This repository contains the implementation of CerebroVision, a project focused on using convolutional neural networks (CNNs) for the early detection and classification of brain tumors from MRI images. Leveraging transfer learning with pre-trained models such as ResNet50 and VGG19, the project achieves high accuracy and ROC AUC scores, demonstrating significant improvement over traditional diagnostic methods.
- Data Preprocessing: Scripts for preprocessing MRI images, including normalization and augmentation.
- Model Training: Implementation of CNN architectures with transfer learning using ResNet50, VGG19, and custom CNN models.
- Performance Evaluation: Tools for evaluating model performance, including accuracy, ROC AUC, and other relevant metrics.
- Visualization: Visualization of training results and model predictions using tools like Matplotlib and Seaborn.
- Tracking: Experiment tracking and logging using Weights & Biases (wandb).
- Python
- TensorFlow
- Keras
- Scikit-learn
- Matplotlib
- Seaborn
- Weights & Biases (wandb)