Research on Underwater Optical Image Target Detection and Network Security Protection Method based on Deep Learning
Year: 2025
Type: article
Abstract: This paper studies underwater optical image target detection and network security protection methods based on deep learning. The YOLOv7-based method is used to solve the color distortion and blur caused by the scattering and absorption of light. By replacing the ELAN structure in the backbone network with the NewDSCLayer structure, the ELAN structure in the detection head module with the GSCLayer structure, and introducing the SimAM attention mechanism, the detection accuracy of overlapping targets and small targets is effectively improved, and the detection speed is also improved. In addition, this paper also designs an adaptive network security protection strategy based on deep reinforcement learning, which can identify multiple network threats through behavior analysis and feature matching, and dynamically adjust defense measures. Experiments have shown that the algorithm proposed in this paper has higher accuracy, faster response speed and higher system stability, and can provide efficient protection for network security.
Source: Computer Fraud & Security
DOI: 10.52710/cfs.474
Author: KuoChai, HeyuWang
Institution: Liaoning Technical University
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4409435438