Welcome to the CNN-CT-BRAIN Project! 🧠

👩‍🔬 Cancer, Tumor, and Aneurysm Classification and Detection

This repository contains the implementation of VGG16 and ResNet-50 models, techniques such as fine-tuning with layer freezing and unfreezing were applied.

Regularization methods like L1 and L2, dropout, and early stopping were employed to prevent overfitting.

Strategies such as data augmentation, transfer learning and one-hot encoding were integrated, and optimization was carried out using algorithms like Adam.

👥 Development Team (Ctrl + Click to view profiles)

🕵🏻 Dataset Overview

This project focuses on developing a classification model to detect brain diseases, such as cancer, tumors, and aneurysms, from CT brain scans.

Scan of a person with aneurysm, cancer, and tumor Our focus on image processing allows us to provide accurate diagnosis 🩺

📖 CNN-CT-BRAIN Libraries

📊 Dataset Structure

The dataset is organized into two main folders:

1. JPEG Images (.jpg) 📷: Contains CT brain scan images in JPEG format.
2. DICOM Files (.dcm) 📁: Contains the same CT scans but in Digital Imaging and Communications in Medicine (DICOM) format.

Requirements

To run this project, please follow these steps:

1. Download the dataset:

   • Visit the dataset page on Kaggle: CT Scan Images of the Brain and download it.

2. Unzip and overwrite the "archive" file:

   • Unzip the downloaded ZIP file.
   • Locate the file named "archive" 🗃️ and overwrite this project's file.

For Local Execution:

• If running locally, ensure the dataset is in the project directory.

For Google Colab:

• If using Google Colab, upload the dataset to your Google Drive.

• Modify the base_dir in the code to point to the Colab Notebooks📂 directory:

  base_dir = "/content/drive/MyDrive/Colab Notebooks/"

Getting Started

1. Clone this repository:

     git clone https://github.com/AlejandroDavidArzolaSaavedra/CNN-CT-BRAIN.git

2. Explore the dataset folders to understand the structure.

3. Use the data to train and test your models and contribute to advancing medical image analysis.

🤝 Contributions

If you wish to contribute to this project, feel free to do so. You can open issues or send pull requests to improve the code. Your collaboration is welcome! 🚀