Fast and Robust Watermarking Method using Walsh Matrix Partition

In the research of watermarking algorithm, the Singular Value Decomposition (SVD) is widely use. It has the capability of maintaining the imperceptibility and robustness at a good level. However, the SVD has a high level of complexity in the process of decomposition and reconstruction. It split a host image into three matrices of U, S, and V. The S matrix contains with the singular value and it is used in the embedding and extraction processes. The other two matrices of U and V are orthogonal and it is used to reconstruct the watermarked image. This scheme can be simplified using a single matrix which is orthogonal and symmetrical to replace the use of matrices of SVD. This paper is proposed a watermarking algorithm using a Walsh matrix which is symmetric, orthogonal and contains signed integer value of 1 and -1. The Walsh matrix is used to transform the host image into Walsh coefficient and reconstruct the watermarked image using a simple equation. The experiment result shows that the proposed method has faster embedding and extraction time with average time of 0.3841 and 0.1854 second compared to the SVD with average time of 0.6235 and 0.2539 second. Meanwhile, the both method has same level of robustness and imperceptibility. The average PSNR value of the proposed method and SVD are 41.4991 and 41.6518 respectively while the average NC values are 0.9416 and 0.9428. The proposed method is able to shorten the processing time without reducing the imperceptibility and robustness.