Research on the Application of Deep Learning-Based Computer Vision in Anomaly Detection in Communication Networks

Network anomaly detection has become an important task for ensuring network security and stable operation. However, due to the increasing scale and complexity of communication networks, traditional detection methods have gradually shown limitations in handling large amounts of network data. This paper investigates the application of deep learning-based computer vision technology in communication network anomaly detection, aiming to enhance the accuracy and efficiency of anomaly detection through innovative algorithm design. To this end, this paper proposes a multi-layer deep model (CLA) that combines CNN and LSTM to efficiently capture the spatial and temporal features of network traffic images. The model converts network traffic data into image representations, utilizes CNN to extract spatial features, and then combines LSTM to capture correlations in the time series, significantly improving the accuracy of anomaly detection. Furthermore, this paper introduces an adaptive attention mechanism to focus on key features, effectively reducing noise interference and false positive rates. We did an experiment in network scenarios, validating its feasibility and advancement in practical applications, while also providing new ideas and solutions for the security protection of future communication networks.