Maleware Classification from their signature

MLCS-WiSe2020 Universität des Saarlandes - Machine Learning in Cyber Security 2020 - Semester Project

Team Name: Alpha

Members:

1. H T M A Riyadh(s8htriya@stud.uni-saarland.de)
2. Fahad Hilal (s8fahila@stud.uni-saarland.de)

My part in this project

It is a two person group project. My part was to develop and implement [1] the program that classifies maleware from their signatures. I use CNN and image processing technique to achieve our result.

Problem Statement

The aim of the project is to perform visualization and classification of malicious code or programs commonly referred to as malware. The central idea is to combine deep learning (specifically Convolutional Neural Networks) and image processing techniques to achieve the objective [1].

Motivation

In the present age of technology, malware attacks are reported quite frequently. The main targets of these attacks being both financial institutions and everyday users. The damage perpetrated by such malware attacks could range from losing critical or personal data or failure in a nuclear power plant. Thus, our computing systems need protection from malware attacks round the clock. Automatic malware identification and detection tools are aimed at addressing this very problem.

Proposed Strategy

We leverage the power of CNNs for image classification. We first transform a malware signature into its binary representation and then convert this binary into an 8 bit vector followed by transformation to a grayscale image. The final step is to feed the image into our network for classification.

Reference:

1. Lakshmanan Nataraj et al. “Malware images: visualization and automatic classification”. In: Proceedings of the 8th international symposium on visualization for cyber security. 2011, pp. 1–7.