how to convert gray scale image into rgb image?

my project is on iridiology...and it requires to convert the gray scale image to an rgb image to find the basic color of the iris. am in need of the code to convert the gray scale image to rgb image.

[TEJASWINI CHINAWALE]

you can use ind2rgb(grayscale image, colormap)...However that cannot convert your grayscale image totally into the coloured one as earlier.

[Image Analyst]

You should put this in the "Answer" section. Right below my Answer that says the same thing you just did.

Answer 1

[Ahmed A. Selman]

Use colormap types, since you are interested with a colored region of interest, I think any colormap format suffices. Example:

   close all
   clear 
   clc % just to clear things up
   I=imread('333.jpg');
   R=rgb2gray(I); % It is gray now
   y=colormap(hot(110));
   imwrite(R,y,'rgb.jpg','jpg'); % Re-change it to colored one 
   imshow('rgb.jpg');

[Ahmed A. Selman]

Hi ug, The above is an example. You must provide your own grayscale images instead of (R=..), then change it to colored one by the lines (y=..) and (imwrite..). Everything above the line of (y) was used to make a gray image only.

The image ('333.jpg') is on MY Matlab path, of course it might not work with you :)

[ug]

i hav done all of this... but still it fails to work... is der ne other method?

[Image Analyst]

Please explain why you think it is even possible to recover "the basic color of the iris". Did you snap the original iris image with a color camera or a monochrome camera?

[ug]

here is the procedure, i hav an image of the iris with all extra features like eye lashes n all..i hav to extract only the iris frm tis image. first i select the two circles of the iris..the resultant image is that the coloured part of the iris is converted to gray scale n the rest is blacked out... but i want to get the original colour....is there ne other method to get the colour of the iris?

[Image Analyst]

Of course. Wish you'd explained that initially. See my new answer.

Answer 2

[Image Analyst]

ug, if your starting image is monochrome, then you cannot determine the original color. You cannot. Not if all you have is the grayscale image. Period.

You can convert a grayscale image into an RGB image using cat():

rgbImage = cat(3, grayImage, grayImage, grayImage);

of course it will look all grayscale even though it is a "true color" RGB image - it's just that all the "colors" are gray. You can apply a pseudocolor lookup table to the gray values, where each gray value gets mapped into some color, to get a multi-colored image, like this:

rgbImage = ind2rgb(grayImage, jet(256));

however the colors are NOT the original colors of the iris as if you had snapped it with a color camera.

[ug]

am getting an error in the ind2rgb function... sayin the dimensions don match

[Image Analyst]

Like I said before see my new answer ( http://www.mathworks.com/matlabcentral/answers/69368#answer_80694). ind2rgb() doesn't apply anymore. You will not be using ind2rgb. Again, see my new answer above.

Answer 3

[Image Analyst]

You ask: "i hav an image of the iris with all extra features like eye lashes n all..i hav to extract only the iris frm tis image. first i select the two circles of the iris..the resultant image is that the coloured part of the iris is converted to gray scale n the rest is blacked out... but i want to get the original colour....is there ne other method to get the colour of the iris?"

Here's how to do it.

% Make a mask from your circle using poly2mask
% xCircle, yCircle are the coordinates of your circle perimeter.
mask = poly2mask(xCircle, yCircle, rows, columns);
% Now mask the image so that only the original RGB part
% inside the circle remains, and outside the circle is set to black.
maskedRgbImage = bsxfun(@times, rgbImage, cast(mask,class(rgbImage)));

Answer 4

[Mahdi]

Gabriel made a script for that. You can find it here.

[ug]

this did nt work too

Answer 5

[ug]

 @Ahmed A. Selman
Am getting the following errors:
??? Error: File: C:\MATLAB7\toolbox\images\images\private\checkinput.m Line: 38 Column: 1
Incomplete or misformed expression or statement.

Error in ==> images\private\checkstrs at 26 checkinput(in, 'char', 'row', function_name, variable_name, argument_position);

Error in ==> imlincomb>ParseInputs at 95 output_class = checkstrs(varargin{end}, valid_strings, mfilename, ...

Error in ==> imlincomb at 83 [images, scalars, output_class] = ParseInputs(varargin{:});

Error in ==> rgb2gray at 58 I = imlincomb(coef(1),X(:,1),coef(2),X(:,2),coef(3),X(:,3), ...

Error in ==> trial at 5 R=rgb2gray(I); % It is gray now

[Image Analyst]

Please make this a comment do Ahmed's answer since your response here is not an answer to your question, but a comment to his answer. Strange though that you are willing to accept any old arbitrary colormap (like "hot" which he uses), and think that this will give you the original color of the iris. What's the reasoning for that?

[Ahmed A. Selman]

Thanks Analyst for the most valuable suggestion, indeed. But first please read my answer correctly, seems you skipped the (Example) and (any color map might work) parts.

[Image Analyst]

I did read it and ran the example before I replied. I figured you must have known that the sample colormap you used, hot, would not be appropriate in his situation. I never did know what you meant by "any colormap format suffices" since the color maps used in ind2rgb must be a prescribed format of a 2D array of N rows by 3 columns with values in the range of 0 to 1. But I took it to indicate that you think there is some colormap out there that would work, and that is where I differ. I don't think there is any colormap out there that would turn his grayscale image into the correctly colored iris image. Where or how would he get such a colormap? You can't take a grayscale image created by, say, rgb2gray() or by taking just one of the color channels, and get back the original color image. You could do an approximate job with an indexed image if you kept the color map returned by rgb2ind(), but the indexed image is not a grayscale image - it looks like garbage if you display it as grayscale. Happy to discuss further....

[Ahmed A. Selman]

True, Analyst, I do agree with the idea that when color is lost it cannot be retained correctly. Thus any color map can do the job mentioned in the question with the same quality, give or take; they all will NOT be correct, but ANY could specifically isolate iris from the rest, at least better than a grayscale image. In fact I do wonder what does (determine the basic color of the iris from gray images) means, I keep reading amazing stuff here. Correctly colored iris image from gray image is, without doubt, not possible unless a true image was used to specify the colormap .. and if true images do exist then why convert them to gray and ask how to retain the color?..

Anyway, industrially colored images do look like garbage, even when using highly sophisticated image software (truth has only one face..) but does it matter if it was using hot or jet or any other map? If there is a difference please specify what kind is better -- new info are useful and appreciated, always. If not, however, then old or new colormaps will be all as crappy as the same, right? :)

[Image Analyst]

We're in agreement here. If all you want to do is to colorize it, you can pick whatever colormap you feel looks best or suits your needs - it's totally subjective.

[Ahmed A. Selman]

My exact point. Thank you.

Answer 6

[G.Liz]

Does anyone know how to segment the sclera of the eye ?

[Walter Roberson]

Please start a new Question for that.