DomainATM: Domain Adaptation Toolbox for Medical Data Analysis

License: Apache License - 2.0

Language: MATLAB

More details can be found in the file "manual.pdf". Note, for Mac, please use the DomainATM_Mac.zip

---

Overview

DomainATM is a MATLAB-based toolbox that provides an intuitive GUI and ready-to-use implementations for feature-level and image-level domain adaptation methods, specifically tailored for medical data analysis. The toolbox supports both real-world and synthetic datasets and enables easy evaluation, visualization, and extension with user-defined algorithms.

---

Installation

Requirements

• MATLAB (2020 or later versions are recommended)
• Image Processing Toolbox
• Statistics & Machine Learning Toolbox

Installation Steps

1. Download the DomainATM.mlappinstall file.
2. On Windows/macOS: Double-click the .mlappinstall file.
   On Linux: Open MATLAB → Apps tab → Click Install App.
3. Set the MATLAB current directory to the folder containing the toolbox.
   Example: E:/DomainATM

---

Folder Structure

• data/: Stores real or synthetic datasets for fast verification.
• algorithms_feat/: Feature-level adaptation methods.
• algorithms_img/: Image-level adaptation methods.
• evaluation/: Stores the output of each experiment.
• tools/: Utility functions used by the GUI.

---

Feature-Level Domain Adaptation

Supported Methods

• Subspace Alignment (SA)
• Correlation Alignment (CORAL)
• Transfer Component Analysis (TCA)
• Optimal Transport (OT)
• Joint Distribution Adaptation (JDA)
• Transfer Joint Matching (TJM)
• Geodesic Flow Kernel (GFK)
• Scatter Component Analysis (SCA)
• Information-Theoretic Learning (ITL)

Usage

1. Click Create Dataset to generate synthetic data or place your .mat dataset in data/.
2. Click Load Data and select your dataset.
3. Click Feature-Level Adaptation and choose an algorithm to apply.
4. Click Feature-Level Metrics to evaluate the results (accuracy, visualization, etc.).

---

Image-Level Domain Adaptation

Supported Methods

• Histogram Matching (HM)
• Spectrum Swapping-based Image-Level Harmonization (SSIMH)

Usage

1. Click Image-Level Adaptation.
2. Select source and target .nii images.
3. Choose an algorithm and click Run.
4. Use Image-Level Metrics to evaluate harmonization quality.

---

Custom Algorithms

To add your own domain adaptation method:

• For feature-level: Place your .m script in algorithms_feat/ using:

  X_adapted = FeatureDA(X, domain_label, Y, param);

• For image-level: Place your .m script in algorithms_img/ using:

  S_adapted = ImageDA(source, target, param);

Use "Add Your Algorithm" in the GUI and click "Refresh".

---

Citation

If you find this toolbox useful in your research, please cite:

@article{guan2023domainatm,
  title={DomainATM: Domain adaptation toolbox for medical data analysis},
  author={Guan, Hao and Liu, Mingxia},
  journal={NeuroImage},
  volume={268},
  pages={119863},
  year={2023},
  publisher={Elsevier}
}