How do I get a colormap for my image?

When I import my image using the imread function, I get a map with only zeros (map = []). I want the RGB components of the image, but cannot get them since this requires the picture's colormap (which is empty). This is the code I'm using:

filename = uigetfile('*.*');
[X, map] = imread(filename);
info = imfinfo(filename)
width = getfield(imfinfo(filename),'Width')
height = getfield(imfinfo(filename),'Height')
figure('Name','Display imported image','NumberTitle','off');
image(X);
rgb = ind2rgb(X,map);

The info of the image is:

info = 

              Filename: '11-11-11 (30).JPG'
           FileModDate: '11-Nov-2011 21:07:46'
              FileSize: 4549123
                Format: 'jpg'
         FormatVersion: ''
                 Width: 4320
                Height: 2432
              BitDepth: 24
             ColorType: 'truecolor'
       FormatSignature: ''
       NumberOfSamples: 3
          CodingMethod: 'Huffman'
         CodingProcess: 'Sequential'
               Comment: {}
      ImageDescription: '                               '
                  Make: 'SONY '
                 Model: 'DSC-W350 '
           Orientation: 1
           XResolution: 72
           YResolution: 72
        ResolutionUnit: 'Inch'
              DateTime: '2011:11:11 21:07:46 '
      YCbCrPositioning: 'Co-sited'
         DigitalCamera: [1x1 struct]
           UnknownTags: [1x1 struct]
         ExifThumbnail: [1x1 struct]

And then I get this error message:

??? Index exceeds matrix dimensions.

Error in ==> ind2rgb at 27
r = zeros(size(a)); r(:) = cm(a,1);

Error in ==> DSP_LCK at 24
        rgb = ind2rgb(X,map);

Some advice please?

Answer 1

[Alex Taylor]

The metadata associated with your image is indicating that your data is already an RGB image with separate RGB planes:

           ColorType: 'truecolor'
     NumberOfSamples: 3

ind2rgb is failing because you are providing input that is already an RGB image (and is not an indexed image).

If you want the RGB components:

R = X(:,:,1);
G = X(:,:,2);
B = X(:,:,3);

[Louis Kok]

Thanks alot Alex, I've tried this, but the images I then get has various other colors mixed with them in each of these three color planes. I can somehow attach the image I'm using so that you can experience this if you'd like?

[Walter Roberson]

The color components are as Alex indicated above. If you want to display them as individual images, you need to put them into appropriate planes.

Rimg = cat(3, R, zeros(size(R)), zeros(size(R)));

Gimg = cat(3, zeros(size(G)), G, zeros(size(G)));

Bimg = cat(3, zeros(size(B)), zeros(size(B)), B);

[Image Analyst]

You accepted, so is it solved or not? Do you want to turn each individual color channel into an RGB image "of its own color" like Walter suggested? Another way is to use colormap() on the original color channel images.

[Louis Kok]

Sorry, the "Accept" button was hit on accident!

[Louis Kok]

Thanks Walter, this was more of what I had in mind to just display each image only in a red, green and blue window. I will try this and see if this solves my problem.

Answer 2

[Image Analyst]

Try this demo:

clc;    % Clear the command window.
close all;  % Close all figures (except those of imtool.)
workspace;  % Make sure the workspace panel is showing.
format compact;
fontSize = 22;

%------------------------------------------------
% Ask user for image file.
% Initialize with some folder to start browsing from.
startingFolder = 'C:/Users/Louis/Pictures';
% Get the name of the color image file that the user wants to use.
defaultFileName = fullfile(startingFolder, '*.*');
[baseFileName, folder] = uigetfile(defaultFileName, 'Select a file');
if baseFileName == 0
  % User clicked the Cancel button.
  return;
end
fullFileName = fullfile(folder, baseFileName)

%------------------------------------------------
% Read in the image.
rgbImage = imread(fullFileName);
% Get the dimensions of the image.  numberOfColorBands should be = 3.
[rows columns numberOfColorBands] = size(rgbImage);
% Display the original color image.
subplot(2, 2, 1);
imshow(rgbImage, []);
title('Original Color Image', 'FontSize', fontSize);
% Enlarge figure to full screen.
set(gcf, 'units','normalized','outerposition',[0 0 1 1]);

%------------------------------------------------
% Extract the individual red, green, and blue color channels.
redChannel = rgbImage(:, :, 1);
greenChannel = rgbImage(:, :, 2);
blueChannel = rgbImage(:, :, 3);

%------------------------------------------------
% Display the individual color channels.
subplot(2, 2, 2);
imshow(redChannel, []);
title('Red Channel Image', 'FontSize', fontSize);
subplot(2, 2, 3);
imshow(greenChannel, []);
title('Green Channel Image', 'FontSize', fontSize);
subplot(2, 2, 4);
imshow(blueChannel, []);
title('Blue Channel Image', 'FontSize', fontSize);

[Louis Kok]

Thanks alot! I will try this to see the results, but the portion of

redChannel = rgbImage(:, :, 1);

greenChannel = rgbImage(:, :, 2);

blueChannel = rgbImage(:, :, 3);

seems alot like Alex suggested above and I've already tried something like this. The results of these three colorchannels are images with other color components mixed into the image and not only filtered by Red, Green and Blue only!

