Asymmetric image encryption algorithm using a new chaotic map and an improved radial diffusion

In order to address the security risks associated with the theft of private images over open Internet of Things (IoT), this paper introduces a new Logistic-Sine-Coupling Map (LoSCM) in three dimensions with a better chaotic behavior, and proposes an image encryption algorithm using an improved Radial Diffusion (ImRD) to achieve a better diffusion effect in image encryption process. Firstly, a preprocessed image is obtained through an XOR operation between the original plain image and a random generated matrix, followed by Secure Hash Algorithm 512 (SHA-512) performed on it, producing a plaintext information with length of 512 bits. Then, a parameter transformation model (PaTM) is constructed, and the plaintext key associated with the plain image is computed by the plaintext property. Secondly, by employing public-key Rivest-Shamir-Adleman cryptogram, i.e., RSA, the ciphertext key is obtained through the plaintext key and opened to public. Then, a new initial value acquisition model (InVAM) is designed to calculate the initial values of LoSCM, thereby generating the corresponding keystream. Finally, XOR, row scrambling, ImRD diffusion, and column scrambling are performed on secret plain image, followed by modular addition with a random matrix to get the final cipher image. Furthermore, the simulation results and performance analysis also prove that the new suggested image encryption algorithm in this paper can effectively oppose chosen or known-plaintext attack. The highlights are: (1) Construct chaotic map LoSCM with better chaotic behavior. (2) Design diffusion ImRD with better diffusion effect. (3) Build models PaTM and InVAM with high security to produce keystream.