Digital watermarking for multimedia security using complex wavelets

Digital data hiding has gained popularity in recent years as a way of settling intellectual property rights disputes, the ease with which digital media can be copied and distributed has led to a need for a method of identifying the original owner ofthe media. Data hiding allows for the embedding of information in a host signal that can later be extracted as proof of ownership or for some other purpose. Robustness, imperceptibility and capacity are the three conflicting attributes of data hiding systems. This thesis is dedicated to the establishment of data hiding algorithms that will achieve new standards of imperceptibility and robustness and consequently lead to higher data hiding capacity. It is proposed to apply Spread Transform embedding that combines the two main methods of data hiding, Spread Spectrum (SS) and Quantization Index Modulation (QIM). Complex wavelets are employed as the watermark embedding domain. These wavelet transforms have several advantages over discretely sampled wavelet transforms such as improved directional selectivity. Visual models are developed for these wavelet transforms. The application of these wavelet transforms to the area of digital watermarking is considered including an information theoretic approach to derive the optimal capacity limits ofthe considered watermarking algorithms. In addition to grey scale image watermarking the application of the considered wavelet transforms and embedding algorithms are extended to the case of video and colour image watermarking.EThOS - Electronic Theses Online ServiceGBUnited Kingdo