Back to Projects
// Personal project
DrowzeeGuard
I built a computer vision safety system that watches a driver-facing camera feed and automatically detects drowsiness, triggering an instant alarm.
Tech Stack
PythonYOLO (Ultralytics)OpenCVReal-Time Computer Vision
How it works
Reaction speed mattered more than anything else here.
- Built the detection pipeline on a YOLO object detection model chosen specifically for its balance of speed and accuracy on live video.
- Made the system react to a lapse in attention within seconds instead of analyzing footage after the fact.
Why I built it
Driver drowsiness is a real safety risk, and by the time it's caught after the fact, it's too late to matter. This needed to detect and alert in real time, not analyze footage afterward.
How I approached it
I built the detection pipeline on YOLO specifically because it balances speed and accuracy well enough to run on live video and trigger an alarm within seconds of detecting a lapse in attention.
