Hands-On Artificial Intelligence for Cybersecurity

Detecting malware from images with CNNs

From the original paper described previously, a tool was developed that leverages CNNs to recognize and classify the images that represent malware codes.

The tool can be downloaded from the GitHub repository, by executing the following command:

git clone https://github.com/AFAgarap/malware-classification.git/

Inside the archive, there is also a dataset of images of malware codes (malimg.npz). To convert your malware codes to grayscale images, you can also use the Python script developed by Chiheb Chebbi, which is available at https://github.com/PacktPublishing/Mastering-Machine-Learning-for-Penetration-Testing/blob/master/Chapter04/MalwareConvert.py.

We show some examples of the tool's usage as follows: ...