Anti-collusion Digital Fingerprinting Technology Research Based On Image

With the development of multimedia technology and the popularity of network, the copyright protection of digital multimedia is becoming increasingly important. It is free to distribute or change when the content is decrypted by the legal user, so encryption technique can’t realize the function of copyright protection. Digital fingerprinting is a key technique for copyright protection. It realizes the copyright protection through embedding the unique identification buyer’s information into original digital products. The buyer’s information include personal information or purchasing information. Then the issuer stores the information in the database. Once the illegal copy is found in the market, the issuer can extract the fingerprint and match it with the fingerprints in the database to find the distributor, then fine to prevent illegal copy. Because the digital products and the position of fingerprint embedded are the same, only the fingerprint is different, several legal users can compare their differs, find the position where the fingerprints are embedded and change the information in that position. This method is called collusion attack. So how to design anti-collusion digital fingerprinting scheme is a question to be resolved.This article focuses on the research of anti-collusion digital fingerprinting technology. First, it introduces the basic concept, the model, the common attack ways and detect methods of digital fingerprinting system. Then we introduce several classic digital fingerprinting schemes and analyse the advantages and disadvantages of them. On the base of existing anti-collusion code and detect method of good performance, We propose some improvements, the following is the main work of my paper.First:With increasing of the number of user, the orthogonal fingerprinting scheme need more orthogonal code and more storage useage. So we proposed a scalable orthogonal fingerprinting scheme which need less orthogonal code to support more users. From theoretical analysis and experiments, it is proved that this scheme maintains the original scheme’s good anti-collusion performance.Second:We propose a improved detection algorithm of orthogonal fingerprinting scheme. It is known that the statistics of collusion fingerprint under different forms of collusion attack and the performance of system. According to the statistical data, we propose a pretreatment technique of extracted fingerprints before detection. Through the experiment results, we find that the improved detection algorithm we proposed have higher probability of detect and lower the probability of false positive compared with the original scheme.