Error-correcting graphs for software watermarking

In this paper, we discuss graph-theoretic approaches to software watermarking and fingerprinting. Software watermarking is used to discourage intellectual property theft and software fingerprinting is used to trace intellectual property copyright violations. We focus on two algorithms that encode information in software through the use of graph structures. We then consider the different attack models intended to disable the watermark while not affecting the correctness or performance of the program. Finally, we present several classes of graphs that can be used for watermarking and fingerprinting and analyze their properties (resiliency, data rate, performance, and stealthiness).