OpenMV Camera Eye opening

Hello everyone,

I am trying to measure the eye aspect ratio by using an OpenMV camera. I want to use that data for calculating the Drowsiness rate. Has anyone worked related to that? Is it possible to do it using OpenMV?

Best regards.

Hey! This is something I wanted to build a few years ago when I got my first OpenMV.
I was thinking this would be useful for people driving a lot.
My thoughts :
You can’t just sit someone down and say They are or are not drowsy. You would have to compare an initial set of frames to a future set of frames, that same day.

You may be able to track blinking as an indicator of drowsiness.

You may be able to track irritability as an indicator - eg. twitchiness or constantly moving

I think you need to run some tests and collect real images of people before they are drowsy and then after.
I think you could tell the difference between when someone starts driving vs a couple hours later.
But I don’t think you could sit someone in front of the camera and immediately say they are drowsy because you don’t have a baseline of that persons normal behavior.
You could train a small Machine Learning set once you have some data to detect general tiredness in people.

Humans are tough this way. I’d be interested in seeing what you do with it.

Hey Kevin!

I made huge research about it and I even talked with a neuropsychologist about this. You are right about everything, but I have a plan and all I have to find is if I can do it without using some libraries like “cv2,scipy” because basically, we cannot use those in MicroPython.

So I was wondering if someone tried it already or not but you were the only one who responded about this topic. I am still looking for it and I will keep you updated if you want :slight_smile: