Enigma Cerebrum is an innovative ensemble model designed for accurate brain tumor detection using artificial intelligence. This repository houses a powerful fusion of pre-trained models, combining the latest advancements in deep learning for robust medical image analysis.
- Ensemble of Pre-trained Models and Custom CNN: Combine multiple pre-trained models with a custom CNN for enhanced accuracy.
- AI-driven Brain Tumor Classification: Utilize artificial intelligence to detect and classify brain tumors.
- Comprehensive Medical Image Analysis: Leverage deep learning for detailed and comprehensive medical image analysis.
- Cerebral Fusion for Enhanced Accuracy: Fuse the strengths of individual models for improved overall performance.
- Easy Integration of New Models: Easily integrate new pre-trained models and custom CNN for continuous improvement.
Download the Dataset: - Brain Tumor MRI Dataset
Download the Pre-Trained Model Weights: - Keras Files
Dataset Information: - Description: This dataset contains 7023 images of human brain MRI images classified into 4 classes: glioma, meningioma, no tumor, and pituitary. - Source: A combination of figshare, SARTAJ dataset, and Br35H. - Note: Pay attention to the varying sizes of images. Resize images to the desired size after preprocessing for improved model accuracy.
Add Pre-trained Models:
- Place your pre-trained Keras models (.keras
files) in the /models
directory.
Run the Ensemble Model Training Script
- Includes an ensemble of all available pre-trained models and a custom CNN.
- Achieves an accuracy of 91% on the provided dataset.
- Includes a subset of pre-trained models and a custom CNN.
- Achieves an accuracy of 91% on the provided dataset.
A brain tumor is a collection, or mass, of abnormal cells in your brain. It can be cancerous (malignant) or noncancerous (benign), causing an increase in pressure inside the skull and potential brain damage.
Early detection and classification of brain tumors are crucial for selecting the most suitable treatment method, ultimately saving lives.
This project employs deep learning approaches for brain tumor diagnosis, including detection, location identification, and classification based on malignancy, grade, and type.
This project is licensed under the MIT License.