Research On Chaotic Image Encryption Method Based On Decryption Location Definition And Self-recovery Strategy

Digital image encryption can transform the meaningful plaintext image into a disorganized noise ciphertext image by the encryption function and the key.It is used to protect the secret image in channel transmission and it plays an important role in image security.Chaos based image encryption algorithms are the hot topics in image security,although the existing encryption algorithms improve the encryption performances to some extent,however there are still many shortcomings:? Some encryption algorithms are mainly designed for the key security while Ingoring the limitations of the decryption location;?Some image encryption algorithms do not take into account the attributes of plaintext image to be encrypted,which results in different images experiencing the same encryption procedures with the same key,thus a malicious attacker can choose a special template to attack some encryption algorithms;?Although some image encryption algorithms consider attributes of the plaintext image to be encrypted,the encryption procedures are fixed and lead to poor security;?Traditional encryption methods always consider deciphering difficulties while neglecting the self-recovery abilities of the attacked encrypted images,although some images scrambling methods have some recover abilities and they can only provide limited abilities.To address these problems,the followings are the finished works:1)An image encryption and decryption method based on the poker shuffle and the fixed decryption range was proposed.In the proposed algorithm,firstly the minimum outsourcing of the effective decryption space was determined according to the ecryption position,and then the intermediate encryption parameters were generated by the random triangles which were produced by the minimum outsourcing center point,the MD5 and SHA-1 values of the plaintext image as well as the user key;Secondly,the initial parameters of Logistic mapping and Henon mapping were generated by the intermediate encryption parameters to drive Logistic map producing a large random image with the same size of the image to be encrypted;Thirdly,Henon mapping and random image were used to randomly select the row and column of pixels to be encrypted and then random poker shuffles and global exclusive-OR encryption were used to generate secret image.Compared with the existing image encryption algorithms,the proposed method limits the scope of the decryption position to improve the security of the intermediate parameter generation strategies,so that the encrypted image can only be decrypted within a specific location range.At the same time,the proposed method introduces attributes of the plaintext image and the random poker shuffle associated with random images,so that different images to be encrypted have different encryption procedures thus improves the security of the image encryption algorithm.2)A self-recovery image encryption algorithm combining zero watermarking and image block classification fusion is proposed.In encryption,firstly,the image blocks were divided into four categories according to the energies of the alternating coefficients in frequency domain and the corresponding AC/DC coefficients were preserved as the restoration reference matrix;Secondly,the deputy information of the reference matrix of the position scrambled blocks and the main information of the original blocks were fused,and then the authentication information of the fusion blocks were obtained according to the coding matrix;Thirdly,the DC component of the image blocks before fusion,the block types and the authentication information of the fusion blocks were scrambled separately as the zero watermarking stored in the third public party;Finally,the reconstructed image is reconstructed from the fused image and the global position scrambling was used to generate the encrypted image.In decryption,Firstly the fusion blocks were constructed,and then the primary and secondary information authentication relationship and the zero watermarking were used to restore the main information of the blocks and to reconstruct the decrypted image.Compared with the existing image encryption methods and the image scrambling methods based on the pixel position transformation,the proposed method can not only encrypt image effectively,but also recover decrypted image by channel attacks,thus the proposed method has higher application values.