Camouflaged Object Detection via Dual-branch Fusion and Dual Self-similarity constraints

Haozhe Yang, Yuan Zhu, Ke Sun, Haoyang Ding, Xianming Lin

Abstract: Highlights•Camouflaged objects hide in environment with high similarity and invisibility.•Self-similarity information helps identify subtle differences of camouflage.•Multi-angle feature fusion handles scenarios with multiple small objects.•Self-similarity constraint improves performance as a plug-and-play module.•Multiple refinement in inter-group interaction contributes to precise localization.