Research On Medical Image Digital Watermarking Algorithm Based On Curvelet Transform

With the advent of the era of information and big data,revolutionary changes have taken place in many fields.The medical field is already developing in the direction of data and informatization.The emergence of patient electronic medical records(EPR),medical images and other data greatly facilitates the communication between doctors and colleagues and promotes the development of medicine.However,the surge in the amount of data year by year makes hospitals and medical institutions more willing to store data on third-party cloud platforms.In the process of transmission and storage of these medical data on the cloud platform,it is vulnerable to attacks by criminals,causing the leakage of patient privacy.Currently,medical data privacy leakage problems occur frequently at home and abroad.With the rapid development of medical data,how to effectively protect the privacy of patient information has become a topic that urgently needs to be solved by researchers.Digital watermarking technology was originally to solve the copyright protection of digital media.Later,it was applied in the medical field.The patient's private data was embedded as a watermark in medical images,and then the watermarked medical images were stored on a cloud platform.These invisible watermarks The problem of patient privacy leakage is well solved.For this reason,many medical image digital watermarking algorithms have been proposed.However,through a large number of literature readings,it is found that these algorithms have better robustness in the face of conventional attacks,but have poor robustness in the face of geometric attacks.How to improve the algorithm's ability to resist geometric attacks,and how to best balance the algorithm's robustness and imperceptibility has always been a research hotspot.This paper presents a research on a robust zero-watermark algorithm based on Curvelet.The main researches are as follows:A zero-watermarking algorithm for medical images based on Curvelet-DCT and RSA pseudo-random sequence is studied.First,the curve wave Curvelet and discrete cosine DCT transform are applied to the medical image,and then in the characteristic region of the transformed low-frequency coefficients,the characteristic vector is obtained by perceptual hashing technology.Then use the RSA pseudo-random sequence to perform spatial scrambling encryption on the watermark.Through the zero watermark technology,the embedding and extraction of the watermark are completed.Experimental simulation results show that the extracted watermark has a high correlation with the original watermark when faced with different degrees of disturbances from geometric attacks and conventional attacks,and the algorithm has a strong ability to resist conventional attacks and geometric attacks.A zero-watermarking algorithm for encrypted medical images based on Curvelet-DCT is studied.The first is to encrypt medical images in the frequency domain.After the medical image is transformed by DWT-DCT,the low-frequency coefficient matrix is obtained.The logistics pseudo-random sequence is subjected to perceptual hashing and dimension transformation to obtain a matrix for encryption.The matrix and the low-frequency coefficient matrix are multiplied by dots to obtain the encryption coefficient matrix,and IDCT and IDWT are performed on it to obtain the encrypted medical image.Then extract its features,and complete the watermark embedding and extraction through the zerowatermark technology.The experimental simulation results can be known.The algorithm not only has high robustness,the watermark system can still effectively extract the watermark when faced with different degrees of geometric and conventional attack disturbances,but also can encrypt and protect the patient's medical images.