Performance analysis of correlation-based watermarking schemes employing Markov chaotic sequences

In this paper, theoretical performance analysis of watermarking schemes based on correlation detection is undertaken, leading to a number of important observations on the watermarking system detection performance. Statistical properties of watermark sequences generated by piecewise-linear Markov maps are investigated. Correlation/spectral properties of such sequences are easily controllable, a fact that reflects on the watermarking system performance. A family of chaotic maps, namely the skew tent map family, is used for verifying the theoretical analysis. Skew tent chaotic sequences are compared against the widely used pseudorandom sequences, indicating the superiority of the former in watermarking applications. The minimum number of samples required for reliable watermark detection is also investigated. Experiments using audio data are conducted to verify the theoretical analysis results.