Robust Watermarking For Medical Images Based On Arnold Scrambling And Transform Domain

With the progress and development of multimedia technology and digital image processing, digital imaging technology has been widely used in medical field. It is an indispensible and effective aid tool, using the X-ray computer tomography (CT) and magnetic resonance imaging (MRI), to diagnose in the modern medical areas, which not only can accurately reflect the detailed information of patients, but also provide the convenience to exchange information between one hospital and the other. However, when the public network realizes the transmission and communication, a new threat of security has been arisen such as falsification, tampering, disorder information and so on, therefore, in order to ensure the security while conveying some important information of medical images and hiding in ones, watermarking technology is put forward as its unique robustness and security, and plays an important role in identifying the authenticity and protecting the copyright, etc.In order to better guarantee the security transmission of digital medical information, in this paper, being aimed at the medical image and three-dimensional medical volume data, two algorithms, including robustness watermarking and multi watermarking, is proposed, which satisfies the requirements of strong transparency and high capacity for watermarking embedding, the innovations are as follows:1. Basing on the characters such as periodicity, reversibility, safety, and effective scrambling, the paper adopts Arnold scrambling technology at the stage of pretreatment watermarking to improve the security and robustness.2. The paper proposes a new algorithm using Arnold scrambling and the transform domain, utilizing to the HASH function, the scrambling watermarking combines with feature vector which is extracted from transform domain and has great robustness, then embedding and extracting of watermarking are realized. The experiment indicates that watermarking always can be extracted whether conventional attacks or geometric attacks, and own the strong robustness. In addition, it is successful for two-dimensional medical images and three-dimensional ones, and meets the visual requirements.3. A multi-watermarking algorithm is proposed, in this paper, which based on Arnold and transform domain. On the basis of the original algorithm, the embedding capacity of watermarking is expanded. This algorithm can achieve between two-dimensional and three-dimensional medical images, and the experimental data show that it still has better robustness and invisibility. Therefore, in the field of medical research, it has more research and practical value.These two algorithms proposed in this paper, which combines scrambling technology, zero watermark technology and traditional cryptography technology, the complex process of looking for a region of interest by human is removed, and watermarking information achieve the dual protection, furthermore, the extraction of watermarking does not require the participation of the original medical image. The high transparency requirements of watermark embedding for the medical image is sufficed, simultaneously, the security of the watermarking and the robustness of the algorithms are all improved.