Drag-based Key-point Control for Myocardial Infarction Echocardiography Video Generation

Recent advancements in data-driven diagnostic systems for myocardial infarction (MI) have been limited by a lack of comprehensive echocardiography datasets. To address this challenge, we propose a key-point drag-based control system that simulates MI in echocardiography videos. Our approach introduces a novel Scatter Tracking technique, specifically designed for echocardiography, which enhances point tracking by leveraging patch-level scatter patterns. The system enables precise control of heart wall movements, allowing for the simulation of region-specific MI with varying severity. Experimental evaluations demonstrate that our method generates high-fidelity, clinically relevant echocardiography videos, showcasing significant potential for improving MI diagnostics.