Brain Tumor Classification using Deep Learning

Table of Contents

• Introduction
• Data Preparation
  • Data Source
  • Data Preprocessing
• Data Exploration
  • Data Visualization
  • Class Distribution
  • Image Dimensions
• Model Architectures
  • Model 1: Linear Layer
  • Model 2: Convolutional Neural Network (CNN)
  • Model 3: Custom ResNet-18
  • Model 4: Pretrained ResNet-34
• Model Training
  • Hyperparameters
• Model Evaluation
  • Evaluation Metrics
  • Confusion Matrices
• Conclusion and Results
• Future Directions

Introduction

Welcome to the Brain Tumor Classification project! This project focuses on using deep learning to classify Brain Tumor MRI images into four categories: Glioma Tumor, Meningioma Tumor, No Tumor, and Pituitary Tumor. The primary objective is to develop models that can assist medical professionals in diagnosing brain tumors accurately and efficiently.

Data Preparation

Data Source

• The dataset used in this project was sourced from Kaggle using the opendatasets library. It consists of MRI images of patients with and without brain tumors.
• The dataset was split into a training set and a testing set, each containing images of the four tumor types.

Data Preprocessing

• Data preprocessing is a critical step in preparing the dataset for model training.
• The following preprocessing steps were applied:
  • Image Resizing: Images were resized to a common size to ensure uniformity.
  • Center Cropping: Center cropping was performed to focus on the central part of the image.
  • Data Augmentation (Training Set): Data augmentation techniques, including random horizontal flips and random rotations, were applied to the training set to increase data diversity.

Data Exploration

Data Visualization

• Visualization of the data is crucial for understanding the nature of the images and the challenges associated with brain tumor classification.
• Sample images from the training dataset were visualized to gain insights into the data.

Class Distribution

• Analyzing the class distribution in both the training and testing datasets is important. This distribution can significantly impact the performance of machine learning models.
• The distribution of the four tumor types was visualized to understand class balance.

Image Dimensions

• Examining the dimensions of the images in the dataset is essential to ensure uniformity in size. This information is crucial for setting the input size for model training.

Model Architectures

In this project, multiple model architectures were explored and implemented for brain tumor classification. Here are the key models used:

Model 1: Linear Layer

• A simple model with fully connected linear layers.
• The architecture included flattening the input image, two linear layers with ReLU activation, and a final linear layer for classification.

Model 2: Convolutional Neural Network (CNN)

• A convolutional neural network (CNN) architecture was used, consisting of convolutional layers with max-pooling, followed by fully connected layers.
• CNNs are known for their ability to capture spatial features in images.

Model 3: Custom ResNet-18

• A custom ResNet-18 architecture was constructed from scratch. This architecture included residual blocks to facilitate the training of deep networks and improve accuracy.

Model 4: Pretrained ResNet-34

• A pre-trained ResNet-34 model from the torchvision library was fine-tuned for brain tumor classification.
• Transfer learning was applied to leverage knowledge learned from a large dataset for improved classification performance.

Model Training

• Model training is a critical phase in developing effective deep learning models. In this project, the models were trained using the one-cycle learning rate policy.
• The one-cycle policy involves gradually increasing the learning rate and then annealing it, resulting in improved training and convergence.

Hyperparameters

• Hyperparameters such as learning rate, weight decay, and gradient clipping were fine-tuned for optimal model performance.
• The models were trained for a fixed number of epochs (12 epochs) to ensure convergence.

Model Evaluation

• Model evaluation is essential to assess the performance of each model. The evaluation phase included the following steps:

Evaluation Metrics

• Evaluation metrics such as accuracy, precision, recall, and F1-score were calculated to measure the effectiveness of the models.
• These metrics provide insights into the model's ability to correctly classify brain tumor images.

Confusion Matrices

• Confusion matrices were plotted to visualize the model's performance in classifying each tumor type.
• These matrices help identify any class-specific issues in classification.

Conclusion and Results

• After extensive model training and evaluation, it was determined that Model 4, which utilized a pre-trained ResNet-34 architecture, achieved the highest accuracy and outperformed the other models.
• This project demonstrates the effectiveness of deep learning in medical image classification, specifically in the context of brain tumor diagnosis.
• The best-performing model can be saved and used for real-world tumor classification tasks, potentially assisting medical professionals in diagnosing brain tumors more accurately and efficiently.

Future Directions

• The project's success opens the door to further improvements and future work. Here are some possible directions:

Advanced Architectures and Techniques

• Explore more advanced architectures and techniques, including ensembling multiple models to improve classification accuracy.

Dataset Expansion

• Expand the dataset by collecting more brain tumor MRI images. A larger dataset can lead to more robust models.

Deployment

• Develop a user-friendly interface for medical professionals to use the model in a clinical setting for real-time brain tumor diagnosis.

This project is a significant step forward in the field of medical image analysis and has the potential to make a meaningful impact on the early and accurate diagnosis of brain tumors.