A Novel Image Encryption Scheme Based On Hybrid Cellular Automata

Digital images are rich in information.With the development of computer information technology and cloud computing,people are increasingly sharing images via the Internet or transferring from local devices to the cloud for convenience.The openness and sharing of the network environment seriously threaten the security of images during transmission.Digital image encryption,as a basic means to ensure information security,is facing the challenge of high image data volume and strong two-dimensional spatial correlation.The existing image encryption algorithms mainly include traditional data encryption algorithms(such as AES,etc.),image encryption based on chaotic systems,image encryption based on cellular automata or DNA encoding,and so on.The traditional encryption algorithm designed for text information encryption is difficult to cope with the characteristics of high data throughput and information redundancy in image encryption,and the efficiency and effect of encryption are not ideal.Image encryption based on chaotic system will suffer from degradation of characteristics such as short periods when implemented on digital devices;when implemented on analog devices,parameter mismatch will occur due to noise interference.Image encryption based on DNA is lacking in security and flexibility,which makes it difficult to use independently.Finite state machine with feedback shift register and cellular automata as special cases have unique advantages in the field of cryptography because of their simple structure,locality,and complex dynamics.The linear feedback shift register is the most widely used,but its linear structure is vulnerable to attack,and the theoretical research of the non-linear feedback shift register is immature.Cellular automata are composed of multiple cells.It has complex dynamical characteristics in time and space.It is fast and easy to implement.Hybrid cellular automata has the advantages of localized operations and high parallelism.It also overcomes the shortcomings of uniform local rule,has more flexible local rules,and the maximum period can be obtained by using the theory of finite fields.This paper proposes an image encryption scheme based on hybrid cellular automata.First,a hybrid cellular automaton with an optimal period is constructed.Combining with the non-linear feedback shift register and substitution-permutation network,a high-throughput pseudo-random number generator based on the hybrid cellular automata is designed.The generator passes NIST randomness test.Subsequently,this paper uses the above-mentioned pseudo-random number generator as a component,combine the hash function and the output feedback scheme in the block cipher to design an image encryption algorithm.Compared with the traditional encryption scheme,this scheme has a high bit utilization rate and an optimal period.Performance tests and analysis of results show that the key space of this scheme can support up to 2²⁵⁶ bits,which is enough to resist brute force attacks and has high key sensitivity.The information entropy is tested to reach 7.9994,which is very close to the ideal value of 8,and is better than most schemes.In addition,it also performs well in resisting linear attacks and differential attacks.In terms of encryption speed,it only takes 0.0063 seconds for an image of 256×256 size.Compared with most solutions,the encryption speed has been greatly improved.Therefore,this scheme has both fast encryption speed and high security performance.The main innovations of this paper are:1.Design and implement a hybrid cellular automata algorithm with optimal period;2.A high-throughput pseudo-random sequence generator is constructed based on the hybrid cellular automata.Based on this,a safe and efficient image encryption algorithm is designed and implemented.