The Design And Implementation On DCT Domain Of Quantization Watermarking Based On Feature Modulation

With the rapid development of Internet technology, it’s more and more fast and convenient to transmit multimedia information like digital images, audio and video.etc.. However, it is easier to get, modify, store and disseminate multimedia information without the author’s consent. During the transmission, digital media suffered some intentional or unintentional attacks, the receiver cannot determine whether the media information received is complete or true. Now, the fourth-generation mobile networks has been very popular, the world has entered the self-media time, people can publish multimedia information on the web at all times and places. A sharp increase occurs in the amount of information on the Internet, it also does in amount of people who demand for copyright protection.Digital watermarking technology hides the information in the text, images, video or audio without affecting the quality of these multimedia, this technology makes it come true that the owner of copyright, integrity and authenticity of multimedia information can be certified. Digital watermarking involves various disciplines and has become the focus of research at home and abroad. This thesis focused on how to implement a kind of watermarking algorithm which is based on random angle quantization modulation and its theoretical performance is excellent in DCT domain.This thesis introduced the research background, research status, general knowledge and typical algorithm about digital watermarking, analyzed the algorithm that will be used and explored the practice based MATLAB. First, we discussed whether the algorithm was implemented in the overall or block of DCT domain. According to comparison of the experimental results, we decided to implement the scheme in the DCT block. Secondly, we discussed what kind of frequency coefficients from each block should be host signal and how long the host signal is. Then, with the analysis of the JPEG compression algorithm, we determined where the optimal position in host signal is to embed the watermarking. This thesis combined Watson model to deduce a new embedding algorithm what made watermark algorithm self-adaptive, so as to enhance the imperceptibility and robustness of watermark. Finally, we summarized the watermarking scheme of watermark in detail. Besides the theoretical description, this thesis made a lot of simulation experiments on the actual picture, and compared performance of this scheme with other classic ones. Experimental results indicated that the implementation scheme we got is more imperceptible and robust.