DL-SCDisassembly

🪪 About | 🪄 Installation | 🗃️ Usage | 🏷️ Features | 🔗 Citation | 📝 Paper

This repository contains the code for the paper: Enhancing deep learning-based side-channel analysis using feature engineering in a fully simulated IoT system

About

The main objective is to create a refined dataset for side channel-based disassembly tasks using DL models. The dataset is enhanced using statistical and temporal log transformation features.

Installation

# Clone the repository
git clone https://github.com/PLASS-Lab/DL-SCDisassembly.git

# Navigate to the project directory
cd DL-SCDisassembly

# Install dependencies
python setup.py install

Usage

After uploading the assembly logs, collected from GDB, to the selected folder, it can be converted to a structured CSV file using assembleyTocsv.py

python assembleyTocsv.py

The generated CSV file can be passed to python-elmo engine for power simulation.

Then, the FeatureEngineering.py file can be used to generate the features using a user-specified window size.

python FeatureEngineering.py

The refined dataset can then be used for the detection of countermeasures or disassembly using predefined feature sets and window sizes

python inference.py
python classification

Features

The main features.

• Assembly to CSV Conversion: Convert raw assembly logs from GDB into structured CSV files, making them easier to manipulate and analyze.
• Feature Engineering for Side-Channel Analysis: Use FeatureEngineering.py to apply statistical and temporal log transformation features on the dataset, allowing for customizable feature sets and window sizes.
• Countermeasure Detection and Disassembly Analysis: Perform classification and disassembly tasks with the refined dataset. Experiment with different window sizes and feature sets.

Citation

If you use this code for your research, please cite the following paper.

Title: Enhancing deep learning-based side-channel analysis using feature engineering in a fully simulated IoT system
Journal: Expert Systems with Applications
DOI: 10.1016/j.eswa.2024.126079

@article{alabdulwahab2025dlscd,
  title = {Enhancing deep learning-based side-channel analysis using feature engineering in a fully simulated IoT system},
  journal = {Expert Systems with Applications},
  volume = {266},
  pages = {126079},
  year = {2025},
  issn = {0957-4174},
  doi = {https://doi.org/10.1016/j.eswa.2024.126079},
  url = {https://www.sciencedirect.com/science/article/pii/S0957417424029464},
  author = {Alabdulwahab, Saleh and Cheong, Muyoung and Seo, Aria and Kim, Young-Tak and Son, Yunsik},
  keywords = {Side-channel attacks, Feature engineering, Hiding countermeasures, Disassembly attacks, Deep learning, Reverse engineering},
}