CRFormer: A cross-region transformer for shadow removal

Open Webpage

Jin Wan, Hui Yin, Zhenyao Wu, Xinyi Wu, Zhihao Liu, Song Wang

Published: 01 Jan 2024, Last Modified: 13 Nov 2024Image Vis. Comput. 2024EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•CRFormer adeptly restores image signals damaged by shadows.•A hybrid CNN-Transformer model is proposed to exploit regional context for shadow removal.•A region-aware cross-attention is proposed to aggregate non-shadow features into shadow regions.•Superior performance can be achieved in both image and video shadow removal tasks.