V3I5P62

Proactive Risk Assessment in CCTV Using Hand Gesture Recognition

Annie Roshal R^1*, Gods Milani C², Ashmin Shibia S³

Abstract

The rapid growth of crimes and suspicious activities in public places has increased the demand for intelligent surveillance systems capable of performing real-time monitoring and automated threat detection. Traditional CCTV surveillance systems mainly focus on recording video footage and require continuous manual observation by security personnel. Human monitoring often leads to fatigue, reduced concentration, delayed response, and missed incidents during emergency situations. To overcome these limitations, this paper proposes a proactive risk assessment system using hand gesture recognition for intelligent CCTV surveillance applications.

The proposed system uses Artificial Intelligence, Computer Vision, OpenCV, MediaPipe, NumPy, and Python technologies to automatically detect suspicious and aggressive hand gestures from live CCTV footage. The system continuously analyzes video frames, identifies hand landmarks, and tracks movement patterns in real time. Whenever aggressive gestures or abnormal activities are detected, the system immediately generates alerts and warning notifications to improve emergency response and public safety.

The proposed model improves surveillance automation, reduces human monitoring effort, enhances detection accuracy, and supports proactive threat prevention. Experimental analysis demonstrates that the system performs efficiently under different environmental conditions and provides reliable real-time aggression detection for modern surveillance systems.