My requirement is to use computer vision to determine the size of an a closeup image of a pupil (for accuracy purposes the pupil needs to take up 80% of the image).
I can’t use a reference object in the image.
I could use the LIDAR in the new iPhone (unreleased) but the people using this product will not all have the latest iPhone so a more mainstream solution is required.
Maybe there is a published scale standard that I’m missing? Maybe there’s an ML dataset of eyes and measurements that I’m missing? Maybe I’m missing some other obvious solution?
Any help/input would be greatly appreciated :)