Abstract

Automated code morphing techniques can make malware recognition difficult. In this article researchers at the University of Lafayette propose a method that can be used to decide whether a binary is a variant of a known item of metamorphic malware by treating the morphing engine as an author.