v1.0.0

Driver Monitoring System

A real-time driver monitoring system implementing Euro NCAP requirements for driver state detection using computer vision and machine learning.

Key Components

1. PoseEstimation.py

• Main processing pipeline for facial analysis
• Uses MediaPipe Face Mesh for facial landmark detection
• Implements:
  • Eye Aspect Ratio (EAR) tracking
  • Mouth Aspect Ratio (MAR) for yawning detection
  • Head pose estimation (pitch, yaw, roll)
  • PERCLOS (percentage eye closure) calculation
  • Real-time visualization

2. DecisionLayer.py

• Implements Euro NCAP compliant driver state assessment
• Features:
  • Multiple driver states (Alert, Drowsy, Microsleep, etc.)
  • Visual Attention Time Sharing (VATS) monitoring
  • Gaze strategy detection (Owl vs Lizard patterns)
  • Risk level assessment
  • Progressive response system

3. Config.py

• Centralized configuration management
• Contains:
  • MediaPipe parameters
  • Facial landmarks indices
  • Detection thresholds
  • Visualization settings
  • Camera parameters

Key Features

1. Driver State Detection

   • Alert/Drowsy state monitoring
   • Microsleep detection
   • Distraction monitoring
   • Phone use detection

2. Head Pose Analysis

   • 3D head pose estimation
   • Gaze direction tracking
   • Looking away detection

3. Drowsiness Metrics

   • PERCLOS calculation
   • Blink rate monitoring
   • Yawn detection
   • Combined drowsiness assessment

4. Visualization

   • Real-time facial landmark overlay
   • Head pose direction indicator
   • Status dashboard
   • Color-coded warnings

Technical Details

• Framework: OpenCV, MediaPipe
• Language: Python 3.x
• Key Dependencies:
  • opencv-python
  • mediapipe
  • numpy
• Performance: Real-time processing (30+ FPS on modern hardware)

Implementation Notes

• Uses rolling windows for temporal analysis
• Implements hysteresis for state transitions
• Euro NCAP compliant warning systems
• Modular architecture for easy extension