Date: Nov 15, 2012 12:30 AM
Author: M. Masud Reza
Subject: Optimum value for scales Iris in Libor masek code for Recognition

I am undergraduate student of EEE  and interested in iris recognition and  trying to simulate libor masek code.

what is the optimum value for 'scales' in
'hd = gethammingdistance(template1, mask1, template2, mask2, scales)'
function. For different persons minimum hamming distance (hd) occurs for different 'scales' value. when i set a threshold value for 'scales'=1 or 5 more than one person is detected and sometimes false person is detected.

I have created template by [template, mask] = createiristemplate(eyeimage_filename)
for both left and right eye for 60 persons and saved the data in two mat file.
I create another template from individual left and right eye (i.e. person 50) and load the both mat file find out hd by hd = gethammingdistance (...) individually for left (hd_l) and right (hd_r) eye from corresponding database. Then
run
[C_l I_l]=find(hd_l<.30);

[C_r I_r]=find(hd_r<.30);

person=intersect(I_l,I_r)