Automated Cataract detection in Images using Open CV and Python— Part 1

Update : Part 2 of this series can be found here

This is gonna be a two part blog series describing a personal project of mine, that I was working on last few days, Automated Cataract detection using Machine Vision.

This thought occurred to me when I took my parents to eye doctors(separate times, of course :) ) and realized the sheer number of people that visit ophthalmologists every day. Now me coming from an relatively urban part of the country have ready access to doctors but I can imagine people in remote areas might not have such easy access. I wondered if we could write something using machine learning to automatically classify a cataract affected eye from an healthy eye and put it in an mobile app for common man to use.

So Something like this :

Notice how it created a bounding box around the cataract when it detected one, but left the healthy eye unmarked

With this in mind there were two ways I could have approached this problem,

• Using Python’s OpenCV module
• Using CNN neural networks.

I tried both :). This first blog is gonna describe in detail the openCV approach and next blog will go over the Kera’s approach. I am also releasing the source code with the hope that someone can take the good work further.