Abstract:

As a new digital copyright protection technology, digital fingerprinting technology plays an important role in plagiarism detection. Still, the fingerprinting generation algorithm directly determines the performance of the digital fingerprinting scheme. In this article, we expound several kinds of commonly used digital fingerprinting algorithms—MD5, SHA1, Rabin fingerprinting algorithm, and introduce the basic theories of the three algorithms. At the same time, we test the performance of each digital fingerprinting algorithm through experiment. To test the difference of the three algorithms in performance, we create many different sample test files whose sizes between 20KB and 20MB. In the experiment, we preprocess all the test files at first, remove all the irrelevant characters. Then we generate digital fingerprints of preprocessed files with block strategy, and test the generation efficiency of the three algorithms in different block strategies and different size files. The result indicates that hash function has better generation performance in terms of large files, while in terms of small files, hash function and Rabin fingerprint algorithm has the same performance. The experiment results provide experimental support for selecting proper fingerprinting generation algorithm in the next new fingerprinting scheme.

Key words: digital fingerprinting, Hash function, Rabin algorithm, plagiarism detection