Brain Tumor Classification using InceptionResNetV2 🧠🔍

Overview

This project focuses on the classification of brain tumor images using the InceptionResNetV2 deep learning architecture. The goal is to develop an accurate model capable of distinguishing between tumor and non-tumor images.

Project Highlights 🚀

• Dataset Description: Brief overview of the dataset, including class distribution.
• Data Generators: Efficient loading and augmentation of data during training, testing, and validation.
• Class Labels: 'Tumor': 0, 'non-Tumor': 1.

Dataset Statistics 📊

• Training Dataset:
  • Number of datasets : 98

GPU and Model Checkpoint 🖥️💾

• Check GPU Availability: Ensure efficient training with GPU acceleration.
• Model Checkpoint: Save the model during training for future use.

Model Architecture 🧠🏗️

• InceptionResNet for Feature Extractor: Utilized the power of InceptionResNetV2 for automatic feature extraction.

Model Training 🚂💨

• Training Starts Here: Begin the training process.
• Save Model History: Save training history in a CSV file for analysis.
• Accuracy and Loss Graphs: Visualize the training and validation performance.

Model Evaluation 🧐📈

• Testing Starts Here: Evaluate the trained model on the test set.
• Evaluation Metrics: Compute accuracy, precision, recall, and F1-score.
• Classification Report: Detailed report of model performance.
• Confusion Matrix: Visualize model predictions.
• ROC Curve: Evaluate model performance.

Model Predictions 📸🔮

• Random Image Predictions: Check the model's predictions on random images from the test dataset.