Hiding Information Based On Gray System Theory

This thesis stressfully analyzes the characteristics and sorts of grey relation analyzing theory, and successfully realizes information hiding methods based on them. These technologies effectively improve the security while transmission, enhance the detective difficulty of attacking and the arithmetic's capability. The main work and innovation are as follows:1. presenting a new image scrambling methodAttackers can easily find the datum rule because of pixels' correlation, so the image second-scrambling method is presented as pre-encryption technology. It includes two main facets-position scrambling and gray scrambling. In the position scrambling, it is presented an unsymmetrical scrambled matrix that can modify the image texture. Compared to Arnold transformation, this method improves the scrambling effect meanwhile reducing the calculation. In the gray scrambling, the thesis adopts the modified Gray coded matrix. After the operation, it can change the image's gray histogram distributing, and reduce the correlation of the pixels, furthermore this method increases the periods of the operation and attacking difficulties. Consequently, the two scrambling methods are both used in one secret image so that the period is enormously increased and the method's capability is obviously enhanced.2. presenting a space domain information hiding method based on Deng's GRA model The main characteristic of space domain information hiding technology is that therealization is convenient, but the drawback is that it's poor in resisting the attack. So the thesis innovates the hiding arithmetic in two facets in order to improve the capacity. Firstly, the choice of embedding method based on Deng's GRA model is presented, as the result it can find the insensitive position for human visual system in the cover image. Secondly, the parity hiding method and grey relation grade are combined in order to form a new embedded arithmetic, each pixel's compared array that is defined in the GRA model is treated as the area for parity detection. When the secret bit is not match with the LSB parity, the middle pixel of the area will be modified. If the data accord, it does not need any operation.3. presenting a DCT domain information hiding method based on variable coefficient GRA modelBecause the transformation domain information hiding methods are more robust and are well in resisting the attack, this thesis mainly researches the DCT domain blind detection hiding arithmetic, and the variable coefficient GRA model is. used for choosing the embedded position, deciding the embedding method. The great innovation in the hiding technology is: the DCT domain coefficient matrix is departed into two parts-the integer matrix which is utilized for calculating the grey relational grade and decimal matrix which is used to detect whether the secret bit match the relative magnitude of choice data or not. If the datum is not match, the two position's decimal should be exchanged. The new hiding method ensures the randomicity of embedded position in each matrix of whole image, and increases the difficulty of attacker's destruction. In the other hand, a significant characteristic is that this technology hardly modifies the matrix value because the decimal changing can only cause little wave. Consequently, this method greatly enhances the information security while transmission since the attackers can infrequence detect the existing of secret information.