Extraction of Cardiac and Respiratory Motion Information from Cardiac X-Ray Fluoroscopy Images Using Hierarchical Manifold Learning

摘要: We present a novel and clinically useful method to automatically determine the regions that carry cardiac respiratory motion information directly from standard mono-plane X-ray fluoroscopy images. demonstrate application of our for purposes retrospective gating Validation is performed on five imaging sequences comprising total 284 frames patients undergoing radiofrequency ablation treatment atrial fibrillation. established end-inspiration, end-expiration systolic with success rates 100%, 100% 95.3%, respectively. This technique images and, unlike many previously proposed techniques, does not require specific catheters be visible works without any knowledge catheter geometry.