EBOF: A fusion-domain color image watermarking scheme based on approximate Hessenberg factorization

Abstract

With the rapid development of mobile networking devices and Internet technology, color images gradually occupy the dominant position of image applications, and the protection of color image copyright is also imminent. In this paper, a blind watermarking scheme for color images based on approximate Hessenberg factorization (AHF) in fusion domain is proposed. The formation rule of Hessenberg matrix coefficients is further researched based on the Householder transform, and the approximate values of Hessenberg matrix at position (1, 2), (2, 1), (2, 2) are directly obtained and used to hide the watermark information in the spatial domain. The proposed scheme not only consists of fractional order Chen’s chaotic system (FOCCS) to encrypt the watermark image, but also combines with Error correction coding application, B-R-G watermark embedding principle, Overflow processing of extreme pixel block and Firefly algorithm (EBOF). Moreover, the embedding position and the amount of embedding information can be also adaptively selected according to the image size. The simulation results show that the presented fusion-domain scheme outperforms state of the art schemes. On the premise of ensuring low time complexity, the presented watermarking scheme has strong robustness, high invisibility and high security.