Graduation Project Report: Malware Classification and Detection Using DenseNet

Introduction

This graduation project report aims to classify and detect malware using the DenseNet model. The project utilizes two main malware datasets: MalImg and Big2015, and benign data sourced from various trustworthy origins.

About the Author

Ngô Phan Tâm Bảo
CSE Student, VNUK
Email: [email protected]
LinkedIn: Ngô Phan Tâm Bảo

Data

MalImg

• Description: The MalImg dataset contains over 9,000 malware images categorized into 25 different classes.
• Download Link: MalImg Dataset

Big2015

• Description: The Big2015 dataset includes various types of malware represented as hex dump files, which are then converted into images.
• Download Link: Big2015 Dataset
• Download Link: Big2015 Image Dataset

Benign

• Source: Dike Dataset
• Download Link: Benign Dataset

Dataset for training detecting model:

• Download Link: MalImg Benign Dataset

Setting Up Streamlit

To install Streamlit and run the project's application, follow these steps:

1. Install Streamlit:

   pip install streamlit

2. Run the Streamlit application:

   streamlit run Homepage.py

Models

The DenseNet models have been trained and stored on Google Drive. You can download the models using the following links:

• Model in Here: Download Model

Conclusion

The project has achieved promising results with the DenseNet model, enhancing the efficiency of malware classification and detection. Thank you for your interest, and best wishes for success in your projects.