Plagiarism Detection for Java Programs without Source Codes

This paper presents a novel dynamic analysis approach to software plagiarism detection. Such an approach is inherently more resilient to code obfuscation techniques such as renaming of program entities, reordering of statements, etc. We develop our technique in the context of a dynamic analysis and visualization system for Java, called JIVE, but the techniques are applicable to other object-oriented languages. Our analyses are based on the execution traces of Java programs (produced by JIVE), and our experimental results confirm that this approach is both efficient and effective in detecting plagiarism of Java programs when their source codes are not available.