The Research On Methods Of Betrayal Checking And Traitor Tracing For Confidential Images On The Internet

With the development of the Internet, a great amount of electronic files including some confidential image documents have being produced by enterprises and institutions in their processes of informatization. Due to the easiness of coping and spreading electronic images, the issue of leaking of secret information is often encountered. Our research aims to develop technologies about betrayal checking and traitor tracing for the purpose of enhancing protection of confidential images. Such technologies are of great theoretical and practical importance to the information security of enterprises and government departments.It is inevitable to face a sea of data when performing betrayal checking on the Internet, and data clustering is unavoidable when dealing a sea of data. Considering a dataset of three objects, if two pairs of the objects are similar but one pair is dissimilar, this will lead to errors in the clustering result of this dataset. To address this problem, we define the concept of IT (Intransitive Triangle), and design a new clustering method which can eliminate the errors coming from ITs if three conditions of the new method are satisfied. We show that the clustering problem addressed by this method is a NP problem. We define the concept of missing error and commission error in clustering and establish standard for two types of errors. It is obvious that conditions of missing error are more complex than the conditions of commission error, and we design the algorithm to calculate missing error.With the help of technologies of digital fingerprint and digital watermarking, we propose a method of confidential image betrayal checking based on digital fingerprint embedded in the images; we design the relevant algorithm and analyze the performance. If there are many users involved, the efficiency of the system is relatively low, so we bring forward the concept of organizational watermarking. The algorithm performance is greatly improved by using organizational watermarking instead of digital fingerprint. Codeword matching is involved in both the method of betrayal checking based on digital fingerprint and the method based on organizational watermarking. To achieve a better performance, we design a method of codeword matching based on LSH (Locality Sensitive Hashing) clustering. We first use LSH to cluster the whole codeword dataset and then just focus on a few codewords from the subsets which is concerning.For the purpose of efficient traitor tracing, it is necessary to select a digital fingerprint solution which can defense collusive attack coming from mutli-users. We propose a two-level concatenated code system which takes group-oriented orthogonal code based on BIBD (Balanced Incomplete Block Designs) as inner code and which takes the error correct code-RS (Reed-Solomon) code as outer code. We design inner code (using BIBD code) and build the inner matrix, with the encode algorithm being given and performance analysis being performed. We describe the methods of encoding and decoding of RS code. We bring forward the encoding method of concatenated code and performance analysis, and show the validity of the fingerprint solution. Compared with famous Boneh-Shaw's code and Dittmann's code used in the same collusive attack situation, our fingerprint code achieves the same effect of defensing with a shorter codeword's lengthWe design a prototype system of betrayal checking and traitor tracing which can be used to protect confidential image in corporations and institutions, and we describe:how to gather images from the Internet using mobile agent, how to evaluate the importance level of the websites on the Internet using their relevant properties, how to adjust the scan scheduling strategy based on their importance levels, how to rapidly find images on suspicion of being confidential based on organizational watermarking, how to detect the source of leaking using image's character matching, and how to quickly exclude the innocent users and trace traitors. In addition, we put up with the principle and method of sub-sector dividing and users grouping, and describe the procession of pretreatment of confidential image. In the end, simulation experiment is done to show the usability of the prototype system.In conclusion, our research is about betrayal checking and traitor tracing used for confidential image protection in the Internet. What we achieve in this research is of both theoretical and practical importance, which shall have positive significance for the protection of confidential digital image.