Unable to detect Iris boundary using imfind circles

Canny_img = edge(newI,'canny');
image(Canny_img)
title('Edge detected')
figure,
imshow(Canny_img)
title('Circle Detected Image')
[centers, radii] = imfindcircles(Canny_img,[20 150],'Sensitivity',0.8);
disp(centers)
disp(radii)
centersStrong = centers(:,:); 
radiiStrong = radii(:);
%metricStrong5 = metric(1:2);
viscircles(centersStrong, radiiStrong,'EdgeColor','b');

this is my code it detects the inner circle not the outer one. pls help

[Jonas]

did you try to use a bigger lower radius bound in imfindcircles? if the lower bound is big enough you will not get the small circle.

it would also help if you provide the original image or the binary image without the circle

[Karthik Krishnan]

thanks

Answer 1

[Matt J]

Increasing the lower radius bound as suggested by @Jonas as well as increasing the sensitivity setting seems to help:

A=any( imbinarize(imread('https://www.mathworks.com/matlabcentral/answers/uploaded_files/690953/image.png')) ,3);

[centers, radii] = imfindcircles(A,[50 150],'Sensitivity',0.96);

imshow(A);
viscircles(centers, radii,'EdgeColor','b');

[Jonas]

parameters for the more outer bound can be e.g.

imfindcircles(Canny_img,[110 150],'Sensitivity',0.98);

[Karthik Krishnan]

why u used imbinarize? @Matt J

[Matt J]

Because your posted image looks binary. It might indeed be better to try to detect the circles in the original color image.