This repository contains the code and resources for a comparative study of Convolutional Neural Networks (CNNs), Residual Networks (ResNet), and Vision Transformers for multi-classification in X-ray images.
You can also refer to our final paper for detailed results and discussion: Paper
Early and accurate detection of diseases is crucial for improving patient outcomes. This project focuses on utilizing deep learning techniques to classify chest X-ray images into different classes of cancerous cells. We compare the performance of CNNs, ResNet, and Vision Transformers to identify the most effective architecture for this task.
We use the NIH Chest X-ray dataset, which comprises 112,120 X-ray images with disease labels from 30,805 unique patients. The dataset is publicly available and includes labels relevant to cancer diagnosis.
- Navigate to
data_download.ipynb
and download the data. Ensure that it is downloaded intoinput
folder. - Delete downloaded zip if needed
- Change the path in models.
image_dir
root_dir
andimage_path
, otherwise where applicable.
To run it on smaller dataset. Navigate to:
https://www.kaggle.com/datasets/nih-chest-xrays/sample
- Download the files
- Change the path in models.
image_dir
root_dir
andimage_path
, otherwise where applicable.
To run the code, follow these steps:
- Clone this repository:
git clone https://github.com/your-username/CSC413-Final-Project.git
- Install the required dependencies:
pip install -r requirements.txt
- Kaushik Murali
- Isha Surani
- Aviral Bhardwaj
- Ananya Jain
Include any relevant references, papers, or resources here.
This project is licensed under the MIT License. See the LICENSE file for details.