A novel image encryption scheme based on multi-directional diffusion technique and integrated chaotic map
Milad Riyahi,Marjan Kuchaki Rafsanjani,Rouzbeh Motevalli
DOI: https://doi.org/10.1007/s00521-021-06077-5
2021-05-16
Neural Computing and Applications
Abstract:In recent years, many researchers have used chaos maps to encrypt digital images due to their unpredictability, ergodicity, and sensitivity to initial values. This paper introduces a novel image encryption algorithm based on an integrated chaos map and bit-level diffusion technique. First, a novel multi-directional diffusion technique is applied to bit levels of the plain image. The innovative technique shifts bit levels of the plain image vertically and horizontally to eliminate visual information of the plain image. The proposed technique is so strong that is able to create high level of diffusion in just one round of implementation. Second, the integrated chaos map, which is a combination of Sine, Tent, and Logistic maps, is applied to the diffused image to encrypt the image. Also, for the objective of evaluating the performance of the proposed method, this paper presents new evaluation criteria, namely ClosenessNPCR and ClosenessUACI, in the field of number of pixel changing rate and unified average changed intensity, respectively. This paper proves that the novel evaluation criteria are more accurate than the preceding one, which is the mean of the achieved values. The implementation results show that the proposed image encryption scheme reaches a high security level, even more than preceding approaches.
computer science, artificial intelligence