Image encryption algorithm based on ElGamal cryptography and selective random diffusion

Abstract

Some current image encryption schemes are independent of the plaintext, leading to a vulnerability to the chosen-plaintext attack under symmetric structure. To address this issue, an asymmetric image encryption algorithm with plaintext correlation is suggested, after analyzing various image encryption schemes using chaotic systems. First, a three-dimensional New Logistic-Sine Map (NewLSM) is designed by coupling Logistic map and Sine map, considering them as seed maps. Analyses prove that the NewLSM has a wider continuous chaotic interval and more complex chaotic behavior than seed maps. Secondly, to enhance the keystream associated with plaintext, a new key acquisition model is constructed, i.e., NewMKG, by combining the hash function SHA-3 with the public key ElGamal cryptography. Then, employing the NewLSM and the NewMKG, a new image encryption algorithm with asymmetric structure is presented using classical framework of confusion-diffusion. In particular, a new diffusion method is proposed after confusion, namely, selective random diffusion (NewSRD). Moreover, experiments and analyses indicate that a good performance can be reached for the proposed algorithm. For example, the value of information entropy for the obtained cipher image is close to eight.