Research On The Technology Of Digital Watermarking In Medical Images

With the continuous innovation of digital technology in China at this stage,and people’s design of medical image technology system,it has greatly broadened the depth and breadth of the development of digital medical image technology.While digitization provides convenience for the access of medical information,the data security risks are becoming increasingly prominent,which greatly reduces the personal information security of patients.Facing the problem of how to more effectively protect the personal safety of patients in the modern medical environment,digital watermarking technology can integrate private data into medical images,which has become the key means of integrity verification and tamper detection of medical images.For this reason,the following work has been carried out:（1）The basic knowledge of digital watermarking is introduced and summarized in detail,including the concept,development history,system structure model,characteristics,classification,attack methods and performance evaluation criteria of digital watermarking.（2）A fragile medical image watermarking algorithm based on prime distribution theorem is proposed.Firstly,an approximate pixel set is constructed according to the prime distribution theorem,and then a random sequence is generated through the chaotic system and hash function to give the parity of pixel values again,so as to enhance the security of the watermarking algorithm.The embedding and extraction of watermark is to adjust the pixel value of the host image according to the parity of pixel value and the parity of prime number contained in pixel value by using logical operation.The algorithm can resist conventional attacks,has high security and low algorithm complexity,has good tamper detection and location ability,and has good practicability.（3）A digital watermarking algorithm for medical images based on Lorenz chaotic system is proposed.In view of the different requirements of medical image on the visual quality of interest regions（ROI）and uninterested regions（RONI）,in order to avoid the serious distortion of ROI of medical image caused by embedding watermark,the algorithm adaptively selects the embedding intensity with the help of information entropy.The highlight of the algorithm is that the Lorenz chaotic system and hash function are used to determine the embedding position of the watermark,which improves the security of the algorithm.At the same time,the live code is used to bind the patient information and its corresponding medical image,which not only avoids the confusion of information,but also greatly improves the flexibility and embedding capacity of the watermark algorithm.Experimental results show that the algorithm can effectively resist conventional attacks,has strong robustness,and the extracted live code can be decoded accurately.