Research On The Key Technologies Of Digital Fingerprinting

With the rapid development of network, more and more electronic documents appear in people’s work and daily life, and become an indispensable part of people’s life. While they bring people lots of convenience, electronic documents also cause many security problems. Because of electronic documents can be tampered easily and copied freely in a large scale as well as the Internet have great ability of spread, the problem of plagiarizing and piracy is becoming more and more serious and it has brought huge impact to the copyright protection. As a new type of digital copyright protection technology, digital fingerprinting technology has its unique digital rights tracking mechanism, which becomes a solid shield for digital copyright protection. For text documents protecting, it is very fast to generate digital fingerprints and detect plagiarism text, which provides effective protection to academic plagiarism preventing and copyright protection. In this paper, we mainly research text digital fingerprint and its key technologies. The detail methods and contents of the research are as follows.1. Researching on text fine-grained selection strategyText fine-grained selection has a great influence on text similarity detection accuracy. In this paper, we design an experiment scheme to study two different fine-grained selection strategies deeply, overlapping selection strategy and non-overlapping selection strategy. We analyze the experiment results got from the tests of different plagiarism styles, and determine a more proper fine-grained selection strategy.2. Researching on digital fingerprinting generation algorithmAs a critical step of digital fingerprinting system, the performance of digital fingerprinting algorithm directly affects the quality of digital fingerprinting system. In this paper, we design different schemes to test the performance of MD5, SHA1and Rabin algorithm in different file size. Results indicate that MD5algorithm perform better while generating digital fingerprints.3. Researching digital fingerprinting selection strategyA proper or an excellent fingerprinting selection strategy can achieve more accurate detection result, even though we get relatively less fingerprints. In this paper, we test four different digital fingerprinting selection strategies, full-fingerprinting, all-substring, hashed-breakpoint and lowest-hash-value, and determine the best fingerprinting selection strategy in different application environments through analyzing the experiment result.4. Designing a new digital fingerprinting scheme and implementing it with programmingBased on the previous studies, we design a new digital fingerprinting scheme for text plagiarism detection, and implement it with programming. Then, we analyze the functions and shortcomings of the system. The new digital fingerprinting system can not only meet the demand of different encoding types, but also can provide different fingerprinting selection strategies, which can meet the demand of high efficiency and high accuracy of different file sizes in similarity detection. But it still needs to strengthen the quality of text preprocessing in the system.