Journal of Information Security and Applications 45 (2019) 117–130 Contents lists available at ScienceDirect Journal of Information Security and Applications journal homepage: www.elsevier.com/locate/jisa A secure image encryption scheme based on cellular automata and chaotic skew tent map Bhaskar Mondal a,∗ , Shrey Singh b , Prabhakar Kumar b a Xavier School of Computer Science and Engineering, Xavier University Bhubaneshwar, Orissa, 752050, India b Department of Computer Science and Engineering, National Institute of Technology Jamshedpur Jamshedpur, Jharkhand, 831014, India a r t i c l e i n f o Article history: MSC: 37B15 68P25 65P20 Keywords: Cellular automata Chaotic skew tent map Image encryption Security a b s t r a c t This paper presents a highly secure image encryption scheme for secure image communication and stor- age. The scheme is based on a chaotic skew tent map and cellular automata (CA). The chaotic skew tent map is used as the initial vector generator for the CA which needs a 128 bit initial bit sequence to gener- ate pseudo random number sequence (PRNS). Using the PRNS, the plain image pixels are first permuted to remove the high correlation among the adjacent pixels in the image. Then the permuted image is en- crypted based on a single random number generated by the chaotic skew tent map. The chaotic skew tent map has a very large key space, on the other hand, CA generates PRNS faster than a chaotic pseudo random number generator (PRNG). Therefore, a combination of both chaotic map and CA gives a sys- tem with higher key space and faster PRNS generator. The experimental results show a good encryption effects which are capable to resist any kind of known attacks. © 2019 Elsevier Ltd. All rights reserved. 1. Introduction Over last few years, several cryptographic schemes have been proposed based on confusion and diffusion. The confusion and dif- fusion based encryption schemes involve two basic phases; first phase is confusion in which the pixels positions are scrambled to destroy inter-pixel correlation and the second phase is diffusion which involves some reversible operations that alter the pixel val- ues [1]. The confusion and diffusion processes may be applied for m and n rounds such that m, n > 0. The confusion and diffusion processes use pseudo random number sequences (PRNSs). PRNS are generated by chaotic maps [2,3]. There are several techniques to generate PRNS. For example, a Chebyshev chaotic map is used to generate PRNS [4], which stores the eight bit planes separately (assuming 8-bit gray level images) and handles a matrix of size eight times the size of original im- age. Then the PRNS permutes the pixel bits. For permutation, this scheme needs to sort the PRNS which adds additional computa- tional overhead for sorting the long chaotic sequence. In the sec- ond part, the scheme again permutes the bit plain based on an Arnold’s cat map. But, the number of rounds needed to achieve the results is not mentioned. As the scheme permutes the bit planes twice, it generates a diffusion effect on encrypted image. As the ∗ Corresponding author. E-mail address: bmondal@xub.edu.in (B. Mondal). permutation-only ciphers are weak to the cryptanalysis, the overall scheme results in high complexity and low security scheme. There- fore, another scheme which involves permutation and diffusion is needed. For example, in [5], Mondal and Mandal proposed an im- age encryption scheme based on cross-chaotic logistic maps and DNA computations. The cross-chaotic logistic maps are used for generating PRNSs which are used for permutation of image pixels. and DNA computations are used for diffusion of the pixel values. It has been proved that in all permutation-only image ciphers, re- gardless of the cipher structure, the correct permutation mapping is recovered completely by a chosen-plaintext attack [6]. The Fu’s scheme [4] and Rahman’s Scheme which are permutation-only ci- pher were successfully broken where the plain-images are of size M × N and L different color intensities, the number of chosen plain- images n required to break the permutation-only image encryption algorithm is n = ⌈(log L (MN)).(O(n.MN))⌉ [6]. In [7], Premkumar and Anand proposed a Secured and compound 3-D chaos image encryption using hybrid mutation and crossover operator. In this proposed scheme arithmetic crossover and multi-point crossovers operator are used for permutation of the image. By using arith- metic crossover, two new individuals sequences are produced and a new sequence is decomposed into bit planes. Multiple crossovers are used to swap these bit-planes. In substitution they have used hybrid mutation i.e. boundary mutation for replacing the array size from top to bottom, non-uniform mutation to add pixel row wise, column-wise or diagonally, and uniform mutation to replace the pixels by circular shift. A 3-D CSI map is used to produce an https://doi.org/10.1016/j.jisa.2019.01.010 2214-2126/© 2019 Elsevier Ltd. All rights reserved.