Abstract

Developing a secure image encryption algorithm that can be implemented in devices with different hardware structures has remained a problematic issue, particularly for the real time applications. In this paper, we introduce two integer versions of chaotic maps, namely the quadratic chaotic map used to generate the parameters of the system and an integer version of Chebyshev chaotic maps used to generate permutations and chaotic sequences of integers. Then, we propose an image encryption algorithm that applies many rounds, each containing confusion and diffusion step in which a permutation is used to shuffle the pixels’ locations and the chaotic sequence is used to change the pixels’ densities. The proposed method is tested against many security measurements and some kinds of attacks for five standard images. It shows better performance over recent methods from the literature. The paper moreover discusses the performance of the decryption algorithm against the salt & pepper and crop attacks on the cipher image. It is found that the decipher image is seriously affected by high levels of noise or crop attacks. To overcome the limitations of the decryption algorithm at high levels of attacks, we proved first that the crop attack has same distribution as the salt & pepper noise attack. Then, an enhancement technique based on the use of median filter is proposed to reduce the effect of the attacks. It is shown from simulations that the enhancement decryption algorithm can double the image quality in terms of the PSNR.