A software birthmark based on weighted k-gram

Abstract

A software birthmark is the invariable features of a program that can used to detect software theft. Software theft can be detected by a birthmark that can cover the whole behavior of a program. To improve the performance of k-grams of resisting semantics-preserving transformations, and also consider a birthmark should cover the whole behavior of a program. A static birthmark based on k-grams and their weights is proposed. Weight is computed by analyzing the rate of change of k-gram frequency between the original program and the transformed programs. Experimental results show that the performance of the proposed birthmark for Java applications is better than static k-gram birthmark against transformations performed by Smokescreen obfuscator, ZKM obfuscator and Jarg optimizer.