Image Watermarking Based on Matrix Decomposition and Gyrator Transform in Invariant Integer Wavelet Domain

Abstract In this paper, an image watermarking algorithm with high security and robustness using matrix decomposition and Gyrator transform is proposed, which is developed in the invariant integer wavelet domain. The novelty of this paper is shown below. We propose the redistributed invariant integer wavelet transform, which has real reversibility and can transform an image into the invariant domain where the pixel values are all integers. Moreover, a new method for finding the singular value of an image matrix by QR decomposition and singular value decomposition (SVD) is given. Compared with SVD, it has lower time complexity. In addition, a new dual encryption scheme based on Gyrator transform and QR decomposition is introduced. This encryption scheme has reversibility and high security. The simulation results show that the proposed encryption algorithm can effectively resist statistical attack and differential attack. Moreover, the proposed watermarking scheme performs well in resisting geometric attacks and has obvious advantages over the existing related algorithms. In addition, this watermarking algorithm satisfies the invisibility, security, robustness and capacity requirements of the watermarking algorithm.