How to add and subtract two color images in matlab?

I-G(I);

where I=original image and G(I)=Gaussian filtered version of I.

[Rik]

Using just the normal symbols will result in an element-wise addition or subtraction. That could cause issues with underflow and overflow if you are using an integer data type, and the result might not be what you mean.

What is your question? What have you tried? How did it fail to match your desired result?

[Sajina Rose]

I'm writing a code for subtract two images and then apply histogram stretching to resultant image.

Here, my sample code

%%%%%%

z=imsubtract(double(X)-double(Y));

N=imadjust(z,[],0.5);

figure,imshow(N),title('Histogram Stretching');

%%%%%%%

But is shows error like,

"Not enough input arguments.

Error in imsubtract (line 34)

scalarDoubleY = isa(Y,'double') && numel(Y) == 1;

Error in method1 (line 83)

z=imsubtract(double(AGW)-double(B));"

Answer 1

[Subhadeep Koley]

Hi, Sajina the function imsubtract expects two real, nonsparse, numeric or logical image. Threfore the below syntax is not right.

z = imsubtract(double(X)-double(Y)); 

Try the code below.

% Read your image here
X = imread('peppers.png');
figure; imshow(X, []); title('Original Color Image');

% Gaussian Filtering
sigma = 5; % Standard deviation of the kernel
Y = imgaussfilt(X, sigma);
figure; imshow(Y, []); title('Gaussian Filtered Image');

Z = imsubtract(X, Y);
N = imadjust(Z, stretchlim(Z));
figure,imshow(N, []), title('Histogram Stretched Subtracted Image');

Hope this helps!

[Sajina Rose]

Thanks. your code works well........can you please help me to add original image with histogram stretched image and then divide it by 2 (unsharp masking procedure)

[Rik]

The imsubtract is already handling the subtraction, so you don't have to use the symbol yourself. See if the line below helps.

z=imsubtract(double(X),double(Y));

[Subhadeep Koley]

@ Sajina Rose Refer the below code

% Read your image here
X = imread('peppers.png');
figure; imshow(X, []); title('Original Color Image');

% Gaussian Filtering
sigma = 5; % Standard deviation of the kernel
Y = imgaussfilt(X, sigma);
figure; imshow(Y, []); title('Gaussian Filtered Image');

Z = imsubtract(X, Y);
N = imadjust(Z, stretchlim(Z));
figure,imshow(N, []), title('Histogram Stretched Subtracted Image');

U = (imadd(X, N)) / 2;
figure,imshow(U, []); title('Unsharp Masked Image');

Otherwise you can use the MATLAB built-in function imsharpen(), which performs Unsharp Masking.

XSharp = imsharpen(X, 'Radius', 1.5, 'Amount', 2);
figure,imshow(XSharp, []), title('Unsharp Masked Image');

[Sajina Rose]

Thank you so much.