Cryptanalyzing a chaotic encryption algorithm for highly autocorrelated data

Recently, a chaotic encryption algorithm for highly autocorrelated data was proposed. By adding chaotic diffusion to the former work, the information leakage of the encryption results especially for the images with lower gray scales was eliminated, and both higher-level security and fast encryption time were achieved. In this study, we analyze the security weakness of this scheme. By applying the ciphertext-only attack, the encrypted image can be restored into the substituted image except for the first block; and then, by using the chosen-plaintext attack, the S-boxes, the distribution map, and the block of chaotic map values, can all be revealed, and the encrypted image can be completely cracked. The improvement is also proposed. Experimental results verify our assertion.