Cryptanalysis of a pixel permutation based image encryption technique using chaotic map

Abstract

Recently, a new image encryption algorithm was proposed by Anwar and Meghana. This encryption scheme uses the Arnold's chaotic cat map to permute the image pixels. The resulting image is then confused with both a secret image provided as part of the secret key and a secret value selected randomly from the permuted image. Using a random image as part of the secret key, gives this algorithm an infinite key space which increases its efficiency against brute-force attacks. In order to help improving the security of modern image encryption schemes, this paper presents a cryptanalysis of the proposed algorithm using a combination of chosen-plaintext and brute-force attacks. First, the infinite key space that the secret image offers is broken using a chosen-plaintext attack. Then, the permutation phase is reversed through a series of chosen-plaintext attacks too. Finally, the secret value chosen randomly from the permuted image is easily brute-forced due to its reduced number of possible values. By applying the above method, it is possible to restore the plain version of any image that was encrypted using the former encryption algorithm. Thus, relying on this algorithm to encrypt real sensitive data is not secure.