sclera based authentication project.

i am doing project on sclera based authentication , first step of project is to detect region of interest ,in my case region of interest is sclera part(white),i used following code to select region of interest but the problem is that i am not getting desired output also the output which i got has noise which i tried to remove using filter which made my image blurry. i am attaching my code , i am attaching two images

A.output which i got

 B.desired output

if true
clc;
clear all;
close all;
I = imread('yo4.jpg');
imageSize = size(I);
row=imageSize(1);
col=imageSize(2);
figure,subplot(3,2,1);
imshow(I);
title('ORIGINAL IMAGE');
I1=rgb2gray(I);
subplot(3,2,2);
imshow(I1);
J=I1;
for i=1 :row
  for j=1 :col
      if (J(i,j)>180 )
          O(i,j)=J(i,j);
      end;
  end;
end;
H=fspecial('average',[5 5]);
O=imfilter(O,H,'replicate');
for i=1 :row
  for j=1 :col
      if (O(i,j)>180 )
          Ex(i,j)=I(i,j);
      end;
  end;
end;
subplot(3,2,3);
imshow(Ex);    
title('region of interest');
K=medfilt2(Ex,[13 13]);
subplot(3,2,4);
imshow(K);  
title('after applying median filter');
end

what changes i should make to get desired output.

Answer 1

[Image Analyst]

Well first of all, you can extract the biggest blob with my attached function.

Next you need to tell us exactly where you want to split the largest blob into two parts - it's not obvious to us where it should be split.

And of course the two images are of completely different eyes - the vein patterns and eyelashes are different!

[Star Strider]

I doubt sclera (or more accurately, conjunctival vessel) based identifications would be robust in practice. See the Wikipedia article on Conjunctivitis for a rather obvious reason.