Histogram-based RGB segmentation

Hi All,

I'm trying to segment an RGB image based on it's histograms; my code is as follows:

    %Split into RGB Channels
    Red = img(:,:,1);
    Green = img(:,:,2);
    Blue = img(:,:,3);
    %Get histValues for each channel
    [yRed, x] = imhist(Red);
    [yGreen, x] = imhist(Green);
    [yBlue, x] = imhist(Blue);
    %Plot them together in one plot
    plot(x, yRed, 'Red', x, yGreen, 'Green', x, yBlue, 'Blue');
    
    histRGB = hist([imhist(Red), imhist(Green), imhist(Blue)]);
    histMask = img.*uint8(histRGB);
    imshow(histMask)

I get the following error when compiling this section: "Array dimensions must match for binary array op." I'm not sure how to fix this, as I'm not well versed in Matlab currently.

Any help will be appreciated!

Thanks

Answer 1

[Subhadeep Koley]

I am not sure about which algorithm you are trying to implement but below is an example of histogram based segmentation using Otsu's method.

clc; close all;

% Read image into the workspace
img = imread('peppers.png');

% Calculate a 255-bin histogram for the image
[counts, ~] = imhist(rgb2gray(img), 255);

% Compute a global threshold using the histogram counts
T = otsuthresh(counts);

% Create a binary image using the computed threshold
BW = imbinarize(rgb2gray(img), T);

% Display the image
figure; imshowpair(img, BW, 'montage');

[Teshan Rezel]

Thanks Subhadeep, much appreciated!