Welcome to the Medical-Projects repository! This collection is dedicated to various biomedical image processing projects, focusing on classification, localization, and segmentation tasks. Our goal is to replicate and build upon state-of-the-art (SOTA) results in these critical domains, advancing the field of medical imaging through deep learning.
Notebook: Classification_Swin_NIH_chestXRay.ipynb
Description:
This project leverages a Swin Transformer-based model for the multi-class and multi-label classification of chest X-rays from the NIH ChestX-ray dataset. The model is designed to detect various thoracic diseases by analyzing X-ray images, pushing the boundaries of diagnostic accuracy.
Notebook: DINO_localization_vindr.ipynb
Description:
This notebook employs the DINO (Self-Distillation with No Labels) approach for image localization tasks on the VinDr dataset. The project aims to enhance localization accuracy by using bounding boxes as references for detecting abnormalities in medical images.
Notebook: TestResults_Localization.ipynb
Description:
This notebook provides detailed test results and evaluation metrics for localization models applied to medical imaging datasets. It includes performance analysis and visualizations, offering insights into model predictions and overall effectiveness.
Notebook: pneumothorax_segmentation.ipynb
Description:
Focused on the segmentation of pneumothorax in chest X-ray images, this project utilizes deep learning to detect and precisely segment affected regions. The model is trained and tested on a dataset specifically curated for pneumothorax detection, contributing to more accurate and timely diagnoses.
Notebook: MedMNIST_2D_3D.ipynb
Description:
This notebook explores both 2D and 3D medical image classification using the MedMNIST dataset. The project replicates and builds upon SOTA techniques for handling diverse datasets, applying deep learning methods to advance the classification of medical images across different dimensions.
To explore these projects, clone the repository and start with the notebook that aligns with your area of interest:
git clone https://github.com/SID-6921/Medical-Projects.git
cd Medical-Projects📈 Results & Contributions Visualizations: Each project includes detailed visualizations of the results, helping you understand the model's predictions and performance. Metrics: Comprehensive evaluation metrics are provided to assess model effectiveness in real-world scenarios. Contributions: Contributions are welcome! Feel free to submit pull requests or open issues for improvements, bug fixes, or discussions. 📄 License This repository is licensed under the MIT License. See the LICENSE file for more details.
Happy coding! If you find these projects helpful or have any suggestions, please consider giving the repository a ⭐ and contributing to the ongoing efforts to enhance medical imaging through deep learning.