CardioVision 3D

the home page of the webiste
the heart model with the prediction input
explaining the heart model
CAD profile heart
another part of the heart
medicine and the symptoms
medicine
explaning the heart part
with the explaing of that particular part
another part with some explanation

Inspiration Cardiovascular disease remains the leading cause of mortality worldwide, claiming nearly 18 million lives each year. Yet, inside clinical consult rooms, a silent crisis of communication exists:

The Abstract Risk Dilemma: Standard cardiovascular calculators (such as Framingham or ASCVD) output a single percentage (e.g., "24% 10-year risk"). This abstract number fails to inform cardiologists which specific coronary artery is vulnerable, and fails to convince patients why they must make urgent lifestyle changes. The "Black-Box" ML Problem: While modern Machine Learning models can detect Coronary Artery Disease (CAD) with high accuracy, they operate as uninterpretable black boxes. In high-stakes medicine, clinicians cannot trust a prediction they cannot explain, and patients cannot follow advice they cannot understand. We asked ourselves: What if an AI could not only predict coronary artery disease, but project that risk directly onto a real-time, interactive 3D model of the patient’s own heart, pinpoint the exact compromised vessel, and explain why in plain English?

That question inspired CardioVision 3D.

🫀 What It Does CardioVision 3D is a multi-vessel CAD risk assessment and explainability platform that unites predictive machine learning, game-engine-grade 3D graphics, and game-theoretic explainable AI (XAI).

Multi-Artery Stenosis Prediction: Instead of just predicting generic CAD, our pipeline predicts stenosis risk individually for each major coronary artery: LAD (Left Anterior Descending): The critical "Widow Maker" artery supplying the anterior wall and apex. LCX (Left Circumflex Artery): Supplying the posterolateral left ventricle. RCA (Right Coronary Artery): Supplying the inferior myocardium and conduction nodes. Interactive 3D Heart Perfusion Mapping: Built with React Three Fiber, the 3D heart dynamically colors each coronary vessel and corresponding myocardial territory based on predicted risk level (Green = Low, Amber = Moderate, Crimson = High). Premature Ventricular Contraction (PVC) Origin Localization: Interactive 3D hotspots allow clinicians to pinpoint ectopic arrhythmia foci across regions like the RVOT (Right Ventricular Outflow Tract) and LVOT. Plain-English Clinical Breakdown: Every prediction is decoded into three fundamental questions: 📍 Where is the problem? (Precise anatomical region & affected artery) ⚠️ What is the problem? (Calibrated probability & physiological hazard) 🔍 Why is this happening? (Top patient-specific clinical drivers)