Research Of Medical Image Security Technology Based On Lossless Digital Watermarking

The rapid development of network technology and the establishment of Digital Imaging and Communications in Medicine (DICOM) standard, promotes the exchange of digital medical imaging information, at the same time, it also cause medical data may be lost or tampered and other information security issues that arise in transit. Therefore, the certification to the authenticity and integrity of the medical data is very important. As digital watermarking technology is a kind of emerging information security technology that can embed private information directly in the carrier information, it widely used in the field of digital information protection and authentication. In order to certificate the integrity of medical image, and avoid lead any distortion into image, to design a kind of lossless watermarking algorithm that can non-destructive restore the original image and certificate the image is of great significance.In this thesis, the valuable research fruits can be summarized as follows:(1)A new kind of watermarking algorithm based on companding technique, the difference expansion technology with error-correcting code technology is presented. The algorithm uses the correlation between the adjacent gray values in a image, to embed the information through the expansion of the difference between adjacent pixels. And use companding technique to control image distortion, and improve the compression rate of the binary location map get in the difference expansion process, thus increase the watermark embedding capacity. At the same time, use error correction code can improve the self-recovery ability of the whole image.(2)Improve a region-based medical image lossless watermarking algorithm, adding tamper recovery part. First of all, determine the region of interest, and set it as a certification region, obtain authentication information as part of the watermark information; In the region of non interest (RONI), automatic select the embedding region; Achieve the watermark embedding in the RONI through the difference expansion of the adjacent pixels; After extract the watermark, to achieve tamper detection and localization through comparing the authentication information, to achieve tamper recovery through error correction code technology. Ultimately, the watermarked image distortion level can be adjusted by entering the threshold. Compared with the previous algorithm, due to without embedding the compressed binary location map, the watermark capacity is no longer subject to binary image compression ratio, and to a certain extend the capacity improved, and this method can limit the embed distortion in the non-interest region, the watermarked image can be directly used for medical diagnosis or other medical application, to facilitate the actual use of watermarked medical images.