Brain Tumor Detection Project

This repository contains the code and resources for a Brain Tumor Detection project using Convolutional Neural Networks (CNN) with PyTorch. The project involves analyzing CT images of the brain to differentiate between healthy and tumor-affected images.

Table of Contents

• Introduction
• Dataset
• Project Structure
• Installation
• Usage
• Model Development
• Training and Evaluation
• Web Application
• Results
• Contributing

Introduction

The Brain Tumor Detection project aims to accurately classify brain CT images as either healthy or tumor-affected. This project leverages deep learning techniques using PyTorch to build and train a CNN model. The trained model is then deployed in a Flask-based web application to provide real-time classification of brain CT images.

Dataset

The dataset used for this project consists of CT images of the brain, sourced from Kaggle. It includes images of both healthy and tumor-affected brains.

Project Structure

brain-tumor-detection/ ├── pycache/ ├── model/ │ └── model.pth ├── static/ │ ├── css/ │ └── uploads/ ├──templates/ │ ├── detect.html │ └── index.html ├── README.md ├── tumor.py ├── app.py ├──Brain Tumor Detection.pptx └──Tumor Detector.ipynb

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/brain-tumor-detection.git
   cd brain-tumor-detection

2. Install the required packages:

    requirements: 
     python
     pyTorch(visual) (checkout from their website for others libraries)
     numpy
     pandas
     sklearn
     matplotlib
     seaborn
     flask
sh
    pip install -r requirements.txt

Usage

1. Preprocess the data:

   cd notebooks
   jupyter Tumor Detector.ipynb.ipynb

2. Train the model:

   jupyter Tumor Detector.ipynb.ipynb

3. Run the web application: flask run

Model Development

The CNN model is built using PyTorch. Key components include:

• Data preprocessing using NumPy and Pandas for resizing and format standardization.
• Custom dataset and dataloader classes to manage image data.
• A CNN model for classifying brain images.

Training and Evaluation

The model is trained over approximately 600 epochs. The training process involves:

• Splitting the dataset into training and validation sets.
• Using data augmentation techniques to improve model generalization.
• Monitoring the model's performance and checking for overfitting.

Web Application

The trained model is deployed in a Flask-based web application. The web app allows users to upload brain CT images and receive real-time classification results, indicating whether the image depicts a healthy brain or one with a tumor.

Results

The model's performance is evaluated based on metrics such as accuracy, precision, recall, and F1-score. Graphical representations of the results are provided using Matplotlib and Seaborn.

Contributing

Contributions are welcome.

Preview