NeuroVision - A Brain Tumour Identification & Classification Model

Hosted and deployed on NeuroVision

This project aims to identify and classify brain tumours from MRI images using deep learning models. The application includes both binary classification (to determine the presence of a tumour) and multi-class classification (to categorise the type of tumour: Glioma, Meningioma, or Pituitary).

Getting started

1. Clone the Repository

git clone https://github.com/Anshxy/Brain-Tumour-Identification-Classification.git
cd Brain-Tumour-Identification-Classification

2. Install Dependencies

pip install -r requirements.txt

Training your own model

The training data used in the pretrained models:

• Multi-class classification
• Binary Classification

You can also train your own custom model by changing the training data pathing in the training files.

Binary Classification

# /trainers/Identify/training.ipynb

# Change training and validation path
train_dir = '../Identify/Data/training'
validation_dir = '../Identify/Data/validation' 

Multi-class Classification

# /trainers/Classification/training.ipynb

# Change training and validation path
train_dir = '../Classification/Data/training'
val_dir = f'../Classification/Data/validation'

Once done training, ensure the models are under the 'Models/' directory

Testing and usage

1. Model preperation Place your pre-trained models (BTBinaryClassification.pth and BTMulticlassClassification.pth) in the Models/ directory.

2. Start the Flask Server

python main.py

3. Open live server Navigate to http://127.0.0.1:5000/ to access the web interface.

4. Upload and Predict Use the web interface to upload MRI images and get predictions!

• Note that this is a demo interface used for purely testing

Author

By Ansh Rawat and Jun Oh