🧠 MCI-to-AD MRI Classification

This project uses EfficientNetB0 Transfer Learning to classify MRI brain scans into 4 stages of cognitive decline:

• EMCI (Early Mild Cognitive Impairment)
• LMCI (Late Mild Cognitive Impairment)
• MCI (Mild Cognitive Impairment)
• AD (Alzheimer's Disease)

📂 Dataset

• Custom MRI dataset with 4 labeled classes.
• Split into training & validation using ImageDataGenerator.

🏗 Model

• Base model: EfficientNetB0 (ImageNet weights)
• GlobalAveragePooling + Dropout + Dense(4, softmax)

📊 Training

• Optimizer: Adam
• Loss: Categorical Crossentropy
• Validation split: 20%

🚀 Results

• Shows per-class accuracy & confusion matrix
• Predicts patient’s stage of cognitive impairment

📦 How to run

pip install -r requirements.txt
python train.py


## 🔮 Future Work
-Add Grad-CAM visualization to see MRI regions that influence the model.

-Experiment with deeper models like EfficientNetB3.