The Research Of Information Hiding And Detection Algorithms Based On Images

With the rapid development of computer networks and multimedia technology, information security has become a society-wide demand. As an aspect of information security, Information hiding technology has got much attention in recent years. There are two main reasons for this. One is illegal copy problem with audio and video products on the Internet, so there are needs for the corresponding copyright protection technology. The other is that digital media is rich in information, which can provide a tremendous space for information hiding.As the opposite of information hiding technology, information hiding detection technology is an important research direction in information security field. Information hiding detection not only helps improve information hiding technology, but also it prevents the abuse of information hiding technology. This thesis is about information hiding and information hiding detection algorithm, the main contents are as follows:The characteristics and applicabilities of information hiding algorithms are evaluated by realizing classical information hiding algorithm in spatial domain and frequency domain. LSB algorithm has the advantage of large embedding capacity and is adopted in many types of software. But histogram of image varies before and after LSB embedding. LSB algorithm has the disadvantage of poor security. So far there are many improved LSB algorithms based on histogram keeping, but these improved algorithms will still change the histogram and the compensation algorithm has smaller capacity than classical LSB algorithms. To overcome the above problem, a dynamic modification and compensation algorithm for information hiding is proposed, which can reduce the histogram changes due to the information hiding and at the same time it won't reduce the embedding capacity.Many information hiding detection algorithms are evaluated by theoretical analysis and experimental verification. Because there is no effective information hiding detection algorithm for detecting LSB matching algorithm, a detection algorithm for LSB matching is proposed. Correlations of bit planes are evaluated before and after LSB matching embedding. The proposed algorithm detects LSB matching, using the gradient energy function of bit planes as measurement parameters, by analyzing the variation of gradient energy of each bit plane before and after the embedding process. The dynamic modification and compensation algorithm and the detection algorithm for LSB matching based on gradient energy proposed in this thesis are asserted by the result of simulations.