The Research Of Image Steganography Based On Syndrome Trellis Codes

Steganography aims to transmit secret information unnoticeably through public channel, and is becoming one of the important research subjects in the field of current information security. Image is the most frequently-used cover form of steganography. Research on image steganography has both academic and applied value. In this thesis, based on Syndrome Trellies Codes(STC), DCT domain and spatial domain steganography has been studied. The main contributions are as follows.1. After analysis of steganography model, steganography characteristic and current image steganography situation, some problems unsolved on this area are pointed out.2. A DCT domain steganography algorithm based on high efficiency code is approached. Combining wet paper codes and STC, a high efficiency code is proposed, which can improve the Steganographic Capacity, while keeping the changing number of DCT coefficients. By modulating the changing probability, statistic abnormity is avoided. Experiments show that the capacity of this algorithm is higher, while the undetectability is stasitfied.3. Based on STC. a spatial domain steganography algorithm minimizing distortion function is approached. Firstly, distortion function is defined, which scales the influence of modification position, mode and number over undetectablity. Then the problem of improving undetectability transforms into the problem of minimizing distortion function. During embedding, suppose the weight of pixel is the distortion in pixel-grade. Then by finding the path of minimizing total weight, the stego-image of minimizing distortion function is made. Experiments show that this algorithm resists the steganalysis effectively.4. A system of image steganography is designed and implemented. The system satisfies the demanding of covert communication, and can be expanded.Finally, the researching work is summarized, and further research topics and directions are prospected.