Malware Detection Method Based on Visualization

The rapid development of information technology and computer networks has led to the emergence of various new applications on both PC platforms and mobile devices. Malware continues to evolve and update, which often developing new variants or changing existing features to evade detection. Traditional feature based malware detection methods are limited in their ability to detect variants, and are computationally resource-intensive. Considering these issues, a new visualization-based and integrated malware detection method, Mal_Vis, is introduced. It decompiles the application software and applies PCA to reduce the feature dimension, then visualises the decompiled data to greyscale and RGB image. A Stacking-based ensemble machine learning algorithm is used to classify the visualized images to detect malware. Experiments show the method achievs detection accuracy of 98.19% and 93.03% in the Windows and Android application software datasets.