Haiwen (Haven) Feng (@HavenFeng)
2025-06-12 | โค๏ธ 108 | ๐ 20
๐ Introducing GenLit โ Reformulating Single-Image Relighting as Video Generation! We leverage video diffusion models to perform realistic near-field relighting from just a single imageโNo explicit 3D reconstruction or ray tracing required! No intermediate graphics buffers, directly in the pixel space! ๐ Dive into the paper: https://arxiv.org/abs/2412.11224 ๐ฅ Project page & demos: https://t.co/fbMBoEukXN ๐ Code coming soon! #GenerativeAI ComputerVision Relighting DiffusionModels Graphics ๐งต 1/5
๋ฏธ๋์ด
![]()
์์ฝ
GenLit์ ๋จ์ผ ์ด๋ฏธ์ง ๋ฆฌ๋ผ์ดํ ์ ๋น๋์ค ์์ฑ ๋ฌธ์ ๋ก ์ฌ์ ์ํด, 3D ๋ณต์ยท๋ ์ด ํธ๋ ์ด์ฑ ์์ด ํฝ์ ๊ณต๊ฐ์์ ์ฌ์ค์ ์ธ ๊ทผ๊ฑฐ๋ฆฌ ์กฐ๋ช ๋ณํ๋ฅผ ํฉ์ฑํ๋ ๋ฐฉ๋ฒ์ ์ ์ํ๋ค. ์ฒจ๋ถ ์ด๋ฏธ์ง์ฒ๋ผ ๊ฐ์ ๋ฌผ์ฒด/์ฅ๋ฉด์ ๋ค์ํ ์กฐ๋ช ์กฐ๊ฑด์ผ๋ก ์์ฐ์ค๋ฝ๊ฒ ์ฌ์กฐ๋ช ํ ๊ฒฐ๊ณผ๋ฅผ ๋ณด์ฌ์ค๋ค.
Tags
domain-vision-3d domain-rendering domain-genai domain-ai-ml domain-dev-tools