IoT Based Face Recognition Using Machine Learning for Women Safety

Abstract

Despite two decades of attention to stalking in developed countries, the issue remains understudied in India, where violence against women is higher. Few reported cases receive little notice, but victims face significant financial, social, and mental losses. Research indicates stalking often precedes sexual offenses and murder, highlighting the potential to reduce overall violence against women by preventing stalking at its early stages. In this paper we have implemented women safety system on Raspberry pi via Ultrasonic sensor and Pi camera. The usage of earlier systems in daily life was both expensive and time-consuming. The voice alert safety system (subject) used by a women is described in this study. This system recognises humans using an ultrasonic sensor network. It accurately calculates the separation between the user and other human come in danger zone (distance measure 20 cm from device) or out of danger zone (distance measure more than 20 cm). It identified humans come in danger zone and give verbal feedback to let the user know. Such voice messages are delivered to the subject using the speaker. It also uses camera and facial recognition algorithms to detect faces and recognize the person and inform the user through audio output. The intention of this  research work is to create a cheap, portable voice alert safety system for women’s.