Learning Dynamic Structural Specialization for Underwater Salient Object Detection

2026-05-14

AI Summary

This work addresses the challenges of inaccurate salient object localization, fragmented regions, and coarse boundaries in underwater images caused by severe degradation. To tackle these issues, the authors propose DSS-USOD, a method that extracts a shared foundational representation from a single underwater RGB image and dynamically decomposes it into two specialized branches—one sensitive to boundaries and the other promoting region consistency—via a dynamic structure specialization mechanism. A spatial coordination module and collaborative structural supervision are integrated to adaptively fuse local structural context. The proposed approach achieves state-of-the-art performance across multiple benchmark datasets and demonstrates practical efficacy on a real underwater robotic platform, significantly enhancing both boundary precision and regional coherence.

Abstract

Underwater salient object detection (USOD) has attracted increasing attention for underwater visual scene understanding and vision-guided robotic applications. However, existing USOD methods still struggle with underwater image degradations, which often lead to inaccurate object localization, fragmented salient regions, and coarse boundary prediction. To address these challenges, this paper proposes DSS-USOD, a novel RGB-based USOD method built upon dynamic structural specialization. DSS-USOD extracts a shared base representation from a single underwater image, decomposes it into boundary-sensitive and region-coherent structural features, and dynamically coordinates their contributions according to local structural context. Specifically, the extracted shared base representation is decomposed into a boundary-sensitive branch for modeling fine-grained boundary details and a region-coherent branch for capturing region-level structural consistency. A spatial coordination module is then introduced to adaptively regulate the relative contributions of the two branches according to local structural context. Moreover, cooperative structural supervision is introduced to promote branch specialization and stabilize spatial coordination, enabling DSS-USOD to better balance boundary precision and region coherence under degraded underwater conditions. Extensive experiments show that DSS-USOD achieves superior performance on benchmark datasets. Finally, real-world deployment on an underwater robot validates the practical effectiveness of DSS-USOD for underwater object inspection.

Problem

• Underwater salient object detection
• image degradation
• object localization
• boundary prediction
• region coherence

Innovation

• dynamic structural specialization
• underwater salient object detection
• boundary-sensitive features
• region-coherent features
• spatial coordination module