Characteristics analysis methods for umbilical artery Doppler velocity waveforms

Several new methods are presented to analyze the characteristics of the velocity waveforms recorded at the umbilical artery using the Doppler ultrasound. A new method to extract the maximum velocity waveform from the Doppler sonogram using mathematical morphology is first introduced. Then three methods are described to detect the characteristics of the velocity waveform: (1) a new automatic algorithm with the morphology used to calculate several traditional indices, (2) a velocity waveform pattern recognition method with a backpropagation neural network, and (3) a three poles model in which the characteristics of the velocity waveform can be shown by the distribution of the poles. From the clinical experiments, it is found that these methods can be used to evaluate the state of fetal circulation.