A watermarking algorithm based on block energy analysis of wavelet transform's coefficients and Lorenz chaotic attractor

Perceptual transparency and robustness are conflicting requirements. To ensure the security of the watermark how to choose the embedding position and intensity is a difficult problem. Chaotic system has characteristics of sensitivity to initial value and unpredictability. High-dimensional chaotic systems have advantage of higher secrecy, so this paper chooses Lorenz chaotic attractor to encrypt binary watermark image. Through the statistic analysis of block energy of the host image, which is applied DWT repeatedly up to the third level, so watermark adaptively embedded into the greatest coefficients which belong to the selected blocks by using the properties of human visual system. Additionally, to realize the blind digital watermarking extraction, chaotic sequence is also applied in selected wavelet coefficients to change the original value of the inherent relationship between wavelet coefficients. Theoretical analysis and experiment results show that the proposed algorithm has high robustness and better performance of perceptual transparency.