A Deep Learning-Based Approach for Malware Classification Using Machine Code to Image Conversion

S. Yaseen; M. M. Aslam; M. Farhan; M. R. Naeem; A. Raza

No Detail Found!

A Deep Learning-Based Approach for Malware Classification Using Machine Code to Image Conversion

Author(s)

S. Yaseen Department of Computer Science, COMSATS University Islamabad, Sahiwal Campus, Pakistan

M. M. Aslam Department of Computer Science, COMSATS University Islamabad, Sahiwal Campus, Pakistan

M. Farhan Department of Computer Science, COMSATS University Islamabad, Sahiwal Campus, Pakistan

M. R. Naeem Department of Computer Science, COMSATS University Islamabad, Sahiwal Campus, Pakistan

A. Raza Department of Computer Science, COMSATS University Islamabad, Sahiwal Campus, Pakistan

Abstract

Malware presents a formidable threat to computer security, with some antivirus companies reporting over five million malware samples per day. Due to the sheer volume of malware, security teams cannot feasibly address each instance individually, necessitating the use of malware classification schemes. The size, type, and complexity of malware continue to grow, as hackers and attackers often create systems that can automatically transfer and encrypt code to evade detection. Classic machine learning techniques that rely on handmade feature vectors are not effective for classifying malware families. In contrast, deep convolutional neural networks have demonstrated efficiency in detecting and classifying malware. This article proposes a new system for classifying malware into families by transforming malware binaries into grayscale images and applying convolutional neural networks. Our approach provides a new method for multi-class classification challenges in the field of cybersecurity, and it outperforms traditional machine learning techniques. By utilizing deep learning models to classify malware, we can enhance our ability to detect and mitigate threats in real-time. Our work emphasizes the potential of advanced machine learning techniques in cybersecurity and calls for further research in this area.

Publication Details

Page(s) 36-46

DOI DOI not available

Published Journal: Technical Journal, Volume: 28, Issue: 1, Year: 2023

Request the full text from PASTIC

Keywords

Cybersecurity, Deep learning CNN, Machine code, bytes code, Image conversion, - Malware classification