Shape-Guided Detection: A joint network combining object detection and underwater image enhancement together

Chao Yang, Longyu Jiang, Zhicheng Li, Jie Wu

Abstract: Highlights•We propose a combinatorial network called Shape-Guided Detection (SG Detection), which innovatively combines the tasks of object detection and underwater image enhancement.•We creatively propose a Shape Prior Enhancement module. This module utilizes the interaction between the prior information and the original skip features to effectively fuse the global and local shape information.•We use a combination of explicit and implicit prior constraints in the model to force the model to focus more on the utilization of shape prior information.