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]

[local minima]

And result is this.

[okay]

[error]

#Daughman's algorithm

It scanned [0:x][0:y][0:r] on image and use this expression

[expression]

##Reference [1] https://www.diva-portal.org/smash/get/diva2:831173/FULLTEXT01.pdf