心动周期性多普勒波形的平均方法

Methods of averaging the Doppler waveforms over cardiac cycles

许多临床上有用的多普勒波形都是随时间变化的，但又近似于周期为心动周期的周期信号．为准确估计这些多普勒波形，并提取出高灵敏度的特征参数，有必要对它们按心动周期进行平均．直观的平均方法可利用外都触发信号如心电图作为同步，本文提出几种只利用多普勒信号本身进行平均的简易方法：数学形态法、小波变换法和自相关函数法．通过对计算机模拟的和临床实测的多普勒信号的分析实验，发现，这些方法均可用于多普勒信号的平均，其中自相关法效果最好．

Many Doppler waveforms useful in clinic are time-varying, which can usuallybe regarded as periodic signals with the period equal to the cardiac cycle. To estimatethese waveforms with high resolution and to extract the high sensitivity characteristicindices, it is necessary to have the waveforms averaged over cardiac cycles. An outertrigger signal, such as that from ECG, can be used for synchronism ill the direct averagnmethod. Several methods in which only Doppler signals themselves are used, such as themathematical morphology, the wavelet transform and the auto-correlation, are describedin this paper. From analysis of the computer simulated and the in vivo Doppler signals,it is found that these methods can be used to average the Dopp1er waveforms. Amongthem, the auto-corre1ation method gives the best results.