Remove white padding from image

I want to remove white padding from an image (see attachment). The size of the padding is random.

So I was thinking of reading in the image and then some looping to check if the entire row/column contains the value 255. However, looping is most of the time not the best way to do stuff. So I'm wondering if there's an easier way to do this, because I've found a similar question where the rows/columns contain 0, but I don't know if this can also be used for this case (with some adjustments).

Thanks in advance.

Answer 1

[Walter Roberson]

I = imread('test-image.jpg');
W = all(I == 255, 3);
maskc = all(W, 1);
maskr = all(W, 2);

J = I;
J(maskr,:,:) = 0;
J(:,maskc,:) = 0;

image(J)

[YT]

While I'll accept this answer because it helped me out, the result of your code will not remove the padding from the image but rather makes it black (changes 255 to 0).

With that being said, I combined this answer and the answer you provided in the question I mentioned earlier to fix the issue.

    %read image
    I = imread('test-image.jpg');
    [~,~,d] = size(I);
    %255 to 0
    W = all(I == 255, 3);
    maskc = all(W, 1);
    maskr = all(W, 2);
    J = I;
    J(maskr,:,:) = 0;
    J(:,maskc,:) = 0;
    % remove 0 rows/columns from R/G/B
    for i = 1:d
        K = J(:,:,i);
        K( ~any(K,2), : ) = [];  %rows
        K( :, ~any(K,1) ) = [];  %columns
        switch i
            case 1
                R = K;
            case 2
                G = K;
            case 3
                B = K;
            otherwise
                warning('Unexpected value');
        end
    end
    IM = cat(3, R, G, B);
    imshow(IM);

[Walter Roberson]

"remove white padding" is ambiguous as to whether it should be transformed to a different color or cropped out of the image.

I = imread('test-image.jpg');
W = all(I == 255, 3);
maskc = ~all(W, 1);
maskr = ~all(W, 2);

J = I(maskr, maskc);
imshow(IM);

[Nate Roberts]

This is exactly what I was looking for. I made a little script that can do this for an RBG image, assuming you only want to remove the padding, and not the white space within an image.

function im2 = removeWhitePadding(im1, dim)
    if nargin<2
        dim = 0;
    end

    WhitePixels = all(im1 == 255, 3);
    maskr = ~all(WhitePixels, 2);
    maskc = ~all(WhitePixels, 1);
    
    if dim==1
        im2 = im1(find(maskr, 1, 'first'):find(maskr, 1, 'last'),:,:);
    elseif dim==2
        im2 = im1(:,find(maskc, 1, 'first'):find(maskc, 1, 'last'),:);
    else
        im2 = im1(find(maskr, 1, 'first'):find(maskr, 1, 'last'),find(maskc, 1, 'first'):find(maskc, 1, 'last'),:);
    end
end

Answer 2

[Image Analyst]

Try this:

rgbImage = imread('test-image.jpg');
subplot(2, 2, 1);
imshow(rgbImage);
title('Original RGB Image', 'FontSize', 20);
axis('on', 'image');

nonWhitePixels = any(rgbImage ~= 255, 3);
subplot(2, 2, 2);
imshow(nonWhitePixels)
title('Non-white Pixels Mask', 'FontSize', 20);
axis('on', 'image');

[maskRows, maskColumns] = find(nonWhitePixels);
% Crop image
croppedImage = rgbImage(min(maskRows):max(maskRows), min(maskColumns):max(maskColumns), :);
subplot(2, 2, 3);
imshow(croppedImage);
axis('on', 'image');
title('Cropped RGB Image', 'FontSize', 20);