Eve-teasing detection from video footage using computer vision and artificial intelligence

Abstract

We present computer vision approaches combined with machine learning techniques to detect eve-teasing from any video material, which may be used in any situa tion. Eve teasing is a colloquial term for public sexual harassment or sexual assault committed primarily against women by men. Eve-teasing is a common and pro foundly distressing reality for young women, particularly young girls. Moreover, in Bangladesh, the number of reported occurrences of Eve-Teasing is increasing at an alarming rate. Already eve-teasing is leading to rape and then murder. Due to a lack of proof and adequate supervision, the vast majority of criminals can get away with their crimes. In order to eliminate this problem, we suggested computer vi sion approaches for detecting eve-teasing in any video clip, which may be used in any critical situation. Our proposed method uses machine learning and computer vision to detect human Gender, expression, posture, and gesture and combine them to verify a matching human behavior in a critical situation to justify. When em ploying the combination of male-female identification, human behavior detection, and CCTV-based monitoring systems or video footage, the system can identify such vital conditions in our suggested method. Also included is the ability to determine who is participating and who is not in the scenario. Women who want to prove eve-teasing or harassment may find the procedures proposed helpful while ensuring the actual offender is punished.