1 paper · 1 filter
Jiamin Xu, Cong Wang, Zheng Dong +4
The paper introduces WildShadowRemover, a system that fine‑tunes a pretrained video diffusion model to remove shadows from real‑world videos while preserving fine details and tempo…