Backgroud Added to images after imwrap transformations

I am performing simple transformations but the Matlab adds backgroud to the image to make it Rectangle. The image has transparent backgroud. I know how to read the image and remove background but I fail to remove it after the transformations

I=imread('d.png');
I=flipud(I);
T = [1  0
     1 1]; 
figure;
T(3,3)=1;
[Iw,R]=imwarp(I, affine2d(T'));
h=imshow(Iw);
axis on xy;
set(h,'XData',R.XWorldLimits,'YData',R.YWorldLimits);
set(gca,'XLim',R.XWorldLimits,'YLim',R.YWorldLimits);

The following is being used

After the transformation the below image is displayed

I do not want this background to appear

Answer 1

[DGM]

Your image has alpha, but you're not using it.

% you need to read the whole image
[I,~,alpha] = imread('https://www.mathworks.com/matlabcentral/answers/uploaded_files/1575777/image.png');

% i don't know why you're flipping this, 
% but the image will be upside-down when displayed (unless you display it upside-down)
% and it will be upside-down when you save it
I = flipud(I);
alpha = flipud(alpha);

% build T
T = [1 0; 1 1]; 
T(3,3) = 1;

% you could attach the alpha and transform this as a single RGBA array
% but imshow() and imwrite() won't know what to do with it.
[Iw,R] = imwarp(I, affine2d(T'));
[alphaw,~] = imwarp(alpha, affine2d(T'));

% display it
h = imshow(Iw);
h.AlphaData = alphaw; % this is the only way you can use alpha in imshow()
axis on xy; % you're displaying it upside-down
set(h,'XData',R.XWorldLimits,'YData',R.YWorldLimits);
set(gca,'XLim',R.XWorldLimits,'YLim',R.YWorldLimits);

% save it with alpha
imwrite(Iw,'out.png','alpha',alphaw)

As the comments state, you're creating an upside-down image, but you're also viewing it upside-down. I don't know what that means to you.

[Farhan]

I didn't understood teh code much, Thanks for such clear comments now I know much of the code. I just wanted to do transformation someone helped me wrote this code. I will change it as you suggested. Thanks

[Farhan]

played a bit with your code, now because of your comments I understand what I am doing. Thanks a lot again, it really helped a lot

Answer 2

[Sulaymon Eshkabilov]

Here is the complete code (your diamond image was used with DIAM.png name):

I=imread('DIAM.png');
I=flipud(I);
T = [1  0
     1 1]; 
figure
imshow(I)
title('Original')

figure;
T(3,3)=1;
[Iw,R]=imwarp(I, affine2d(T'));
% Convert the Image to Grayscale:
Gray_img = rgb2gray(Iw);
% Create a Mask (called B_mask) for the Background Color:
B_mask = Gray_img ==0; % Adjust the threshold based on the transformed image

% Apply the mask (B_mask) to remove the background color
Final_img = Iw;
Final_img(repmat(B_mask, [1, 1, 3])) = 255; % Set background pixels to white

h=imshow(Final_img);
axis on xy;
set(h,'XData',R.XWorldLimits,'YData',R.YWorldLimits);
set(gca,'XLim',R.XWorldLimits,'YLim',R.YWorldLimits);
title('Transformed Image')