-
-
this is our patient test history interface
-
-
Inspiration
Cardiovascular diseases, particularly Coronary Artery Disease (CAD), are the leading cause of death worldwide. However, preliminary evaluations leave critical warning signs buried inside complex lab reports, ECGs, and echocardiograms. Traditional AI tools often act as non-intuitive "black boxes," offering a raw risk percentage without explaining where the blockage is located or why the risk is elevated. We created Heart Check 3D to bring visual clarity, AI transparency, and interactive 3D modeling to preliminary cardiac screening.
What it does
Predicts LAD, LCX, and RCA stenosis using SHAP Explainable AI and interactive 3D heart visualization.
How we built it
Built using Flask, Chart.js, and WebGL for real-time 3D cross-sectional clipping.
Challenges we ran into
Solved cloud GPU crashes by migrating from server-side VTK to client-side WebGL.
Accomplishments that we're proud of
Achieved a 0.942 ROC-AUC score while making AI predictions fully transparent.
What we learned
Transparent XAI is essential for gaining clinical trust.
What's next for Track A
Integrating custom GLB models and automated PDF medical reports.
