Research On The Crucial Techniques Of The Image Information Security Under Regional Medical Collaborative Service

Abstract:The target of regional medical collaborative service is to realize the share of medical resources and diagnostic information among hospitals. Its major work is to create an information sharing platform in a region, which platform uses a hospital have the advantage of medical resources and information construction as the leader, can radiates community hospitals and grassroots’medical institutions. To construct the regional collaborative medical service system has a bearing on people’s health rights and interests, has become an important part of medical reform in our country, and it has been pay more attention by the government. It helps to promote the reform of public hospitals, and promote equality of medical service, solve the problem of "difficult to see a doctor, expensive to undertake medical treatment". It can provide a safe effective, convenient, and affordable medical and health services for the people.However, the most important reason for impacting the regional medical service system is really applied into practical, market-oriented and accepted by the masses of ordinary people is the information security problems. Due to the lack of security in public network, medical information transmitted during network is facing enormous challenges, such as the problem of patient’s privacy information protection, copyright protection and integrity authentication of medical information are needed to solve. The introduction of digital watermarking technology is a good solution to the security problem of the series.This paper uses medical image as the research object, which is the main component of medical information, and utilizes lossless digital watermarking technology as the main tool to do some research works, which focus on the key problem of secure transmission of the medical image information under the network environment. This work hopes to provide the powerful technical support of constructing the regional medical collaborative service system. The main works of this paper are as follows: (1)For the patient’s privacy protection in medical image, a medical image reversible information hiding method based on region and histogram shifting is proposed. The proposed method takes full account of medical image’s regional features, uses a maximum between-class distance of single threshold segmentation method to obtain the region of interest, and then utilizes the polymeric polygonal approximation algorithm to get the approximate polygon of region of interest. Finally, an image model fitting method is utilized to get the regular embedding region. For region of interest, the adapted difference histogram circular shifting method is employed to embed electronic patient records and other privacy data. On the other hand, for region of non-interest, an improved histogram shifting method based coding is utilized to embed the original medical image’s hash authentication codes. The proposed algorithm realizes high capacity embedding of electronic medical record and effective authentication of extracting watermark.(2)To address the problem of copyright protection of medical image, a lossless robust watermarking algorithm for medical image’s copyright protection is proposed. For the medical image with multiple channels, the existing lossless robust watermarking algorithm based on block difference is improved. The histogram compression technology is introduced to prevent overflow and underflow, and reduces the watermark payload; the Huffman encoding is utilized to encode the block identifier for reducing the block type identifier length, thus, the watermark’s embedding capacity is increased. In addition, the redundancy theory is introduced to embed the same watermark information in R, G, B three color channels respectively, and the voting mechanism is employed to judge the final extracted watermark to improve the robustness of watermarking. The algorithm can withstand a certain degree of JPEG compression and noise attack, realizes medical image’s copyright protection in lossy transmission environment.(3)For the integrity authentication problem of medical image, a tamper detection method based on quad-tree decomposition for medical image is proposed. Aiming to the shortcomings of the method based on fixed block, the quad-tree decomposition method is employed to divide the medical image into blocks with variable size, and a linear weighted interpolation is selected to computer the feature values of each block. These feature values as watermark information is embedded by a simple reversible integer transform based on chaos. Compare with the existing fixed size method, the proposed method has a larger embedding capacity. In addition, the chaotic system enhances the security of the algorithm. For the tampered area with complex texture, using the linear weighted interpolation as feature values is prior to the mean method in view of image quality and tamper detection accuracy. The proposed can be used for medical image’s tamper detection and recovery.(4)For the preventing overflow and underflow problems in medical image lossless digital watermarking algorithm, a thorough analysis and contrast experiment of the existing solution methods are given, the effects produced by the different methods on performance of reversible watermarking algorithm is illustrated. In order to solve the overflow and underflow problems, a histogram compression method is proposed. Analysis and experimental results points out that the preventing overflow and underflow algorithm is the key performance of medical image reversible watermarking, which depends on the choice of reversible watermarking algorithm and the different modality of medical image.