A blind watermarking optimal detection based on the wavelet transform domain

A performance of watermarking scheme relies heavily on the design of the detection. However, most of watermark detection algorithms in the literature are not optimum. Based on the generalized Gaussian distribution model of wavelet coefficients, an adaptive optimal blind watermarking detection is deduced and presented in this paper. Used an asymptotically optimal detection, it is an adaptive blind watermarking detection with the estimate of the shape parameter of wavelet coefficients. A series of experimental results which demonstrate the effectiveness of detector show that it is more validity and practicability than the correlation detector.