Breaking a chaotic image encryption algorithm

  • PDF / 4,955,009 Bytes
  • 21 Pages / 439.642 x 666.49 pts Page_size
  • 47 Downloads / 286 Views

DOWNLOAD

REPORT


Breaking a chaotic image encryption algorithm Saeed Noshadian1 · Ata Ebrahimzade1 · Seyed Javad Kazemitabar1 Received: 7 June 2018 / Revised: 8 June 2020 / Accepted: 24 June 2020 / © Springer Science+Business Media, LLC, part of Springer Nature 2020

Abstract Image encryption is an important issue in information security. Due to sensitivity to initial values, randomness and simple computation, chaos-based image encryption algorithms efficiently provide security. In a recent paper a chaos-based image encryption was proposed that employs genetic algorithm to optimize correlation between adjacent pixels. We show via simulations a security exploit that can find the key and decrypt the image. To fix the the aforementioned vulnerability we harness Knuth shuffle where we replace the rand function with the logistic map. We then evaluate the security of our proposed algorithm by testing the impact of several attacks on it. We show that while improving the security, our algorithm keeps the same level of pixel entropy as the previous method. Moreover, the correlation coefficient among adjacent pixels of the cipher image was reduced by an order of magnitude. Keywords Image encryption · Chaotic · Genetic algorithm · Knuth shuffle · Logistic map

1 Introduction Information security has gained a lot of attention among researchers in recent years. Cryptography has been the major player in providing security for stored and transmitted data. Several encryption algorithms such as DES, IDEA, AES and RSA have been proposed and studied in the literature [6, 8, 35, 36]. Due to the inherent features of image, however, such as large volume, high redundancy, low entropy and high correlation among adjacent pixels conventional algorithms are not always proper for image encryption [8, 11, 22, 27, 31– 33, 38]. Two principal techniques to avoid high redundancy and high correlation in image encryption are confusion (or substitution) and diffusion (or permutation). In substitution, pixel values go through complex changes that in effect increase the complexity of encryption. In permutation, the location of pixels changes which leads to an increase in redundancy. While each of these techniques alone is a good encryption method, their combination will create an even stronger cipher.

 Seyed Javad Kazemitabar

[email protected] 1

Babaol Noshirvani University of Technology, Babol, Iran

Multimedia Tools and Applications

In the literature, different encryption methods are introduced. One of these methods is the quad tree algorithm [11]. While this method reduces processing time, it is only suitable for the quad tree compression algorithm, which is not a global standard. Another method for image encryption is cellular automata (CA) [10, 19]. In [7], cellular automata is shown to be suitable for the production of the bit stream of the key in Vernam cipher. In [9], an image encryption method based on the permutation of the pixels and replacement of the pixel values is discussed. The permutation is performed by scan patterns that were generated by