A Color Image Watermarking Algorithm Based on Sparse Representation

This study presents a novel method to embed and extract watermarks for an image via on sparse representation. First, the decomposition coefficient matrix is obtained by the sparse decomposition of the color carrier image based on the minimum gray value. Then, a watermark will be embedded into the coefficient matrix to generate a color image with the sparse reconstruction. Finally, the watermark can be extracted by the sparse decomposition of the image green and blue channels. The experimental results from our simulation show that the proposed watermark embedding schema is robust to multiplicative noise attacks and Wiener filtering attack, the embedded watermark can meet the requirement of watermark invisibility and the extracted watermark information could be correctly identified.