Advanced Malware Detection Using Convlutional Neural Networks: A Visual Approach
INTERNATIONAL ENGINEERING AND TECHNOLOGY MANAGEMENT SUMMIT 2024, Ankara, Türkiye, 17 - 19 Ekim 2024, ss.55-58, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Basıldığı Şehir: Ankara
- Basıldığı Ülke: Türkiye
- Sayfa Sayıları: ss.55-58
- Ankara Üniversitesi Adresli: Hayır
Özet
In the rapidly evolving field of cybersecurity, the detection and classification of malware are essential for protecting systems and networks from malicious attacks. Traditional signature-based malware detection methods often struggle to keep up with the increasing sophistication and variety of malware. This project explores the use of advanced machine learning techniques, specifically Convolutional Neural Networks (CNNs), to address this challenge by transforming malware samples into visual representations for more accurate and efficient identification and classification. The primary objective is to train a CNN model to accurately predict and classify malware samples based on their grayscale image representations. This involves developing a robust CNN model that distinguishes between benign and malicious software with high precision, identifying distinctive patterns of various malware families through their visual representations, and enhancing the speed and reliability of malware detection processes. The dataset consists of 5,386 samples, including 2,929 malware and 2,457 benign samples, represented as grayscale images. Each byte of a file is depicted as a pixel value, capturing intricate structural and behavioral patterns crucial for effective classification. By leveraging this innovative approach, the project aims to improve cybersecurity defenses and provide a dynamic solution to the growing threat of cyber attacks.