VDSB-GWSyn: Diffusion Schr\"{o}dinger Bridge for Controllable and Anatomically Feasible Guidewire Synthesis in Coronary Angiography
Researchers have developed VDSB-GWSyn, a novel framework utilizing a Diffusion Schrödinger Bridge model to synthesize realistic guidewire images for coronary angiography. This method addresses the scarcity of annotated data and improves the accuracy of guidewire endpoint localization, a critical step in computer-assisted and robot-assisted percutaneous coronary intervention (PCI). By generating controllable, anatomically feasible guidewire samples, the framework significantly enhances downstream localization performance, reducing mean positional error and increasing correct localization rates. AI
IMPACT Enhances data availability for medical AI, potentially improving robotic surgery precision and reducing operator radiation exposure.