Watermark Embedding And Detection

The popularity of Internet access has enabled the wide spread of digital multimedia contents in the form of image, video and audio; however, it also makes unauthorized copying and distribution easier. Researchers from both the industry and the academy have been trying to address this dilemma by watermarking techniques. Watermarking, as a prospective weapon against piracy, embeds ownership information into the host contents without degrading the perceptual quality of the host contents. The embedder and the detector (or decoder) are the two most important components of the digital watermarking systems. Thus in this work, we discuss how to design a better embedder and detector (or decoder). Spread spectrum (SS) and quantization techniques are the two most appealing embedding techniques in the literature.We first explore the optimum detector or decoder according to a particular probability distribution of the host signals. The optimum detection or decoding is not new in this work since it has already been widely investigated in the literature. However, our work offers new insights into its theoretical performance. First, the theoretical analyses presented in the literature are unreasonable since their analyses (with the watermark sequence as a random vector) are not in accordance with the prerequisite in their deriving the optimum decision rules that the watermark is a fixed sequence. Second, we found that for Multiplicative Spread Spectrum (MSS) schemes in both Discrete Cosine Transform (DCT) and Discrete Fourier Transform (DFT) domains, its performance depends on the shape parameter of the host signals. Third, without perceptual analysis, the Additive Spread Spectrum (ASS) scheme also has a performance dependent on the host signals and outperforms MSS at the shape parameter below 1.3.For spread spectrum schemes, the detector or the decoder's performance is reduced by the host interference. Thus, we came up with a new host-interference rejection idea for MSS schemes. In this work, we call this new host interference rejection idea an Enhanced Multiplicative Spread Spectrum (EMSS) scheme. Moreover, in our scheme, we also consider the probability distribution of the host signals, and match the embedding rule with the optimum detection or decoding rule. We particularly examined EMSS's performance in the DCT domain and found that it produced a nicer performance than the traditional non-rejecting schemes. Furthermore, this scheme can be easily extended to the watermarking in the Discrete Wavelet Transform (DWT) or the DFT domain.Though the host interference rejection schemes enjoy a big performance gain over the traditional spread spectrum schemes, their drawbacks that it is difficult for them to be implemented with the perceptual analysis to improve the fidelity of the watermarked contents discourage their use in real scenarios. Thus, in the final several chapters of this work, we introduced an innovative idea—double-sided embedding and detection to tackle this drawback. This idea differs from the host interference rejection technique in that it does not reject the host interference at the embedder. However, it also utilizes the side host information at the embedder. Moreover, for most of the spread spectrum methods, the detector reports the existence of the embedded watermark if its response is above a given threshold. However, our double-sided detector reports the existence of the embedded watermark if the absolute value of its response exceeds a given threshold. Though our technique does not reject the host interference, it can also achieve a great performance enhancement over the traditional spread spectrum schemes. Most important of all, it has a big advantage over the host interference rejection techniques in that we can embed the watermark with the perceptual analysis to achieve a maximum allowable embedding strength.In summary, this work contributes in1. Theoretical performance analyses for optimal detectors in both the DCT and the DFT domains, and finding that the performance of MSS depends on the shape parameters of the host signals,2. Proposing a host interference rejection scheme whose embedding rule is tailored to the corresponding optimal detection or decoding rules,3. Most important of all, presenting a new model of watermark embedding and detection—double-sided embedding and detection.