Research On Digital Image Encryption Based On DNA Sequences And Compressive Sensing

Since ancient times,information security has been an attractive and practical topic.Along with the rapid development of cloud computing and extensive application of Internet technology,various types of data are transmitted over the network.Especially,due to the characteristics of intuitiveness and fast transmission,the digital images gradually become one of the important information carriers for people's daily communication and are widely applied in various fields such as Internet communication,medical treatment and military etc.Unfortunately,the information during transmission is prone to be leaked or illegally intercepted because of the wide openness of the Internet and some defects in its protocols.Therefore,the fundamental measure is to encrypt the information to be transmitted.In this thesis,several image encryption methods are designed based on some excellent features of DNA sequences,for instance,huge storage capacity,highly parallel performance,Extremely low energy consumption and so forth.Due to the intrinsic features of the digital images such as large amount of information,un-uniform energy distribution,high data redundancy and so forth,the color image encryption algorithms are proposed by using DNA sequence operations,chaotic systems,and the confusion and diffusion technologies.Furthermore,since compressive sensing can simultaneously sample and compress the signals with lower rate and computational cost,an image compression-encryption algorithm based on compressive sensing and by using the chaotic systems to observe or measure the continuous signals.The main research work of this thesis are summarized as follows:1.A new color image encryption scheme is proposed by using CML based NCA mapping,DNA sequences and one-time pad.Firstly,the initial values and system parameters of the NCA-CML are updated by using the SHA-256 hash function and the initial secret keys.The keystreams are generated by the plain-text image and secret keys.Then,the color image is decomposed into three color component matrices,i.e.,red,green and blue.Based on the DNA encoding rules,three color matrices are converted to DNA sequence matrices,respectively.Further,three DNA matrices are merged into a new DNA matrix,and the new matrix is scrambled by using the keystreams.Next,the confused DNA matrix is divided into three blocks with equal size.The DNA sequence operations are performed on these blocks.Finally,the DNA matrices is decoded to obtain the encrypted image according to the DNA decoding rules.In order to improve the security of the cryptosystem,the encrypted image is further diffused by using the keystreams.Thus the resulting cipher-image is obtained.Experimental results and security analysis show that the presented algorithm has good encryption effect and can resist various common image attacks.2?A lossless color image encryption scheme based on DNA sequence and information entropy is introduced.In the first stage,together use two simple improved chaotic systems and the information entropy of the plain-image to produce the pseudorandom keystreams.Next,transform the plain-image into the DNA sequence matrices by the DNA encoding rules,and utilize the binary keystreams to shuffle the DNA matrices.Thirdly,implement a row and column diffusion processes on the confused DNA matrices by the keystreams and DNA XOR operation.Eventually,convert the DNA matrices into the encrypted image by the DNA decoding rules,and perform a cipher-text diffusion in crisscross pattern to enhance the security and sensitivity of the encryption scheme.The final cipher-image is attained.Experimental results and security analysis demonstrate the excellent performance of the proposed image method.3?A double color images encryption algorithm based on block compressive sensing is designed.First,generate chaotic sequences by the plain-images and three improved chaotic systems.Then,the plain-images are decomposed into the red,green and blue grayscale images.The NDWT technology is employed to sparse the image blocks and three sparse matrices are obtained.In what follows,the chaotic systems are iterated to obtain the required measurement matrices,and three sparse matrices are measured by compressive sensing.One can get the measurements.Finally,the measured values are combined and further encrypted through the Latin square matrix,image diffusion and the keystreams.The cipher-image is finally obtained.The receiver can recover the plain-images by the decryption process and image reconstruction mehtod.Simulation results have verified the effectiveness and security of the proposed image cryptosystem.