How do i Crop out the Only sclera portion from image of an eye ?

i have used Hough Transform to detect the radius of the Limb & its Centre. after that i used Chen-vese Active Contour to detect the Limb Portion & i got the its mask. but i need Help in subtraction of image so that i can get Only Sclera portion from the Original image of eye.

So help me Get the Only Schlera portion so that the output image have the Sclera with veins ? Thanks.

##Original Image

##Limb Detection with its Position & Radius.

## Active Contour Detection of Limb to Crop the Schlera Out.

##Mask Output after 1000 iteration

I=imread('K.jpg');
L=rgb2gray(I);
BW1 = edge(L,'Canny');
imshow(BW1);
radii = 78:1:92;
h = circle_hough(BW1, radii, 'same', 'normalise');
peaks = circle_houghpeaks(h, radii, 'nhoodxy', 15, 'nhoodr', 21, 'npeaks', 1);
imshow(I);
hold on;
for peak = peaks
    [x, y] = circlepoints(peak(3));
    plot(x+peak(1), y+peak(2), 'g-');
end
hold off
figure;
imshow(I);
a=peak(1);
b=peak(2);
c=peak(3);
% Customerlized Mask
m = zeros(size(I,1),size(I,2));
m(20:120,20:120) = 1;

% Built-in Mask
seg = chenvese(I,'small',1000,0.02,'chan'); % ability on gray image
%-- End 
imshow(seg);

[snehal sule]

what is chenvese? is it function file ?

[Image Analyst]

It's "Chan-Vese" after the authors of an algorithm for active contours. It's implemented in the activecontour() function of the Image Processing Toolbox.

Answer 1

[Image Analyst]

Have you tried a search of the forum http://www.mathworks.com/matlabcentral/answers/?term=sclera&sort=updated+desc

If you want the iris only on that particular image, I'd use imopen() on the binary image to separate it from the rest of the clutter. Or try imfindcircles(). Then call bwareafilt(binaryImage, 1) to extract only the iris since it will be the largest blob.

For finding the sclera, you can check the literature: http://www.visionbib.com/bibliography/contentsmedical.html#Medical%20Applications,%20CAT,%20MRI,%20Ultrasound,%20Heart%20Models,%20Brain%20Models. My guess is that they first convert to HSV colorspace and look for regions with low saturation and high value - that would segment out white regions.

Use the Color Thresholder app on the Apps tab on the tool ribbon. Or see the HSV segmentation demo in my File Exchange: http://www.mathworks.com/matlabcentral/fileexchange/?term=authorid%3A31862

[Dipesh Gupta]

Thanks for Revert, as Search link was really helpful.

i am using chen-vese Active contour to find out the sclera, but i want to use 2 initial seed points which lies outside of the iris on the sclera region from where the contour grow & it stops just at the eye-brows region.

please suggest active contour such that i am able to define the 2 initial seed Points Location ??

[Image Analyst]

Use the color thresholder to find the sclera and tell it to generate/export code. Then tell regionprops() to get the centroid of each region.

[Dipesh Gupta]

Hi Sir, Really Great tool (The color thresholder) work flawlessly on image with different colors, but in case the color in the selected region is existing in the undesired area as well,then it won't give the desired result.

as i have to crop out the sclera i selected the region, the veins whose color is quite similar to the outside skin is also coming after applying threshold.

so please suggest how can i apply active contour with two points

[Image Analyst]

Yes, that's why you have to do some spatial processing too, not just color segmentation alone. Attach your m-file and I'll see what I can do, but I'm pretty busy this week with a course on multivariate curve resolution and some other things so I can't guarantee I'll get to it.

[Dipesh Gupta]

Hi Sir,

Greetings !!

Thanks for your revert. Please find the createmask.m file for the applied Color threshold on the inputimage & also find the My Sclera Recognition Code (Sclera_Recognition_Testcode.m).

InputImage_Color_thresholding.jpg

maskedRGBImage_Color_thresholding.jpg

BW_Color_thresholding.jpg

Sorry to ask this of you when you are busy but I appreciate your help.

[mak carlos]

can u help me i has try run the code by matlab but it show out error on this sentence h = circle_hough(BW1, radii, 'same', 'normalise');

[Walter Roberson]

circle_hough is possibly https://www.mathworks.com/matlabcentral/fileexchange/26978-hough-transform-for-circles

[Image Analyst]

@mak carlos or Akshaya, try imfindcircles if the link Walter just gave does not work.

If you have any more questions, then attach your data and code to read it in with the paperclip icon after you read this:

TUTORIAL: How to ask a question (on Answers) and get a fast answer

Do it in a new discussion thread, not here in @Dipesh Gupta's 7 year old question.

Answer 2

[Prince Jindal]

hi Dipesh

Just take pixel by pixel product of original image and complemented binary mask you generated.

[Image Analyst]

This is how you'd do that:

% Mask the image using bsxfun() function
maskedRgbImage = bsxfun(@times, rgbImage, cast(mask, 'like', rgbImage));

This will give the middle image of the three images he posted in a comment to my Answer, so he's already doing this. I'm not sure what or if he had any questions after that.

Answer 3

[RIDZA]

%This code can be used if you know the x,y and r %only functioning using .jpg image % [filename, pathname] = uigetfile({'*.jpg';'*.bmp';'*.tif'}); [filename, pathname] = uigetfile({'/*.*'}); S = imread([pathname filename]);

%I = imread('088_1_2.jpg'); imageSize = size(I); % center and radius of circle ([c_row, c_col, r]) [xx,yy] = ndgrid((1:imageSize(1))-ci(1),(1:imageSize(2))-ci(2)); mask = uint8((xx.^2 + yy.^2)<ci(3)^2); croppedImage = uint8(zeros(size(I))); croppedImage(:,:,1) = I(:,:,1).*mask; croppedImage(:,:,2) = I(:,:,2).*mask; croppedImage(:,:,3) = I(:,:,3).*mask; imshow(croppedImage);

[Laisa Fernochio]

Hello Dipesh Gupta Can you give me the source code, I only need a code that detect the iris please

Answer 4

[Emilia Badescu]

Hello!! How can detect a value in this white area of the iris to distinguish it from the image that does not show this white area around the iris

[Image Analyst]

Do a search of the Image Processing Literature http://www.visionbib.com/bibliography/contents.html I know you can find articles there on finding sclera.

[Emilia Badescu]

I don`t want to detect the sclera, I want to detect the area around the iris (all white, but another shade, it's a ring) - it's called arcus senilis, because I'm doing a program for detection of arcus senilis and through this I want to differentiate my face healthy eyes, without arsenic senilis. Can you give me an idea how to do this? Thank you!

[Image Analyst]

Use the Color Thresholder on the Apps tab of the tool ribbon. Use hsv color space to segment out bright white. Then take the largest blob with bwareafilt() to get rid of things like noise and the specular glint in the middle.

[Emilia Badescu]

how can I detect two or three circles on the eyes as in the example attached: a pupil circle, one for the iris and another area, the area of interest from the extremity of the iris to the sclerotic, is a circle of cholesterol

[Image Analyst]

I'd first find the pupil - easy since it's the largest dark thing. Then use improfile to send out rays to get the average radial profile. Examine that to get the other rings. See similar demos, attached.

[Emilia Badescu]

thank you, I'm reading the transformed Hough, but I just find the circle of the iris