Deep learning for medical segmentation and classification
Spring 2023, Cogito NTNU.
In this project we use convolutional neural networks (CNNs) for computer vision. Using the U-Net architecture (Olaf Ronneberger, Philipp Fischer, Thomas Brox, 2015) as a basis, we tackle both classification-problems and segmentation-problems in the medical field.
Using the U-Net arcitechure we perform binary classification of
Dataset: Labeled Chest X-ray .
Segmentation of lungs
Dataset: LGG_MRI segmentation .
We performed lung segmentation achieving a F1 score of 0.957 and a IoU/jaccard score of 0.918.
Segmentation of Pneumothorax
Dataset: https://www.kaggle.com/datasets/vbookshelf/pneumothorax-chest-xray-images-and-masks
As you use Doctor-AI for your own use, please cite the authors of the package:
@misc{doctorai2023,
title={Doctor AI - U-Net for medical segmentation and classification},
author={Vilhjalmurson, Vilhjalmur and Myhre, Sveinung and Bohne, Erik and Constantinos, Joel and Zhao, Ine},
year={2023},
publisher={Cogito NTNU}
}
This project is licensed under the MIT License. See the LICENSE file for more information.