An Image Encryption Algorithm Based on Self-Adaptive Method and Hash Function

In this paper, a novel algorithm for image encryption based on self-adaptive method and hash function is proposed. In the suggested self-adaptive method, an image is equally divided into four sub-images, all of which are encrypted by using the masks that is generated by other part. This will significantly increase the resistance of the cryptosystem against known/chosen plaintext and differential attacks. In the next process, the salsa20 hash function is modified to generate a key stream. Then the final encryption key stream (by correlating the keystream and plaintext) is used to encrypt 8×8 block images in only two rounds of diffusion process that resulting in both key sensitivity and plaintext sensitivity. Experimental results such as key space analysis (>2512), differential analysis ((NPCR>99.61% and UACI>33.45%)), information entropy (entropy 7.9975) and … show that the proposed image encryption algorithm is secure and reliable against the all attacks.