Research And Implementation On The Technology Of Electronic Fetal Monitoring Analysis

Fetal heart rate monitoring is a necessary routine examination item in obstetric clinic,which has important significance in the health examination of the perinatal fetus.Electronic fetal heart rate monitoring uses ultrasonic and pressure sensors to obtain fetal heart rate,uterus shrink pressure and fetal movement tags,which are the important basis for obstetricians’ clinical diagnosis of intrauterine fetal distress.Therefore,the extraction of fetal heart rate,pattern recognition of cardiotocography and accurate tagging of fetal movement are the key technologies for fetal electronic monitoring.This thesis mainly studies the technology of the extraction of fetal heart rate and the automatic recognition algorithm of fetal movement,and designs and realizes a obstetric central monitoring software system.The research on the extraction of fetal heart rate.According to the features of doppler signal envelope curve and average magnitude difference function curve,this thesis proposes a set of extreme point search schemes and a machine learning model of fetal heart cycle recognition based on ensemble learning,which aims to improve the accuracy of the calculation of fetal heart rate.The experimental results show that the proposed method effectively screens out the optimal fetal heart cycle.Subsequently a model of signal source recognition based on fast Fourier transform(FFT),which uses the Support Vector Machine(SVM)to determine whether the signals are from abdominal aortic by extracting and analyzing the spectrum feature of the signals,is proposed to solve the problem of the error fetal heart rate obtained from the abdominal aortic signals.The experimental results show that the recognition model can achieve the best recognition effect with the rule that the signals are regarded as from abdominal aortic if more than 88% of the signals get the negative output by the model within the time window of more than 6.4 seconds.The study on fetal movement signal recognition.This thesis first calculates the baseline of fetal movement signals acquired by actography from Doppler ultrasound and uses a dynamic threshold method to detect fetal movement.Subsequently a detection algorithm based on multicriteria is used to identify fetal movement by integrating the information of fetal heart rate acceleration,uterine contractions and fetal movement information from actography,and the tagging method of fetal movement has been improved to adapt to the need of real-time tagging from obstetrics experts.The experiment results show that 98.57% of the fetal movement tags recognized by the proposed algorithm are consistent with the manual tags from obstetrics expert and 75.36% of the fetal movement tags meet the real-time tagging requirements,which prove the practicality and effectiveness of the proposed algorithm.Based on the above work and the requirements from hospital obstetrics,a obstetrics fetal monitoring central station system is presented,which can meet the demands of clinical requirements for fetal heart rate monitoring in hospital obstetrics and realize the fetal heart rate monitoring of multi-bed.