Eve-Teasing Detection from Video Footage Using Computer Vision and Artificial Intelligence

Eve-teasing, a pervasive form of public harassment targeting women, remains a significant social issue in Bangladesh, often leading to serious crimes and distress among young women and girls. Despite the widespread use of surveillance, effective methods for detecting such incidents in real-time are lacking, allowing many offenders to escape due to insufficient evidence. This paper proposes a novel approach leveraging computer vision and machine learning to detect instances of eve-teasing in video footage, focusing on gender classification, expression analysis, and gesture recognition. Our system integrates gender identification, human behavior analysis, and CCTV-based monitoring to detect critical behaviors indicative of harassment. By accurately identifying potential offenders and victims, the proposed model provides valuable evidence, supporting law enforcement in apprehending offenders and enhancing public safety. This solution offers a promising tool for real-time harassment detection, contributing to a safer public environment.