily Papers · Oct 8, 12:00 AM

  We present SpaceFlow, a training-free pipeline for locally controllable 3D generation from text descriptions and a collection of geometric primitives.

  Current 3D generation methods lack explicit local control: geometric adherence is often defined by a global control strength, and appearance cannot be specified locally.

• [2]SpaceFlow: Locally Controllable 3D Generation
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 8, 05:42 PM · same content

Extractive summary: sentences quoted from the sources.