Study On Data Hiding Algorithm For Encrypted Image

With the rapid development of the network communication, cloud computing, distributed processing and multimedia technology, more and more users depend on network to transmit and process information. Digital image is widely used due to its vivid and intuitive features. However there is potential safety hazard in the transmission, due to the information is easy to be stolen, tampered maliciously, replicated illegally at any time. Therefore, ensuring the security of digital image, especially the ones with confidentiality is an imperative problem to be solved.In order to achieve the security of information in transmission and processing, a large number of research results are proposed. Cryptography and information hiding technology are the two most important branches. Cryptography makes the content invisible to protect the security of information, and data hiding technology makes the existence invisible to protect the security of information, the combination of the two can achieve the confidentiality and integrity of the image protection. At the same time, data hiding in encrypted image which has wide application prospect can protect the image security, carry on effective control and management of the image.In this thesis, a technique for data information hiding for encrypted image, based on the cryptography and data hiding technology, has been discussed thoroughly. The main contents and results are as follows:Firstly, the algorithms of image encryption and data hiding are summarized, and the existing data hiding algorithm based on encrypted images is studied and analyzed.Secondly, in view of the problem of low embedding capacity of existing algorithms and the problem of unstable effect when applied to medical image, a novel reversible data hiding method in encrypted medical images based on slash grouping and gradient operator is proposed. After encrypting the medical image by bitwise exclusive-or operator, the encrypted image is divided into blocks and each block contains four specific groups. Then each block is embedded two bits by flipping three least significant bits of corresponding group. With an encrypted medical image containing additional data, it is firstly decrypted using the encryption key. Then the image restoration and data extraction can be achieved according to the data-hiding key and the evaluation function. The experimental results show that the original medical image can be perfectly recovered and the embedded data can be successfully extracted. When the block size is the same, the method offers double embedding capacity, higher PSNR of stego-image and lower extracted-bit error rate.Thirdly, combination of block compressed sensing and the separating the block's edge pixels, an improved data hiding algorithm for encrypted images based on block compressed sensing is proposed. Firstly the original image is divided into non-overlapping blocks, then the lower right edge pixels in the block are encrypted by XOR operation, and the remaining pixels are encrypted by block compressed sensing, the two parts are combined into the cipher text to send to the channel owner. Then the channel owner embeds the secret information into the cipher text according to the information hiding keys. The receiver extracts information and restores the original image according to the encryption and information hiding keys. The experimental results show that the improved algorithm has higher embedding capacity, the lower extracted-bit error rate, and the quality of finally restored image is better.