摘要
射频导管消融手术已经开始成为一种常见的针对心律失常手术的治疗方案。从动态X线图像中自动跟踪导管尖端可以为计算机辅助诊断和导管三维姿态重构提供基础。本研究的目标是发展一种稳健的能够自动识别和跟踪消融大头电极导管的方法。为此,本文提出了一种结合阈值法和形态学运算识别消融大头电极导管尖端,然后利用光流法对动态X射线图像中消融大头电极导管尖端进行自动跟踪的方法。利用临床上34个动态X射线图像序列验证该算法,跟踪结果的平均误差为1.126 mm±0.748 mm,跟踪每帧图像耗时大约7 ms,因此该方法能比较精确地实时跟踪消融大头电极导管尖端在动态X射线图像中的运动。
Radio frequency catheter ablation （RFCA） has become a common treatment option for atrial fibrillation. Automatic tracking of the catheter tip from the fluoroscopic images may provide a basis for computer-aided diagnosis and 3D pose reconstruction of the catheter. Our goal was therefore to develop a robust algorithm that can automatically detect and track the ablation catheter tip. Our method uses threshold in combination with morphological operations to detect the catheter tip, then applies the Lucas - Kanade algorithm to track the catheter tip in the following images. Our method allows for an accurate and robust tracking of the ablation catheter tip with a mean error of 1. 126 ± 0. 748 mm for all frames. This is demon- strated by 43 series of fluoroscopic images from different patients.
出处
《中国体视学与图像分析》
2012年第1期12-19,共8页
Chinese Journal of Stereology and Image Analysis
基金
国家自然科学基金项目(81127003)
973项目(2011CB707701)
关键词
光流法
导管跟踪
射频消融
optical flow method
catheter tracking
radio frequency catheter ablation
rfea
cardiac atrialfibrillation
