AI-MLPythonMediaPipeYOLOv8OpenCV
Driver Drowsiness Detection
Realtime computer-vision system that detects driver fatigue from a webcam feed using facial landmarks and object detection.
A realtime safety system that watches a driver-facing camera and raises alerts when fatigue indicators appear — no special hardware required.
Approach
- Facial landmarks — MediaPipe Face Mesh tracks eye aspect ratio and mouth openness to detect microsleeps and yawning.
- Object detection — YOLOv8 confirms head pose and detects distractions like phone use.
- Pipeline — OpenCV handles capture and frame processing, keeping inference realtime on CPU-class hardware.
Highlights
- Eye-closure and yawn detection tuned with per-driver calibration.
- Escalating alert logic that distinguishes a blink from a microsleep.