API 호출 빈도 기반 소프트웨어 버스마킹 시스템

초록

In this paper, we propose a software birthmarking system using frequencies of API-calls. The birthmark is representative features of a program, which can identify program. We use the frequencies of API-calls used in a Windows program as a birthmark of the program. The system forms the frequencies of API-calls as a vector each of which element is the frequency of the corresponding API-call. Weight of each API is assigned as it's IDF score. Then the system computes the similarity between two programs through cosine similarity and Pearson correlation coefficient. When the similarity of two programs is higher than a given threshold, our system regards the two programs as plagiarism candidates. Through intensive experiments, it was found that the proposed system outperforms the conventional birthmarking system.

키워드