Inspiration
Accidents due to drivers sleeping on the road is a common phenomenon all over the world and this project aims to alert the driver when he is drowsy based on the features of the driver's face like his eyes and eyebrows.
How we built it
The project includes direct working with the 68 facial landmark detector and also the face detector of the Dlib library. The 68 facial landmark detector is a robustly trained efficient detector which detects the points on the human face using which we determine whether the eyes are open or they are closed.
The project includes direct working with the 68 facial landmark detector and also the face detector of the Dlib library. The 68 facial landmark detector is a robustly trained efficient detector which detects the points on the human face using which we determine whether the eyes are open or they are closed.
- As you can see the above screenshot where the landmarks are detected using the detector.
- Now we are taking the ratio which is described as 'Sum of distances of vertical landmarks divided by twice the distance between horizontal landmarks'.
- Now this ratio is totally dependent on your system which you may configure accordingly for the thresholds of sleeping, drowsy, active.
What's next for OpenCV Project - Driver Drowsiness Detection
Next I would like to implement sound so that it alerts the driver when he dozes off.
Log in or sign up for Devpost to join the conversation.