AI for Congenital Heart Defects in Virtual Reality

Discipline: CS

Advisor: Matthew Gerber

We are developing an AI-assisted, immersive visualization system to solve the difficulty clinicians and students face in accurately understanding complex, personalized anatomy from static 2D CT scans, particularly for Congenital Heart Defects (CHDs). The system works by using running the scan through our AI classifier to diagnose and categorize common CHDs, and using a deep learning pipeline to automatically segment and isolate organs, reconstruct them into 3D digital twins, and place them inside a Virtual Reality (VR) environment. We then allow users to analyze these models by dynamically toggling visibility and controlling opacity of individual heart segments in real-time. This offers a major advantage over traditional methods by providing patient-specific 3D visualization and interaction, crucial for noninvasive diagnosis, surgical planning, and advanced medical training.

Back to Showcase