Real-Time Gesture-Based Emergency Assistance System Using MediaPipe and Artificial Neural Networks

The purpose of this study is to design an emergency assistance system based on real time gestures and gesture based emergency response systems. The goal is to create a mechanism for conveying an emergency in a non contact manner through the use of established gesture signals. As a result of the possibility that the majority of people may have difficulty accessing verbal communication, written communication, and or other standard types of communication to request emergency assistance, the need for a non contact gesture method is essential. People may not be able to use verbal or traditional methods of communicating due to physical limitations, noise from the environment, social isolation, or situational constraints. Therefore, the proposed system will utilize a traditional webcam to capture video in real time and use Computer Vision and Deep Learning techniques to detect and classify established emergency gestures. This system will use MediaPipe for accurate hand landmark extraction and will require the use of an artificial neural network for detection of gestures in real time. Once an emergency gesture has been detected with a sufficient level of confidence, an alert will be generated by the system, and the monitoring interface in which the alert is generated will be updated in order to allow for a timely response to the emergency gesture. This system is intended to be lightweight, inexpensive to deploy, and easily deployable across multiple public places, campuses, workplaces, and smart environments, thereby improving the safety and efficiency with which individuals are able to respond in the event of an emergency situation.