DataMonitor: Tool for Out-of-Distribution Detection and Data Drift Monitoring

Welcome to the DataMonitor documentation. This tool is designed for:

• Detecting Out-of-Distribution (OOD) inputs in medical imaging datasets.
• Monitoring dataset drift over time for AI/ML models.

DataMonitor is modular and contains three core modules:

1. Feature Extraction: Supervised and unsupervised learning methods for image feature extraction.
2. OOD Detection: Employs similarity and distance metrics to identify OOD inputs.
3. Data Drift Monitoring: Leverages Statistical Process Control (SPC) techniques to flag data drift.

---

Code Repository

The source code for DataMonitor is publicly available on GitHub.
👉 GitHub Repository: DataMonitor

Feel free to explore the code, contribute, or raise issues.

---

Citation

If you use DataMonitor in your work, please cite the following paper:

Zamzmi, Ghada, et al. "Out-of-Distribution Detection and Radiological Data Monitoring Using Statistical Process Control." Journal of Imaging Informatics in Medicine (2024): 1-19.

---

Quick Links

Here are the key sections of the DataMonitor documentation:

• Installation Guide
• Modules Overview
• Data Drift Scenarios
• Limitations
• Contact

---

Data Monitoring Overview