SDUV方法中组织振动信号提取算法比较

组织振动信号的提取是剪切波频散超声振动成像技术(SDUV)方法中的重要步骤.目前,SDUV方法中有两种常用的组织振动信号提取算法,正交解调法(QDM)和互功率谱法(CSM),但是未见比较这两种算法提取质量差异的相关研究.因此,构造了不同信噪比(SNR_U)的参数化仿真超声回波信号模型,分别使用QDM和CSM从超声回波信号中提取组织振动信号,比较了两种方法的提取效果与计算效率.实验结果表明,当SNR_U≥35 dB,两种算法在相同信噪比下提取出的信号所分离出的振动相位结果相近,标准差均小于1°,对于剪切波波速的计算结果没有太大影响.CSM的计算效率低于QDM的计算效率.因此,当SNR_U<35 dB,为了减小振动信号初始相位的提取误差,应该使用CSM提取组织的振动信号.当SNR_U≥35 dB,应该选择QDM提取组织的振动信号,以减少信号处理时间.本研究的发现有助于提高SDUV方法的检测效率.

Comparison of tissue vibration signal extraction algorithms in shearwave dispersion ultrasound vibrometry

Vibration extraction is an important step in shearwave dispersion ultrasound vibrometry (SDUV). There are two primary algorithms for vibration extraction in SDUV, the quadrature demodulation method (QDM) and the cross-spectrum method (CSM). However, the extraction qualities of QDM and CSM are under appropriate comparison. This study aimed at comparing the performance of QDM and CSM for tissue vibration extraction based on a parameterized model under various signal-to-noise ratio of ultrasound echoes (SNR_U). Results show that when SNR_U≥35 dB, the standard deviations of the estimated vibration phase using the vibration extracted by the two methods are close, which has no significant influence on the calculation result of the shear wave speed. The computation efficiency of CSM for vibration extraction is lower than that of QDM. As a conclusion, when SNR_U<35 dB, the tissue vibration should be extracted by CSM to suppress the error of vibration phase estimation. However, when SNR_U≥35 dB, the tissue vibration should be extracted by QDM to reduce the signal processing time. The findings may provide some strategy in SDUV, which can optimize the examination protocol and make it more efficient.