Related Experiment Video
Updated: Aug 8, 2026
03:31
End-To-End Deep Neural Network for Salient Object Detection in Complex Environments
Published on: December 15, 2023
Underwater multi-sensor information fusion for salient object detection
Yan Mou1, Zhaolong Gao1, Jinjiang Li1
1School of Computer Science and Technology, Shandong Technology and Business University, Yantai, Shandong, China.
Plos One
|August 6, 2026
Summary
This study introduces a new framework for underwater salient object detection using an information cross-fusion network. The method effectively fuses RGB and depth data to improve accuracy and reduce noise in underwater environments.
Area of Science:
- Computer Vision
- Image Processing
- Robotics
Background:
- Underwater salient object detection is crucial for various applications like object tracking and recognition.
- Challenges in underwater imaging include multimodal information fusion and noise reduction.
Purpose of the Study:
- To develop a novel framework for underwater salient object detection.
- To improve multimodal feature aggregation and refinement using RGB and depth data.
Main Methods:
- Introduced an information cross-fusion network with cross-attention feature injection and information embedding modules.
- Modeled complementarity between RGB and depth data at global and local scales.
- Employed multi-level and multimodal information fusion to mitigate noise and uncertainty.
Main Results:
- The proposed method achieved superior performance on multiple underwater datasets.
- Demonstrated enhanced representation of salient regions and effective background noise suppression.
- Highlighted efficacy in multimodal feature integration for salient object detection.
Conclusions:
- The novel framework effectively addresses challenges in underwater salient object detection.
- The approach shows significant improvements over state-of-the-art methods.
- The method is highly effective for multimodal feature integration in underwater scenarios.