| Recently, with the rapid development of computer communication technologies, digitalinformation is being applied widely in all fields of the society. Digital image, one of the mostwidely used and important expression forms of information, has many advantages, such asintuition, vividness and imaginal performances. Therefore, the security of image information isbecoming the world’s attentive focus. The study of image encryption is of profound theoreticaland practical significance, because image encryption is not only the key technology in imagesecurity, but also indispensable in the fields of information hiding and digital watermarking.Based on chaos theory and characteristics, this dissertation focuses on the analysis andstudy of anti-attack of typical chaotic image cryptosystems. Meanwhile, based on the analysisabove, several new image cryptosystems are proposed. The main achievements in thisdissertation are as follows:1. An improved image encryption based on pseudorandom matrix is proposed.The security of an image encryption scheme with a pseudorandom permutation is analyzed,which belongs to a class of image encryptions based on permutation-diffusion mode. Firstly, ifthe elements of the plain-image matrix are the same, the cipher-image matrix is an all-zeromatrix. Therefore, the receiver cannot decrypt correctly and the plain-image cannot be recovered.Secondly, if the plain-image is an all-zero matrix, the cipher image is an all-zero matrix too.Thus the attacker can get the plain-image without any computing, because the plain-image is justthe cipher image. Thirdly, when the control parameter is relatively small, statistics attack can beused to reveal some visual information of any other plain-images encrypted by the same secretkey. Finally, when the cycle number is, each plain-text word is only related to cipher-textwords, but independent of other cipher-text words, which cannot meet the principles of algorithmdesign-diffusion and confusion. In order to solve the problems of this kind of algorithm, animproved image encryption based on pseudorandom matrix is proposed according to fourprinciples of image permutation. And the security is verified by differential analysis, correlationanalysis, sensitivity test and information entropy analysis.2. An image encryption algorithm based on hyper-chaos is deciphered successfully and anew cryptosystem based on hyper-chaos is developed.The chosen plaintext attack and chosen ciphertext attack are given to an image encryptionalgorithm based on hyper-chaos. Many cryptosystems have weak anti-attack performance facing chosen plaintext attack and chosen ciphertext attack, and generally the solution is cipher blockchaining (CBC) that makes the serial operation necessary in encryption process. However, theparallel operation is useful to improve operation speed and is valuable to encrypt large data.Therefore, in view of the computer operating system in which the serial and parallel operationscoexist currently, we propose a new image encryption based on hyper-chaos. The proposedalgorithm not only can solve the problem of weak anti-attack performance, but also can supportboth serial and parallel operations. Theoretical analysis and simulation results show that thealgorithm not only can resist chosen plaintext attack and chosen ciphertext attack effectively, butalso can obtain better statistical characteristics and difference characteristics.3. A block image cryptosystem based on discrete wavelet transform is presented.A block image cryptosystem based on discrete wavelet transform (DWT) is presented.Image preprocessing based on DWT is applied to originality image to decrease key sequencesand improve efficiency. Image scrambling in low-frequency sub-image can make each transitionpoint in frequency domain have influence on all data set in space domain rapidly. Then based onthe requirement of digital image, the preprocessed image is grouped according to lossyprocessing or lossless processing. Some binary key streams generated by logistic map are usedfor group scrambling and the other key streams are used for substitution. And the iterationinterval of key streams is controlled by sub-band group to resist chosen plaintext attack andchosen ciphertext attack. Theoretical analysis and simulation results show that the algorithm canhide statistical characteristics, improve encryption efficiency and resist chosen ciphertext attack,statistical attack and difference attack effectively.4. A self-adaptive parallel encryption algorithm based on discrete2D-Logistic map isdeveloped.A self-adaptive parallel encryption algorithm based on discrete2D-Logistic map isdeveloped according to the position scrambling and diffusion of multi-direction in variablespace of spatial chaos. The binary sequencesb1b2b3...bn are obtained according to the user key, inwhich the binary sequence0and1denote distribution mode of processors and the number ofbinary sequence n denotes cycle number. Then the pseudorandom2D matrix is generated by2D-Logistic map, and adaptive segmentation is applied in image matrix and pseudorandommatrix according to the value and the number of binary sequence. The parallel operation is usedamong blocks to improve efficiency and meet real-time demand in transmission processes.However, the pixel permutation is applied in partitioned matrix through ergodic matrixgenerated by pseudo-random matrix-block to decrease the correlation of adjacent pixels. Then the pixel substitution is used for fully diffusing through cipher block chaining mode until ncycles. The proposed algorithm can meet the three requirements of parallel operation in imageencryption and the real-time requirement in transmission processes. The security is proved bytheoretical analysis and simulation results.Finally, the research works of this dissertation are summarized, and the future researchdirections are indicated. |
PDF Full Text Request