Skip to content

Latest commit

 

History

History
111 lines (78 loc) · 4.46 KB

README.md

File metadata and controls

111 lines (78 loc) · 4.46 KB

Brain-Tumor-MRI-using-VGG16-ResNet50-EfficientNet-CNN-Xception

Welcome to the Brain Tumor MRI Classification project! This repository demonstrates the use of advanced Convolutional Neural Networks (CNNs) to classify brain MRI images into tumor and non-tumor categories. The models utilized include state-of-the-art architectures like EfficientNet, ResNet50, Xception, and VGG16 to achieve high accuracy and performance in image classification.

🎯 Project Overview

The goal of this project is to classify brain MRI scans into tumor or non-tumor categories using pre-trained CNN models. The dataset used contains MRI images of brain tumors, and the models are trained and evaluated to help improve automated diagnostics in medical imaging 🏥.

🔍 Key Highlights:

  • Multiple CNN Models: We explore and fine-tune several popular CNN architectures including:
    • EfficientNet
    • ResNet50
    • Xception
    • VGG16
  • Data Augmentation: We implement powerful data augmentation strategies to increase the variability of training data and prevent overfitting.
  • Evaluation Metrics: The models are evaluated on a wide range of metrics, including:
    • Accuracy
    • Precision
    • Recall
    • F1 Score
    • Confusion Matrix 📊
    • AUC-ROC Curve 📈
    • Support (class distribution)
  • Visualizations: Comprehensive visualizations for confusion matrices, loss/accuracy plots, and sample predictions.

📂 Dataset

The Brain Tumor MRI Dataset used in this project is sourced from Kaggle. It consists of:

  • Training Data: Brain MRI images categorized into tumor and non-tumor.
  • Testing Data: Separate set of MRI images used for testing model performance.

🛠️ Technologies and Tools

The following technologies and libraries are used in this project:

  • Deep Learning Framework: TensorFlow, Keras
  • Data Handling: Pandas, Numpy
  • Image Processing: PIL, OpenCV
  • Visualization: Matplotlib, Plotly
  • Evaluation: Scikit-learn

📦 Python Packages

To install the necessary dependencies, run:

pip install -r requirements.txt

Requirements

  • tensorflow
  • keras
  • scikit-learn
  • matplotlib
  • plotly
  • pandas
  • numpy

🚀 Models and Training

1. EfficientNet

EfficientNet is a family of neural networks that are optimized for performance and efficiency. In this project, EfficientNet is fine-tuned to classify MRI images into tumor/non-tumor categories.

2. ResNet50

ResNet50 is a deep residual network that overcomes the vanishing gradient problem by using skip connections. It allows the model to go deeper with better accuracy.

3. Xception

Xception is a powerful model built on the depthwise separable convolution idea. It's highly efficient for image classification tasks and is fine-tuned for our specific dataset.

4. VGG16

VGG16 is a classic CNN architecture that uses a stack of convolutional layers followed by fully connected layers for classification.

Each model is trained using the Adamax optimizer and evaluated on the test set. The models are compared based on accuracy, precision, recall, and F1 scores.

📊 Evaluation Metrics

The models are evaluated using the following metrics:

  1. Accuracy: Overall accuracy of the model in classifying the images.
  2. Precision: How many of the predicted positive cases were correct.
  3. Recall: How many of the actual positive cases were correctly identified.
  4. F1 Score: A balance between precision and recall.
  5. Confusion Matrix: To visualize the model's performance in a matrix form.
  6. AUC-ROC Curve: Evaluates the true positive rate against the false positive rate across thresholds.

📜 Results

The table below summarizes the performance of each model across different evaluation metrics:

Model Accuracy Precision Recall F1 Score AUC-ROC
EfficientNet 96.3% 0.96 0.95 0.98 0.97
ResNet50 94.8% 0.94 0.93 0.93 0.96
Xception 89.5% 0.95 0.94 0.94 0.96
VGG16 97.3% 0.92 0.91 0.91 0.94

👏 Acknowledgments

Special thanks to the Kaggle community for providing the dataset and to the developers of TensorFlow and Keras for their powerful tools for deep learning.