A Malware Detection Method Based on Rgb Image

Abstract

In recent years, with the development of the Internet, information security has become the focus of our attention. With the advent of the era of big data, the detection of large-scale malicious code has attracted a lot of researches' attention. For solving the problem, we propose a malware detection method based on operation and data flow of instructions, which is used by malicious code. It combines the operation and data flow of the instructions being used by malware, then reflects itself in an rgb image. Then, it uses the convolutional neural network that has advantages in image processing for deep-learning to detect the rgb image of malicious code. We have carried out a series of experiments. And through these experiments, it is proved that this kind of rgb image, which is generated by the fusion of the operation and data flow of instructions used by malware, could be well applied to the detection of malicious code. The experiment shows that the highest detection accuracy could be as high as 97.95% and the false positive rate could be as low as 2.618%.