[Image Analyst]

Sorry, but you are wrong. There is no mixing of the color channels. Each color channel is a gray scale image, not a color image, just in case you were incorrectly thinking that if you display the redChannel it should appear red. It's just a 2D array of numbers. Think about it.

If you want each color channel to appear in "it's own" color, then you have to apply a pseudocolor look up table with the colormap() function. But that just affects the display only, not the underlying image, which is still just a 2D monochrome array of numbers.

[Louis Kok]

Thank you Analyst, I agree with you that the underlying image is still a 2D array of numbers. I got confused when I displayed only the redChannel as I did, and got only that plane's data displayed. What I'm actually interested in, is the resulting images to display only the red, green and blue color components of my RGB image. I have to do edge detection (Sobel's operator) on each of these three monochrome images -- which originates from the original RGB image.

Answer 3

[Michael Barrow]

Got here by googling how to read in a colormap for an image that does not have one. My goal was to import an indexed image directly as I needed a colormap to texture a 3D mesh.

The solution for me was to re-save the image in indexed mode using an external tool (gimp, an open source image editor).

When I used [X,map]=imread('newfile.png'); I had a usable map that allowed me to colour my mesh

[DGM]

To avoid having to touch disk or use external tools, you could always use rgb2ind().

% a RGB image
rgbpict = imread('peppers.png');

% convert to indexed color using whatever settings you want
% most formats which support indexed output should support up to 256
[indpict CT] = rgb2ind(rgbpict,256);
imshow(indpict,CT)
xlim([200 500]); ylim([100 300]) % zoom in

% try a shorter table to help visualize the quantizaton
% the default is to use error-diffusion dithering
[indpict CT] = rgb2ind(rgbpict,16);
imshow(indpict,CT)
xlim([200 500]); ylim([100 300])

% but the dithering can be disabled
[indpict CT] = rgb2ind(rgbpict,16,'nodither');
imshow(indpict,CT)
xlim([200 500]); ylim([100 300])

... though it can't do a patterned dither, so if that's something you want, using external tools would make sense.

Answer 4

[Louis Kok]

Here is a link to the image I'm referring to. Thank you for all your help and insight!!

http://dl.dropbox.com/u/66925667/DSP_Image.rar

[Image Analyst]

Can you make it easy and just post the image so we can see it immediately, and not something I have to download, extract, then load into MATLAB?

[Louis Kok]

Sorry for the inconvenience! I'm having trouble only posting the image as it is, but I can give and example of what I'm expecting to get from this image processing in Matlab. Here is a link to a combined image of the results I'm expecting.

http://dl.dropbox.com/u/66925667/results.jpg

F.l.t.r. The original RGB image, the green image, a gray-scale image, the red and blue images as well as a gray-scaled image that has been passed through a low-pass filter after taking the fft of that gray-scale image.

A friend of mine sent me all these images, but he processed them from C++ software in Qt-environment.

I aim to produce these type of images in Matlab, since I also need to do edge detection on each of the three fundamental color images as well as to filter out high frequency components from the gray-scale image after applying an FFT algorithm. The catch is, I'm not allowed to use Matlab's FFT() functions or edge detection algorithms, I have to code them myself. I'm progressing slowly, since I'm having trouble to get these fundamental color images. I am able to get the gray-scale image of the original RGB image.

Thanks a lot for your insight, help and tips!! My apologies for the file download schlep.

Answer 5

[Louis Kok]

Walter, I've implemented your suggestion and it works perfectly fine as is!

Thanks a lot for all your contributions: Alex, Image Analyst and Walter. I've learned a lot about RGB images, some properties of RGB image processing we're not learning about on varsity as well as new coding techniques and tools I now have to create better algorithms for my practicals and assignments! I kind of feel stupid not thinking to zero-pad the color planes I do not want.

My apologies also for incompetent questions and statements regarding image processing. This semester is our firs DSP course and we've tackled mostly the math behind CTFT, DTFT, DFT and a quick overview of the FFT and the accompanying background foundation for DSP. Color image processing was the assignment and mostly self-study.

Answer 6

[Louis Kok]

This is my derived and modified solution:

        [filename filepath] = uigetfile('*.*');
        [X, map] = imread(filename);
        info = imfinfo(filename)
        width = getfield(imfinfo(filename),'Width');
        height = getfield(imfinfo(filename),'Height');

        figure('Name','Display imported image','NumberTitle','off');
        image(X);

        R = X(:,:,1);
        G = X(:,:,2);
        B = X(:,:,3);

        Rimg = cat(3, R, zeros(size(R)), zeros(size(R)));
        Gimg = cat(3, zeros(size(G)), G, zeros(size(G)));
        Bimg = cat(3, zeros(size(B)), zeros(size(B)), B);

        figure('Name','RED component of the imported image','NumberTitle','off');
        image(Rimg);
        figure('Name','GREEN component of the imported image','NumberTitle','off');
        image(Gimg);
        figure('Name','BLUE component of the imported image','NumberTitle','off');
        image(Bimg);

        G = rgb2gray(X);    
        figure('Name','Gray-scale image of the imported image','NumberTitle','off');
        imshow(G);

Again, thanks a lot to everyone contributing!