Improving Salsa20 Stream Cipher Using Random Chaotic Maps

The efficiency of a stream cipher is an important property that measures its robustness against the attacks and analysis. The robustness of the stream cipher based on the complexity of the generated keystream that is used to secure text and images via digital transition through open networks and such as emails and use images that are used in computers and phones. The randomness and complexity of the keystream had great attention in many studies to improve the stream ciphers. A new generation of stream ciphers based on the chaotic map is presented. This paper proposed a new method to generate a keystream based on the Salsa20 algorithm with a set of chaotic functions that are chosen depending on our previous evaluations. The selected chaotic map is implemented in the inner processes of Salsa20 to obtain more complexity. The encrypting text, gray images, and the colored image has been tested by Histogram comparison that resulted no informative data about plain image, three correlation of image in the vertical, horizontal and diagonal way of the plain and ciphered image also have been tested, that results larger than 0.15, plain and ciphered images are compared by the concept of peak signal to noise ratio and mean square error with no significant differences. The gray and colored images are tested by Structural Similarity Index (SSIM), information entropy with significant differences. Finally, the NIST is applied to the outputs. The result of the tests showed high randomness with p value of 0.9. the encryption time is in about 38 sec.