A zero-watermarking for color image based on LWT-SVD and chaotic system

A zero-watermarking algorithm for color image based on lifting wavelet transform (LWT), singular value decomposition (SVD) and chaotic system is proposed in this paper. Firstly, the characteristics of a 5-D chaotic system are analyzed by phase diagrams, Lyapunov exponent spectrums and bifurcation diagrams. The initial conditions of the chaotic system are changed by hash value calculated from the watermark image. Then, feature information of low-frequency region of color image is extracted by LWT, and the maximum singular value is obtained by improved SVD. Based on chaos correction factor, the feature matrix can be extracted and encrypted by SVD simultaneously. And the watermark image is composed of meaningful single watermark or combined watermark. Finally, zero-watermarking is generated by encryption and logical operation of feature matrix and watermark image. During authentication, the watermark image can be extracted from the color image and the registered zero-watermarking, and the copyright ownership can be determined according to the registration time. This algorithm can effectively protect image copyright without changing image and has good robustness. Experiments verify the effectiveness of the algorithm and its ability to resist filtering, compression, rotation, noise and clipping attacks, and the NC value of each attack is higher than 89%.