摘要：The autocorrelation algorithm is the most commonly used method for extracting fetal heart rate from ultrasound Doppler fetal monitors. The traditional autocorrelation algorithm can not always extract the detection cycle accurately. During the calculation process, the heartbeat cycle may not be recognized, or the cycle may be doubled or halved recognized. Combining the characteristics of envelope curve with average magnitude difference function curve, this paper designs a set of extreme point search scheme and a fetal heart cycle recognition model based on ensemble learning to assist in screening the best fetal heart cycle. The aim of this study is to improve the precision of the fetal heart rate calculation. The experimental results show that the proposed method can effectively screen out the best fetal heart cycle with enhanced reliability and robustness.
关键词：autocorrelation algorithm; ensemble learning; fetal heart rate(FHR); average magnitude difference function
