Robust Image Encryption Based on Improved Logistic Chaos and Cross Chaotic Diffusion

To address the problem of weak robustness of some current image encryption algorithms, this paper proposes a strong robust encryption algorithm. The Logistic chaotic system is improved to produce more random sequence distribution. The plaintext hash is used to construct the initial key of chaotic system and the spiral matrix transformation is used for the first scrambling. A novel chaotic sequence cross-diffusion method is constructed by introducing two values of plaintext correlation val1 and val2. Through plaintext transformation of chaotic sequences, it improves the plaintext sensitivity, which can resist the selective plaintext attack, and has strong robustness. In order to avoid the position of some pixels not changed in the first spiral matrix transformation, cross-scrambling is carried out again after cross-diffusion. Therefore a scrambling – diffusing - scrambling encryption framework is built. The experimental results show that the proposed algorithm can resist brute force attacks and selective plaintext attack. It has strong robustness and high security. Especially it has perfect performance in anti-shear attacks, which has significance for image security transmission and is very suitable for image encryption.