Researtch Of Adaptive Steganography Based On Minimal Distortion Model And Embedding Codes

Steganography aims to hide secret messages using the public carrier to achieve covert communication. Current trends of image steganography is adaptive steganography, namely steganography select the embedding position and define the distortion characteristics according to the carrier, so as to modify the embedded cover in safer areas to focus on in order to improve the security of steganography. Coding and distortion by the definition are two parts of adaptive steganography, one is to reduce the number of embedding changes as much as possible, while the other features that characterize the carrier well. JPEG image, the most popular digital images on Internaet,are used as steganography carrier in this thesis. It Focuses on how to make steganographic encoding has more practical and more secure JPEG steganography. Research results are as follows:1. Constructs a dynamic distortion adaptive steganographic coding based on pixel-chain. In order to solve the failures in the presence of STCs embedded ternary situation and large computational complexity, this thesis pixel chain technology and current best STCs steganographic encoding combined to reduce the number of modified embedded object is achieved by a special modification mode, thereby increasing the efficiency of the embedding, and reduces the computational complexity. Experiments show that the coding efficiency at the expense under the premise of a small embedded greatly reduce the computational complexity of the situation and embedding failure does not exist, with a strong practical.2. Presents a minimal two-dimensional embedding distortion model and JPEG steganography based on it. In order to solve the existing problem of various issues on interaction in JPEG image compression on, a minimize dimensional distortion model is proposed. By defining the vertical and horizontal distortion distortion and specify the longitudinal and transverse load capacity embedding ratio, calculated theoretically Optimal modification probabilities. Its application to have a hidden side information write JPEG, through experiments show that the algorithm has a high resistance to detection capabilities.3. Propose a JPEG image edge information to predict recovery algorithm and its application to the boundless information JPEG steganography. In the same length of the message embedded conditions, side- informed JPEG steganography has a higher security than the JPEG steganography without side information. If we can recover a portion of the side information, apparently it can increase JPEG steganography security. By C haracteristic calculated of the DCT transform used in JPEG compression, the relationship between the energy of the image and the DCT coefficients has been found. Based on the relationship, a side information prediction recovery algorithm is proposed. By experiment on the performance of the recovery algorithm and security of JPEG steganographic algorithm based on the recovery algorithm, the experiment result shows that the algorithm has a certain resilience and can have a more significant upgrade on security at large embedding rate.