An Improved UNet Algorithm Based on Multiscale Features and Attention Modules for Underwater Salient Multi-Target Detection

Haibin Han, Honghao Yan, Haodi Zhu, Na Wei, Shaojian Yang, Zhenyu Ge, Xingqian Zhu, Fengzhong Qu
June, 2025
Cite

Abstract

This paper improves UNet for underwater salient multi-target detection by combining multiscale feature extraction and attention modules. The method targets the difficult visual conditions common in underwater scenes, where object scale variation, turbidity, and background interference make reliable salient object detection challenging.

Type
Conference paper
Publication
In OCEANS 2025 - Brest
Underwater object detection Salient object detection UNet Attention mechanism
Shaojian Yang
Shaojian Yang
Assistant Research Fellow / Postdoctoral Fellow

My research interests include underwater communications, distributed acoustic sensing and ocean Internet of Things.