Image Watermarking Algorithm Using 2d Cellular Automata Transform

Xiao-Wei Li,Sung-Jin Cho,Seok-Tae Kim

2012-01-01

Abstract:Cellular automata are discrete dynamical systems that provide the basis for the synthesis of complex emergent behavior. Starting from a basic unit, the cell, the use of cellular automata not only provides insight into the evolution of life but also opens vistas into the creation of intelligent computer systems. This paper proposes a secure and novel digital watermarking system based on Maximum-Length Cellular Automata (MLCA) and a Two-Dimension Cellular Automata Transform (2D CAT). First, the original image is decomposed into a level-1 CAT algorithm and the sub-bands LL1, HL1, LH1 and HH1 are obtained, and further using level-1 CAT to sub-bands HL1 and LH1, the middle-frequency: HL2 and LH2 are obtained. Then, the original watermark, which is encrypted by the Maximum-Length Cellular Automata (MLCA), is embedded into the CAT domains of the original image. Finally, the watermarked image is obtained using the inverse Cellular Automata Transform (ICAT) for the CAT-transformed image. This is a novel scheme because it has the capability of robustness and invisibility simultaneously. However, the traditional watermarking algorithm based on CAT cannot achieve. The experimental results show that our CAT-based watermarking system can simultaneously improve the security, robustness and image quality of the watermarked image.