Driver Drowsiness Recognition System

Model is successfully able to identify "awake" state
Labels
Correlogram
Training data

Inspiration

Everyone who drives is susceptible to fatigue. We wanted to create a system that is able to identify if a driver is drowsy to increase road safety and help reach GM's target of zero crashes.

What it does

Identifies whether a driver is drowsy or awake.

How we built it

YOLOv5 model and OpenCV.

Challenges we ran into

A lot.

Accomplishments that we're proud of

A lot.

What we learned

A lot. We were inspired by Nicholas Renotte's tutorial on using YOLOv5 (https://www.youtube.com/watch?v=tFNJGim3FXw)

What's next for Driver Drowsiness Recognition System

This has many implementations behind the wheel, and we can expand from drowsiness to distracted driving as well as driving under influence. Facial recognition and recognition of emotional state also has applications to other industries, for example integrating AI with healthcare.