AlzAware

Inspiration

Witnessing how Alzheimer’s silently impacts my father & family inspired me to explore how AI could help detect the disease earlier. I believed early intervention could still make a difference. My motivation came from the idea that early awareness can save memories, and even lives.

What it does

How we built it

1/ Collected and cleaned open-source Alzheimer’s datasets (e.g., ADNI).
2/ Extracted key biomarkers using CNNs for MRI scans and ML models for tabular data.
3/ Combined the outputs using a weighted ensemble:
        "Final Score"=α⋅P_"MRI" +β⋅P_"Cognitive" 
4/ Deployed a simple streamlit dashboard to visualize prediction confidence and progression risk.

Challenges we ran into

Working with limited labelled medical data and ensuring ethical model use were major challenges. I also faced issues with data imbalance, model generalization, and computational limits for 3D MRI processing.

Accomplishments that we're proud of

1/ Built a working prototype that detects early Alzheimer’s risk with decent accuracy. 2/ Learned to handle medical datasets responsibly and apply AI for social good. 3/ Turned a personal motivation into a functional, data-driven healthcare solution.

What we learned

I deep-dived into neuroscience datasets, MRI imaging, and cognitive-test features, learning how subtle changes in brain structure and activity can indicate early decline. I also strengthened my skills in Python, TensorFlow, and data pre-processing for real-world healthcare data.

What's next for AlzAware

1/ Expand training with larger clinical datasets for improved reliability. 2/ Collaborate with healthcare experts for validation and review. 3/ Develop a mobile-friendly screening tool for larger user base community.

Built With

• python
• streamlit