Alzheimer’s MRI Detection

Inspiration

I wanted to explore how AI can support early detection of Alzheimer’s from MRI scans. Subtle brain changes are hard to notice manually, so I built a model that can classify MRI images automatically.

What it does

• How to prepare medical imaging datasets
• How CNNs learn features from MRI scans
• Preprocessing, augmentation, and evaluation
• Working with limited hardware and avoiding overfitting

How i built it

1.  Cleaned and preprocessed MRI images
2.  Trained a CNN classifier
3.  Evaluated using accuracy, precision, recall
4.  Built an inference pipeline for predictions

Challenges i ran into

• Large data → slow training
• Cleaning and balancing the MRI dataset
• Avoiding overfitting

Accomplishments that i'm proud of

• Built a working MRI-based Alzheimer’s classifier with high accuracy
• Successfully cleaned and organized a messy medical dataset
• Implemented a full training → evaluation → prediction pipeline
• Managed everything within hardware limits and time pressure

What i learned

• How to preprocess and standardize medical imaging data
• How CNNs detect patterns in MRI scans
• Importance of proper augmentation and balanced datasets
• How to evaluate models using accuracy, precision, and recall

What's next for Alzheimer’s MRI Detection

• Train on larger, multi-center MRI datasets
• Deploy as a simple web app for demo usage
• Extend model to multi-class Alzheimer’s stages

Built With

• google-colab
• matplotlib
• numpy
• opencv
• python
• tensorflow/pytorch