Research On Image Compression-Encryption Using Compressive Sensing And Chaos Theory

Digital image,which is a significant aspect of multimedia information,has played a very important role in peoples' daily life because of its large amount of information,rich content and good comprehensibility.However,in the process of storing and transmitting digital images,some serious security problems are easy to arise.At the same time,a large amount of redundant data in digital images needs to be eliminated to reduce storage space and improve transmission efficiency.Therefore,how to maintain a high compression ratio while ensuring the security of digital images has become an urgent problem in the field of information science.Cryptography is an important means to ensure the security of multimedia information.In recent years,with the continuous development of chaos theory and its close relationship with cryptography,chaotic cryptography has emerged.Compared with traditional cryptography(such as DES,AES,etc.),chaotic cryptography is more suitable to provide efficient protection for multimedia information.With the rapid development of compressive sensing theory and its related applications,image compression-encryption technology based on compressive sensing has attracted extensive attention from researchers.Compressive sensing can realize image compression and encryption simultaneously,and the computational cost of compression and encryption process is low.Compressive sensing framework is often combined with chaotic cryptography to improve the security and compressibility of compression-encryption algorithms,while reducing storage space and transmission costs.In this paper,on the basis of the research of chaos-based image encryption technology,we further explore the image compression-encryption methods based on compressive sensing,and propose two chaos-based image encryption algorithms and two compressive sensing-based image compression-encryption algorithms using compressive sensing,chaos theory and the inherent characteristics of digital image.Finally,the validity,security and compressibility of the proposed algorithms are verified by the experiments.The innovations of this paper are summarized below:(1)In order to solve the problem of low plaintext sensitivity in chaos-based image encryption algorithms,a key scheme based on weighted summation is designed on the basis of analyzing the existing key schemes.Compared with the existing key schemes,the new scheme uses the characteristics of chaotic system to calculate the pixel summation of plaintext to update the initial keys,thus providing a strong plaintext sensitivity and high computing efficiency for encryption algorithm.On the basis,an image encryption algorithm based on key scheme is proposed by combining a bit-level confusion algorithm based on Sudoku matrix and a lightweight diffusion algorithm based on chaotic system.Simulation results show that this encryption algorithm has high security and can effectively resist the known plaintext attack or the chosen plaintext attack.(2)On the basis of the above research,key scheme and substitution operation are combined to improve encryption efficiency while enhancing plaintext sensitivity.First,a simpler key scheme based on weighted summation is given.Then,after analyzing the encryption process of traditional substitution methods and the characteristics of substitution results,a random substitution algorithm based on S-box is designed.The ciphertext image obtained by the new substitution algorithm has a uniform distribution of pixel values.Thus,the problem that traditional substitution methods can not resist statistical analysis attack is solved.Finally,according to the above work,an image encryption algorithm based on key-substitution architecture is implemented.Experimental results show that the proposed encryption algorithm has high encryption efficiency and security.(3)To solve the problem of low plaintext sensitivity caused by quantization operation,an image compression-encryption algorithm based on compressive sensing and S-box is proposed.In this algorithm,all the initial keys in the substitution operation after quantization are updated by the measurements before quantization,so as to enhance the ability to resist the known plaintext attack or the chosen plaintext attack.Experimental results show that this algorithm is sufficiently sensitive to slight changes in plaintext,its security is further enhanced,and its compressibility remains unchanged.(4)Aiming at the problem of low compressibility in image compression-encryption algorithm based on compressive sensing,lossless coding is introduced to compress the higher bit-planes based on the analysis of the inherent characteristics of digital image,while the lower bit-planes are processed by lossy coding based on compressive sensing.In this way,the compression ratio of the compression-encryption algorithm can be increased.Therefore,combining with chaotic encryption technology,a lossless compression-encryption algorithm based on Huffman coding and a lossy compression-encryption algorithm based on compressive sensing are designed to compress and encrypt the higher bit-planes and the lower bit-planes respectively,so as to implement an image compression-encryption scheme based on compressive sensing and Huffman.Simulation results show that the proposed compression-encryption algorithm can improve the compression ratio while guaranteeing security.