The University of British Columbia

In this paper a wavelet-based watermarking is proposed according to which watermark is embedded on a selected sets of wavelet coefficients with two different strength factors. The method, based on a criteria defined in the paper, distinguishes between dominant and non-dominant coefficients in each sub-band in an attempt to increase the embedded watermark power as much as possible It is shown that the use of two different strength factors - instead of a single global factor, results in an improved image quality preservation and in a more resilient watermark against different attacks. In this paper a new parameter, referred to statistical detection threshold SDT, was introduced and used to enhance the performance of watermarking. Several illustrative examples are given including a comparative analysis of the performance of the proposed method with the standard single global factor method and MAP-based watermark detection algorithm.