Underwater image enhancement via cross-wise transformer network focusing on pre-post differences

Zhixiong Huang, Jinjiang Li, Xinying Wang, Zhen Hua, Shenglan Liu, Lin Feng

Abstract: Highlights•We introduce a novel cross-wise transformer module (CTM), which employs cross-attention to capture dependencies.•We devise a feature supplement module (FSM) that employs an information complementary mechanism.•Our proposed network shows outstanding results across multiple datasets and effectively handle various visual tasks.