IllumiCraft: Controllable Video Generation Dataset for Geometry and Illumination

IllumiCraft is a dataset released with a research paper on unified geometry and illumination diffusion for controllable video generation. The dataset was created by Yuanze Lin, Yi-Wen Chen, Yi-Hsuan Tsai, Ronald Clark, and Ming-Hsuan Yang. It was last updated on the Hugging Face platform in June 2026.

Use Cases

• Training controllable video generation models based on geometry and illumination concepts.
• Benchmarking video synthesis algorithms based on the described unified diffusion framework.
• Researching disentangled control of scene geometry and lighting in video generation.

Strengths

• Dataset is associated with a peer-reviewed research paper published on arXiv.
• Dataset is hosted on a major AI platform (Hugging Face) with a public project page and code repository.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file size are unknown, which may limit suitability assessment.

Provenance

Source

YuanzeLin

Collection Method

Released with the IllumiCraft research paper.

Freshness

Last updated 2026-06-07 03:41:07; freshness should be verified.