On the security of an image encryption scheme

This paper studies the security of a recently-proposed image encryption scheme based on chaos, and points out the following problems: 1) there exist a number of invalid keys and weak keys, and some keys are partially equivalent for the encryption/decryption processes; 2) given one chosen plain-image, a sub-key K10 can be guessed with a smaller computational complexity than that of the simple brute-force attack; 3) given O(10) (at most 128) chosen plain-images, a chosen-plaintext attack may be able to break the following part of the secret key: ({Ki mod 128}i=4,K10), which works very well when K10 is not too large; 4) when K10 is relatively small, a known-plaintext attack can be mounted with only one known plain-image to recover some visual information of other plain-images encrypted by the same key.