E
Eye Blink Detection Module
A Computer Vision project that detects the eye blink using the 468 face landmarks obtained using the Mediapipe Library.
22
Created on 13th March 2022
•
E
Eye Blink Detection Module
A Computer Vision project that detects the eye blink using the 468 face landmarks obtained using the Mediapipe Library.
The problem Eye Blink Detection Module solves
I was able to successfully:
- Differentiate between a normal blink and a tired eye closure by keeping a note of the frequency and duration of these blinks.
- Find an optimal value for the number and duration of these eye blinks which helps us to have an accurate measure of when the person starts feeling drowsy.
PS: This module can act as a good fit to be integrated to several large-scale and can be modified as per the need.
Challenges I ran into
The challenge with this module was to have an accurate measure of the number of eye-blinks and the frequency of these blinks which would help me differentiate between a normal eye closure and a tired eye closure
This challenge was overcome with continuous trial and error of the eye-blink detection module on a group of people. The threshold value for each one of them was recorded and a cumulative measure suggested to me the value which should be finally taken for this module.
Technologies used
Discussion
Builders also viewed
See more projects on Devfolio