NeuroVision AI

Dashboard
Stroke Risk Prediction
Analytics
Accessibility Center
Medical Report

Inspiration

Every minute of stroke delay costs 1.9 million neurons, yet millions lack access to radiologists. We wanted to democratize stroke detection through accessible AI that anyone can use—regardless of visual ability, location, or technical expertise.

What it does

NeuroVision AI analyzes brain CT/MRI scans in seconds, providing instant stroke risk assessment (Low/Medium/High) with visual heatmaps explaining exactly where the AI is looking, plus generates clinical reports and personalized recommendations.

How we built it

We trained a 4-block CNN with TensorFlow/Keras on brain imaging data, integrated Grad-CAM for explainability, and built a fully accessible Streamlit interface with high-contrast, large-text, and colorblind-friendly modes.

Challenges we ran into

Making complex AI explainability (Grad-CAM) visually intuitive while maintaining accessibility standards, and ensuring the interface remained fully keyboard-navigable while preserving a modern, professional medical aesthetic.

Accomplishments that we're proud of

Achieving a built-in accessibility score of 70-100/100 with six distinct accessibility features, while building an end-to-end clinical workflow—from scan upload to PDF report generation—in a single, beautiful interface.

What we learned

AI transparency through visual explainability transforms trust in medical AI, and accessibility isn't an add-on—it's fundamental to making technology truly serve everyone, especially in healthcare.

What's next for NeuroVision AI

Speech-to-text integration for voice-driven navigation, multilingual support, EHR system integration via FHIR, and multi-modal analysis combining CT, MRI, and patient vitals for even more accurate predictions.