How to create histograms for L a b channels separately in CIELAB color space.

How to create histograms for L a b channels separately in CIELAB color space.

Answer 1

[Image Analyst]

Try this:

rgbImage = imread('peppers.png'); % Read in example image shipped with MATLAB
labImage = rgb2lab(rgbImage);
[LImage, AImage, BImage] = imsplit(labImage);

Then call histogram on each image.

figure;
subplot(4, 2, 1);
imshow(rgbImage);
title('Original RGB image')
subplot(4, 2, 3);
imshow(LImage, []);
title('L Image')
subplot(4, 2, 4);
histogram(LImage);
title('Histogram of L image');
subplot(4, 2, 5);
imshow(AImage, []);
title('A Image')
subplot(4, 2, 6);
histogram(AImage);
title('Histogram of A image');
subplot(4, 2, 7);
imshow(BImage, []);
title('B Image')
subplot(4, 2, 8);
histogram(BImage);
title('Histogram of B image');

Or see my attached demo and adapt it to LAB color space.

Answer 2

[DGM]

Let's actually plot some histograms. I'm using fixed xlimits so that the histograms are presented consistently between images and it's readily apparent where the chroma distribution is in relative terms. The limits I chose are wider than the gamut of sRGB as projected into LAB using the applied transformation parameters buried in rgb2lab(), though I've made no effort to mark those extents. I just wanted the neutral axis to be centered.

inpict = imread('peppers.png');

labpict = rgb2lab(inpict);
limits = [0 100; -108 108; -108 108]; % oversize and symmetric
for c = 1:3
    subplot(3,1,c)
    histogram(labpict(:,:,c));
    
    % use fixed limits so that presentation is consistent
    xlim(limits(c,:))
end

MIMT now has a tool to do this.

% an RGB image
inpict = imread('peppers.png');

% use defaults
cshist(inpict,'lab');

% use a filled patch instead of the default bar plot
cshist(inpict,'lab','style','patch');

% or use a stem plot like imhist()
cshist(inpict,'lab','style','stem');

See also:

https://www.mathworks.com/matlabcentral/answers/1936059-how-can-i-plot-the-ycbcr-histograms-with-the-correct-colour-points-0-255