Medical Image Encryption Algorithm Based On Memristive Chaotic System And Compressive Sensing

With the rapid development of network communication technology and digital medical image technology,telemedicine has entered people's lives,and it has solved the problem of seeking medical treatment without leaving home,and has been highly concerned by people.However,medical image involve personal privacy as well as sensitive information.During their transmission and storage over public network,there are insecure problems such as theft,peddling and tampering.Therefore,it is especially important to protect the security of medical image information.In addition,large size and high-volume image transmission reduces transmission efficiency.Therefore,how to ensure the security and efficient transmission of medical image information has become a topic studied by many scholars,and image encryption is an effective technology to protect image information security.a secure and efficient medical image encryption algorithm is designed for medical image information protection,which has high research value.Medical image has strong neighboring pixel correlation,high redundancy and large data volume.Traditional encryption methods,such as Data Encryption Standard(DES)and Advanced Encryption Standard(AES),are designed for text information and are not suitable for medical image encryption.The development of chaos theory provides enlightenment for medical image encryption,chaos is a nonlinear system with complex pseudo-randomness.The memristive chaotic system is highly sensitive to initial values and control parameters,and the sequence is generated with highly random and unpredictable.At the same time,the cryptosystem is designed with memristive chaotic system is easy to implement,and the image encryption speed is better than the traditional encryption method,which is very suitable for medical image encryption of large data volume.According to the relationship between the plaint image and the obtained cipher image size,the image encryption is divided into completely encryption and compressiv encryption,if the plain image and the obtained cipher image have the same size,it is completely encrypted;if the plain image size is larger than the obtained cipher image size,it is a compressive encryption.In this paper,under the scrambling-diffusion framework of chaotic image encryption,the study and research are carried out.The main contents are as follows:(1)A medical image encryption algorithm based on Latin square scrambling and two-way diffusion is proposed.Firstly,the Latin square matrix is generated from the index sequence of the chaotic sequences,then the plaintext pixels are scrambled by the Latin square matrix and the plaintext information,then the scrambling matrix is rotated clockwise,and finally the two-dimensional diffusion is performed to obtain the cipher image.Since the initial value of the memristive chaotic system is highly correlated with the plaintext information,the proposed encryption algorithm has the ability to resist known-plaintext and chosen-plaintext attacks.Simulation results and security analysis verify the feasibility and effectiveness of the encryption algorithm.(2)A medical image encryption algorithm based on 3D orthogonal Latin square is proposed,which includes three stages of preprocessing,scrambling and diffusion.Firstly,a random number is generated by the 4D memristive chaotic system and embedded in the plain image,and then the AXM(Alphabetic xor method)operation is performed on the embedded data image to obtain a pre-processed plain image.Then,an index sequence is generated from the chaotic sequence,and a 3D orthogonal Latin square is generated by the index sequence and the control parameters.Then,the pre-processed plain image is converted into a 3D bit matrix,and the 3D orthogonal Latin square is directly scrambled and diffused to obtain a final cipher image.Simulation results and security analysis prove the security and robustness of the encryption algorithm.(3)A medical image encryption algorithm combined game-of-life-based scrambling and compressive sensing is designed,which uses scrambling-compressive-diffusion encryption process.First,the plain image is subjected to discrete wavelet transform(DWT)to obtain a sparse coefficient matrix.Then,the GSMHDP(Game-of-life-based scrambling method of high dependent on plaintext)is used to scramble the sparse matrix,and then the scrambled matrix is performed.The measurement and quantization processes are performed,and finally the diffusion operation is performed to obtain a noise-like cipher image.Among them,the chaotic sequence is obtained by the 5D memristive chaotic system and used in the above three encryption processes.Simulation results and performance analysis show that the encryption algorithm has a certain security and compressive.