AbusoGAN: GAN-Based AI System for Detecting Abuse, Unlawful Activities, and Suspicious Behavior in Surveillance Video Streams for Security Applications

Abstract

AbusoGAN is a sophisticated surveillance system that applies deep learning techniques to identify abuse, illegal behavior, and unusual activity in video feeds. By integrating Generative Adversarial Networks and Convolutional Neural Networks, it processes security footage and automatically identifies unusual activities. The process includes extracting frames from uploaded videos and processing them using trained AI models to identify unusual activities. A secure web interface is used for user login and provides real-time prediction output. It reduces manual surveillance and improves prediction accuracy, providing a detailed description of activities and video playback for confirmation. KEY WORDS: Deep Learning, Surveillance System, GAN, CNN, Activity Detection, Computer Vision, Abnormal Behavior Detection.