一种抑制IVUS图像序列中运动伪像的方法

摘要

针对连续回撤超声导管采集的冠状动脉血管内超声(intravascular ultrasound,IVUS)图像序列中存在由周期性心脏运动所致运动伪像的问题,提出一种抑制方法,以改善IVUS纵向视图的视觉效果.首先,在综合分析运动伪像产生机制和表现形式的基础上,对相邻帧之间血管壁的刚性运动进行定量估计；然后,采用谱分析方法完成刚性运动参数中运动分量和几何分量的分离；最后,通过对各帧图像中血管壁的运动分量进行补偿,完成对运动伪像的抑制.该方法既不需要专门的ECG门控图像采集装置,也无需记录ECG信号,同时不需要抛弃有用帧,在抑制运动伪像的同时保证图像数据集合的完整性.对模拟图像和临床图像的实验验证了方法的可行性.%A method for suppressing motion artifacts due to cyclic cardiac motion existing in continuous pullback coronary artery intravascular ultrasound (IVUS) image sequence acquired in vivo is proposed. Firstly, rigid motion parameters of cross-sectional vessel wall in tomographic image sequence are quantitatively estimated based on comprehensive analysis of both mechanism and expressions of motion artifacts. Then, dynamic and geometrical components of rigid motion parameters are separated with spectrum analysis technique. Finally, geometrical components of vessel wall motion in successive tomographic frames are quantitatively compensated to implement suppression of motion artifacts. Neither special ECG-gating acquisition device nor ECG signals are required. It is not required to discard useful frames to guarantee the integrity of the image data set. The validity has been demonstrated with computer-simulated and clinically acquired image data.