Title :
Development of fetal cardiac intervals throughout 16 to 41 weeks of gestation
Author :
Marzbanrad, Faezeh ; Kimura, Yuichi ; Funamoto, Kiyoe ; Sugibayashi, Rika ; Endo, Miyuki ; Ito, Takao ; Palaniswami, Marimuthu ; Khandoker, Ahsan
Author_Institution :
Univ. of Melbourne, Parkville, VIC, Australia
Abstract :
In this paper a new noninvasive method is proposed for automated estimation of fetal cardiac intervals from Doppler Ultrasound (DUS) signal and fetal electrocardiogram (fECG) as a reference. The proposed method is based on a combination of Wavelet analysis and hybrid Support Vector Machines- Hidden Markov Models (SVM/HMM). This method provides automated beat by beat identification of cardiac valves´ opening and closing which are used to estimate the fetal cardiac intervals. The range of these intervals in different gestational age from 16 to 41 weeks was evaluated in this study. The correlation between beat to beat intervals was also investigated. Results show a significant negative correlation between Pre-Ejection Period (PEP) and Ventricular Ejection Time (VET) (r = -0.61, p <;0.0001), a positive correlation between VET and Systolic Time Interval (STI) (r =0.51, p <;0.0001) and a significant correlation between PEP and Isovolumic Contraction Time (ICT) (r =0.67, p <;0.0001).
Keywords :
biomedical ultrasonics; electrocardiography; hidden Markov models; medical signal processing; obstetrics; support vector machines; wavelet transforms; Doppler ultrasound signal; ICT; PEP; STI; SVM-HMM; VET; automated beat by beat identification; cardiac valve closing; cardiac valve opening; fECG; fetal cardiac intervals; fetal electrocardiogram; gestational age; hybrid support vector machine-hidden Markov models; isovolumic contraction time; noninvasive method; preejection period; systolic time interval; ventricular ejection time; wavelet analysis; Frequency modulation;
Conference_Titel :
Computing in Cardiology Conference (CinC), 2013
Conference_Location :
Zaragoza
Print_ISBN :
978-1-4799-0884-4