Malware Visualization Based on Deep Learning

Abstract

In this paper, we propose a new visualization analysis method based on the binary sequence of malware. The method uses SFCs (space filling curves) to visualize malware files and differentiates the displayable characters from non-displayable ones by different colors. This method resolves the problems that other methods cannot orient characters and shield analysis system from the ZipBomb attack risk aroused by huge malware. We randomly selected 7162 Kaspersky malware files and used the deep fusion networks to extract image signatures. Experiments obtained classification accuracy 98.24% and detection accuracy 99.02%.