← Back to Selected Work
02 / COMPUTER VISION & SAFETY UX
Driver Drowsiness Detection System
A non-invasive, real-time computer vision safety system combining 68-point facial landmarks, Eye Aspect Ratio (EAR), Mouth Aspect Ratio (MAR), and temporal frame analysis to detect fatigue and trigger life-saving alerts.
Computer Vision & Dlib
Python & OpenCV
Mathematical Modeling
IIT Patna Capstone
D
THE MISSION & SYSTEM GOALS
Catch fatigue before the driver loses consciousness.
Driver fatigue causes millions of highway accidents worldwide. This project engineerings a low-latency, non-invasive system using standard webcam hardware. By mathematically monitoring eye closure rates and prolonged yawning patterns across continuous frames, the system detects micro-sleep episodes without causing false-alarm fatigue.
0.21 EAR
Calibrated Eye Aspect Ratio Threshold
> 15 Frames
Temporal Window to Eliminate Normal Blinks
1000 Hz Alert
High-Frequency Multi-Modal Warning Tone