Karhunen-Loeve Transform for watermarking

Does someone have any code for digital watermarking using the KLT transform? Any help would be welcome!

Answer 1

[Harsh Sanghai]

Hi,

Here is a sample code which you can use for digital watermarking using KLT transform in using MATLAB:

% Read the original image
I = imread('original_image.jpg');

% Convert the image to grayscale
I_gray = rgb2gray(I);

% Generate the watermark image
W = rand(size(I_gray)) > 0.5;

% Perform KLT transform on the original image
C = dct2(I_gray);

% Calculate the covariance matrix
R = cov(C);

% Calculate the eigenvalues and eigenvectors
[V,D] = eig(R);

% Sort the eigenvalues in descending order
[~,ind] = sort(diag(D),'descend');
V = V(:,ind);

% Use the first N eigenvectors to embed the watermark
N = 5;
W_embedded = C + 0.1*W.*repmat(V(:,1:N)*V(:,1:N)',size(C,1),size(C,2));

% Perform inverse KLT transform
I_watermarked = idct2(W_embedded);

% Display the original and watermarked images
figure;imshow(I_gray);title('Original Image');
figure;imshow(I_watermarked);title('Watermarked Image');