This repository is the official implementation of Brain Tumor MRI Images Classification with EfficientNet - B7 as the final project of CMP5550 Computer Vision course. In this study the Brain Tumor Dataset dataset has been used.
In this project Python virtual environment has been utilized. To create a virtual environment for the project run the following code:
python -m venv <path-to-virtualenv>
After virtual environment is created, to activate the virtual environment run the following code:
call <path-to-virtualenv>/Scripts/activate.bat
source <path-to-virtualenv>/bin/activate
After activating the virtual environment your terminal should look like below:
(venv) C:\<path-to-project>
(venv) machine-name:path-to-project username$
(venv) username@machine-name:path-to-project$
After activating the virtual environment, to install the requirements run the following code:
pip install -r requirements.txt
To train the model described in the paper, run the following command:
python train.py -d <device-type> -mv <model-version> -e <number-of-epochs> -lr <learning-rate> -op <optimizer> -tb <train-batch-size> -vb <validation-batch-size> -dp <dataset-path> -ci <checkpoint-save-interval> -r <resolution> -nc <number-of-out-classes>
To get more help and see the options run the following command:
python train.py --help
Train creates a directory named as "artifacts" and saves the following artifacts in this directory:
- Train history plot
- Model checkpoints
- Optimizer checkpoints
- Learning rate scheduler checkpoints
- Scaler checkpoints
To evaluate the model run the following command:
python eval.py -dp <dataset-path> -tbs <test-batch-size> -mp <model-checkpoint-path> -mv <model-version> -nc <number-of-classes>
To get more help and see the options run the following command:
python eval.py --help
Eval program saves an ordered dictionary storing the following outputs:
- Misclassified images as numpy arrays
- Predicted labels
- Ground truth labels
- Confusion matrix
Our model achieves the following performance:
| Model | Accuracy |
|---|---|
| EfficientNet B7 | 95.728% |
Versions of the modules in the requirements.txt file may have dependencies or they may require a Python version newer than a specific version (>Python 3.9).
While this study thoroughly examines the effectiveness of deep convolutional neural networks for medical images, its output solely serves research purposes and is not intended for medical decision-making.
We welcome contributions to this project! To ensure a smooth contribution process, please follow these guidelines:
Areas of Contribution:
- Bug fixes and improvements to the codebase
- New features related to the paper or model
- Documentation improvements and tutorials
- Unit tests and code coverage enhancements
- Sharing experimental results and analyses
Contribution Workflow:
- Fork this repository and create a new branch for your changes.
- Implement your changes and update the relevant documentation.
- Run unit tests and ensure your code adheres to the project's style guide.
- Create a pull request and clearly describe your changes.
- Be prepared to address any feedback or suggestions from the maintainers.
Additional Notes:
- Please adhere to the PEP 8 coding style guide.
- Include unit tests for any new code you add.
- Use descriptive commit messages and pull request titles.
- We appreciate contributions in any form, even if they are small bug fixes or suggestions.
Thank you for your interest in contributing!
For citation you can use the following BibTeX
@misc{effnet_brain_tumor_classifier,
author = "Murat Samanci, and Volkan Bakir",
title = "Brain Tumor MRI Images Classification with EfficientNet - B7",
year = "2024",
howpublished = "\url{https://github.com/creaturerigger/efficientnet_brain_tumor_classifier}",
note = "Implements the EfficientNet model on brain tumor MRI images dataset. Includes training and evaluation scripts."
}