Research On Copyright Protection For Medical Images Based On Digital Watermarking

In the modern medical health industry, the digital information system is playing a more and more important role. Digitization of medical images reduces usage and management of traditional films and improves the efficiency of images retrieval, and it’s also convenient for applying medical images in remote diagnosis, teaching and scientific researches. In order to acquire medical images, kinds of delicate and expensive imaging instruments are needed. However, it is fairly easy for attackers to intercept or copy for distributions illegally when medical digital images are transmitted in the public networks. Nowadays, the government pays more attention to constructing the medical information sharing platform, thus the copyright protection for the sharing of medical images becomes an urgent problem needed to be solved.Digital watermarking as an important branch of information hiding techniques can be applied to the protection of the confidentiality of medical images, the diagnosis hiding of patients, identification of integrity control for medical images, copyright protection and so on. The thesis mainly conduct the research on the copyright protection for the medical images based on digital watermarking, which provides an available method about the copyright protection for remote diagnosis, teaching and studying or something else. The main research contents include the following aspects:1. The thesis introduces digital watermarking related theory knowledge, characteristic and classification, and introduces the characteristics of medical images. Then, watermarking schemes related to medical images are introduced in detail.2. A discrete cosine transformation domain watermarking algorithm base on Chinese remainder theorem and region of non interest (RONI) is proposed. The algorithm uses the pseudo-random sequence to select low frequency coefficients in DCT domain dynamically, and embed the watermarking information in those coefficients. By dividing regions of interest, the algorithm can restrain the distortion of watermark embedding beyond it. What’s more, the algorithm could resist attacks from JPEG compression and adding noises. The thesis provides an available way for copyright protection for medical images in teaching and researching situations.3. An improved lossless robustness algorithm based on arithmetic differences of each blocks is proposed. According to the characteristic of arithmetic differences in medical images, adjusting the embedding strategy can enhance the transparence of watermarking. Meanwhile, the algorithm embeds the watermarking information into several channels according to the characteristic of multi-channels for color medical images, and then extracts the redundant watermarking information from separate channels in the extracting stage. At last, it generates watermarking information based on voting theory. The algorithm holds abilities of resisting JPEG compression, the Gaussian noise, salt and peppers noise attack, and it can be well applied to copyright protection for medical imaging in remote diagnosis situations.