Skip to content

inesdv26/Bone-Tumor-Classification-and-Explainability

Repository files navigation

Uncertainty in Bone Tumor Classification: Enchondroma vs. Atypical Cartilaginous Tumor

Overview

This repository contains the official manuscript and a Jupyter Notebook for my master's thesis, focused on distinguishing Enchondroma and Atypical Cartilaginous Tumor (ACT) using conventional radiographs.

Contents

  • manuscript.pdf: The complete written thesis, including research background, methodology, results, and conclusions.
  • Uncertainty-in-Bone-Tumor-Classification.ipynb: A comprehensive Jupyter Notebook demonstrating data processing, model development, inference, and explainability techniques.

Objective

The goal of this project is to leverage a Deep Learning model to classify two types of bone tumor lesions that appear similar on X-ray images. The study uses Explainable AI (XAI) techniques, specifically SHapley Additive exPlanations (SHAP), to improve model interpretability and identify the most critical features contributing to predictions.

How to Use

  1. Read through the manuscript.pdf to get a comprehensive understanding of the research.
  2. Explore the notebook for detailed code, analysis, and visualizations.

Requirements

Install dependencies by running:

pip install -r requirements.txt

Contact

For questions or further discussions, feel free to reach out at: [[email protected]].

License

This project is licensed under the MIT License.

Acknowledgments

This thesis was completed with the support of the Lab for AI in Medicine at the Technical University of Munich and the Department of Orthopaedics and Sports Orthopaedics at the Klinikum rechts der Isar.

Data Privacy and Usage

The dataset used in this project is private and belongs to the hospital. It cannot be shared publicly due to confidentiality agreements and patient privacy regulations. Any results or code shared in this repository are based on this dataset but do not contain any identifiable or sensitive patient information.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published