This project is a real-time women safety monitoring system built using Python, OpenCV, and MediaPipe. It analyzes human behavior and silent hand gestures through a webcam to detect stress, danger, and SOS situations without requiring voice input or external devices.
Webcam captures live video.
Body pose, hands, and face landmarks are detected.
Movement intensity is analyzed over time.
System classifies behavior into:
NORMAL – calm movement
STRESS – repeated unstable motion
DANGER – sustained panic motion
Silent hand gestures can manually trigger:
HELP or SOS
Output is shown in real time and recorded automatically.