MRI Brain Tumor Detection Project

Overview

This project implements a brain tumor detection system using Gabor filters and a 3D U-Net architecture. The model is trained on the BraTS 2020 dataset, which contains multimodal MRI images (FLAIR, T1, T1CE, T2) and corresponding segmentation masks. The goal is to accurately segment brain tumors from MRI scans.

Table of Contents

1. Requirements
2. Dataset
3. Installation
4. Usage
   • Gabor Filter Visualization
   • Model Training
   • Model Evaluation
5. License

Requirements

To run this project, you will need the following Python packages:

• numpy
• opencv-python
• matplotlib
• nibabel
• tensorflow
• keras
• sklearn
• segmentation-models-3D

You can install the required packages using pip:

pip install numpy opencv-python matplotlib nibabel tensorflow keras scikit-learn segmentation-models-3D

Dataset

This project uses the BraTS 2020 dataset, which can be downloaded from the BraTS Challenge website. The dataset contains the following modalities:

• FLAIR
• T1
• T1CE
• T2
• Segmentation masks

Make sure to place the dataset in the appropriate directory as specified in the code.

Installation

1. Clone this repository or download the notebook file.
2. Ensure that the BraTS dataset is downloaded and placed in the specified directory.
3. Open the notebook in Google Colab or Jupyter Notebook.

Usage

Load and Preprocess Data

The code loads the MRI images and corresponding masks, normalizes the pixel values, and prepares the data for training.

Gabor Filter Application

The code applies Gabor filters to the MRI images to enhance feature extraction.

Model Definition

A 3D U-Net model is defined for semantic segmentation of brain tumors.

Training

The model is trained using the training dataset, and the training history is recorded.

Evaluation

The model's performance is evaluated using metrics such as Intersection over Union (IoU) and accuracy.

Gabor Filter Visualization

The project includes a section that visualizes Gabor filters with varying parameters (theta and gamma). This helps in understanding how different filter configurations affect the feature extraction process.

Example code to visualize Gabor filters:

# Example of creating Gabor filters
ksize = 50
sigma = 3
thetas = [0, np.pi/4, np.pi/2, 3*np.pi/4]
gammas = [1, 0.5, 0.01]

# Code to visualize Gabor filters

Model Training

The model is trained using the following parameters:

• Batch size: 2
• Learning rate: 0.0001
• Number of epochs: 100

The training process includes data augmentation and uses a custom loss function that combines Dice loss and focal loss to handle class imbalance.

Example code for model training:

# Example of model training
history = model.fit(train_img_datagen,
                    steps_per_epoch=steps_per_epoch,
                    epochs=100,
                    validation_data=val_img_datagen,
                    validation_steps=val_steps_per_epoch)

Model Evaluation

After training, the model's performance is evaluated on a validation dataset. The code calculates the Mean IoU and visualizes the predictions against the ground truth masks.

Example code for model evaluation:

# Example of model evaluation
IOU_keras = MeanIoU(num_classes=n_classes)
IOU_keras.update_state(test_pred_batch_argmax, test_mask_batch_argmax)
print("Mean IoU =", IOU_keras.result().numpy())

License

This project is licensed under the MIT License. See the LICENSE file for more details.