【摘要】 目的:提出一种基于独立分量分析的多普勒胎心音瞬时心率的检测方法。方法:首先对由超声多普勒胎心仪采集到的心音信号进行预处理,再对预处理后信号进行时频分析,然后将胎心音信号的时频谱进行实数化处理和对数化处理,接着将对数化以后的胎心音信号时频谱以不同的频率分量进行独立分量分析分离出独立分量,最后选取与原信号相关性最强的一个独立分量并对该独立分量进行峰值检测计算出胎儿瞬时心率。结果:该方法不仅能识别胎心音信号的幅值信息,同时能识别其频域信息,多次实验表明,该算法准确率高达95%以上。结论:该方法能够在强噪声背景下快速地分离出理想胎心音信号的包络信号并计算胎儿瞬时心率,对围产阶段胎心监护具有较为重要的意义。更多还原

【Abstract】 Objective To develop a detection method for instantaneous heart rate of Doppler fetal heart sound signal based on independent component analysis(ICA). Methods The heart sound signal acquired by color Doppler fetal heart rate monitor underwent pretreatment, and then went through time-frequency analysis. The obtained time-frequency spectrum of the fetal heart sound signals were turned into real numbers and logarithms, then independent components were extracted from the logarithmized time-frequency spectrum with different frequency components. The independent component related the most to the original signal were found to calculate instantaneous heart rate by peak detection. Results This method could recognize the information on the amplitude and frequency domain of the fetal heart sound signal, and the trials proved the accuracy of the method was more than 95%. Conclusion The method can isolate the envelope signal of desired fetal heart sound signal and calculate instantaneous fetal heart rate under strong noise conditions, and thus is of great significance for the monitoring of fetal heart during perinatal period.更多还原