This application leverages a PyTorch-based EfficientNet model to classify MRI brain scans into one of the following categories:

• Glioma Tumor
• No Tumor
• Meningioma Tumor
• Pituitary Tumor

Simply upload an MRI scan image, and the model will analyze it to predict the tumor type.

• User-friendly interface: Upload MRI scans in commonly used formats (JPG, PNG, JPEG).
• Efficient and accurate classification using a pre-trained EfficientNet model.
• Real-time predictions for quick analysis.

1. Choose an MRI scan image

   • Supported formats: JPG, PNG, JPEG
   • File size limit: 200MB

2. Upload the image

   • Drag and drop your file into the app.

3. View Results

   • The app will display the uploaded image along with the raw model predictions and the tumor classification result.

Example output from the model:

[19.53183937072754, -53.19132995605469, -115.8001480102539, -4.080233097076416]

Tumor Type

The model assigns scores to each category, and the category with the highest score is selected as the predicted tumor type.

The application is deployed and accessible via Streamlit. You can try it out here:

🔗 MRI Tumor Classification Deployment

• Deprecation Notice: The use_column_width parameter for image display is deprecated. Future updates will replace it with the use_container_width parameter.
• For best results, ensure the uploaded MRI scan is of high quality and correctly formatted.

To run the application locally, you'll need:

• Python 3.8+
• PyTorch
• Streamlit
• EfficientNet library

Install dependencies using the following command:

pip install -r requirements.txt

Special thanks to the developers and contributors who made this project possible by providing tools and frameworks for deep learning and web deployment.

This project is licensed under the MIT License. See the LICENSE file for details.