Smart Surveillance with Hand Gesture Detection for Silent Emergency Alerts

Student safety in our learning environments is challenged with ragging, mental heath challenges and medical emergencies. Pre-existing types of monitoring like audio surveillance and wearable devices are frequently versatile because of concerns with privacy, false alarms, and outlets. This paper proposes Silent Alert, a real-time gesture-based emergency detection system that can use computer vision and machine learning to recognize discrete distress.To accomplish that, the system uses MediaPipe to detect the locations of hands and OpenCV to process the live video. Several classifiers such as Random Forest, SVM, KNN, DNN and CNN+LSTM are trained upon a custom hand gesture dataset, as well as how to effectively integrate the deep learning-based YOLOv8 network in finding the gestures. When distress signals are detected, Silent Alert records image frames and indicates to the authorities via secure email alerts so that immediate action can be taken. Revolving around very low intrusiveness and being able to easily meld into the current campus infrastructure, Silent Alert presents a scalable solution with a minimal privacy overhead that increases the effectiveness of emergency response in educational environments as well as other crucial spheres.