Drowsiness Detection

Overview

A real-time driver monitoring system that uses facial landmark analysis to detect drowsiness through Eye Aspect Ratio and Mouth Aspect Ratio signals, triggering instant audio alerts.

Problem

Driver fatigue remains a critical safety risk. The system needed low-latency visual cues that could detect drowsiness without specialised hardware.

Research and requirements

Facial landmark geometry provided a practical route to estimate eye closure and yawning patterns for continuous monitoring in cabin-like conditions.

Solution

Implemented a Python and OpenCV pipeline with facial landmark detection, EAR/MAR thresholds and real-time audio alerts for fatigue events.

Features

• Real-time facial landmark tracking
• Eye Aspect Ratio based drowsiness cues
• Mouth Aspect Ratio yawn detection
• Instant audio alerts
• ADAS-oriented safety concept

Challenges

• Maintaining detection stability under changing lighting
• Tuning thresholds to reduce false positives

Results and impact

• Demonstrated practical CV-based fatigue detection for driver safety concepts
• Provided a foundation for ADAS-oriented alert systems

Gallery