시각화 기법을 이용한 악성코드 탐지 방법

Abstract

The number of malwares found on the internet increases significantly. To detect a number of malwares fast, Image Matrix techniques is introduced, which can be analyzed visually and calculate the similarity fast. This paper introduces the advanced Image Matrix technique by introducing Representative Image Matrix. This technique reduces the amount of samples should be calculated to detect malware. Two methods to generate Representative Image Matrix is suggested, Selected Representative Image Matrix and Mixed Representative Image Matrix. A method to keep the Representative Image Matrix updated is also introduced. A measurement of running time without and with applying Representative Image Matrix is presented to prove the effect of Representative Image Matrix. Updating overheads is also compared between Selected Representative Image Matrix and Mixed Representative Image Matrix.