Intelligent Malware Detection Using CNN and OSINT

Madathil Abdul Majeed, Anjal Muhammed (2025) Intelligent Malware Detection Using CNN and OSINT. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (604kB) | Preview

Abstract

High levels of malware sophistication present an ever-growing risk to global cybersecurity. Signature-based detection methods are proving to be less effective for polymorphic/metamorphic threats; therefore, advanced/intelligent techniques are required. This research project evaluates how effective it is to combine visual malware analysis performed through CNNs with contextual data from Open Source Intelligence (OSINT). The primary methodology consists of converting malware binaries into grayscale images these images are then used as visual fingerprints to train the deep learning model. First, a CNN model was created and used for malware family identification. Later, an innovative model that incorporates OSINT prepared the model by combining visual attributes with simulated metadata, such as the threat level of a malware and the propagation mechanism, into a single model. After this was prepared, both models used the Malimg dataset for training and testing, which contained 9,339 instances across 25 families of malware. The results of the experiment showed that the baseline model has high prediction accuracy (test accuracy) of 95.30%. In contrast, by integrating visual information from the analysed image with contextual OSINT information, the proposed OSINT-integrated model achieved improved accuracy (test accuracy) of 95.51% over the baseline model with significant improvements in classifying under-represented and ambiguous families of malware. This research confirms the hypothesis that improving visual analysis with contextual data from OSINT increases the accuracy and reliability of malware detection methodologies, thereby enhancing future directions for building the augmentation OE cyber threat intelligence platforms.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Mahajan, Kamil UNSPECIFIED
Subjects:	Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence Q Science > QA Mathematics > Computer software > Computer Security T Technology > T Technology (General) > Information Technology > Computer software > Computer Security Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Cyber Security
Depositing User:	Ciara O'Brien
Date Deposited:	03 Sep 2026 11:28
Last Modified:	03 Sep 2026 11:28
URI:	https://norma.ncirl.ie/id/eprint/9803

Actions (login required)

View Item