Iris segmentation using Daughman's Algorithm
I used CASIA Data set. You can download from Here
Argument
- r: Radius of iris ex) --r 60
- e: extension of image format ex) --e bmp
- p: path of root image folder i.e: python main.py --r 100 --p ~/Downloads/CASIA-IrisV4(JPG)/CASIA-Iris-Lamp/001/L --e jpg
Requirement library
- OpenCV
- Numpy
Process
Image preprocess and reduce the target pixel, finally adopt Daughman's algortihm for finding IRIS & Pupil.
In image preprocess, I used threshold and mopology for alleviating light reflections.
[mopology]
[binary]
And then, used crop and local minima for reduce target pixel
[crop]
#Daughman's algorithm
It scanned [0:x][0:y][0:r] on image and use this expression
##Reference [1] https://www.diva-portal.org/smash/get/diva2:831173/FULLTEXT01.pdf