Visual Analysis of Computer Security Data Using Artificial Neural Networks

Artificial neural networks are used in various fields of human activity, including information and computer security, for example, in biometric systems such as face and fingerprint recognition. However, in the field of visual analysis of security incident data, it is assumed that the visualization is intended for the human operator since the person makes the final decision. This paper investigates whether computer vision systems can perform the tasks of a human operator during visual data analysis. To study the question posed, the paper proposes a model architecture for classifying visualizations of traffic with anomalies and normal traffic. Experiments were also carried out using a neural network and involving subjects. Experiments show that a trained neural network can recognize anomalies in images with visualization of security events and, in some cases, even better than a human.