A Hybrid EMD-Kurtosis Method for Estimating Fetal Heart Rate from Continuous Doppler Signals

Abstract

Monitoring of fetal heart rate (FHR) is an important measure of fetal wellbeing during the months of pregnancy. Previous works on estimating FHR variability from Doppler ultrasound (DUS) signal mainly through autocorrelation analysis showed low accuracy when compared with heart rate variability (HRV) computed from fetal electrocardiography (fECG). In this work, we proposed a method based on empirical mode decomposition (EMD) and the kurtosis statistics to estimate FHR and its variability from DUS. Comparison between estimated beat-to-beat intervals using the proposed method and the autocorrelation function (AF) with respect to RR intervals computed from fECG as the ground truth was done on DUS signals from 44 pregnant mothers in the early (20 cases) and late (24 cases) gestational weeks. The new EMD-kurtosis method showed significant lower error in estimating the number of beats in the early group (EMD-kurtosis: 2.2% vs. AF: 8.5%, p < 0.01, root mean squared error) and the late group (EMD-kurtosis: 2.9% vs. AF: 6.2%). The EMD-kurtosis method was also found to be better in estimating mean beat-to-beat with an average difference of 1.6 ms from true mean RR compared to 19.3 ms by using the AF method. However, the EMD-kurtosis performed worse than AF in estimating SNDD and RMSSD. The proposed EMD-kurtosis method is more robust than AF in low signal-to-noise ratio cases and can be used in a hybrid system to estimate beat-to-beat intervals from DUS. Further analysis to reduce the estimated beat-to-beat variability from the EMD-kurtosis method is needed.

Hypertensive Disorders of Pregnancy: Kurtosis-Based Classification of Fetal Doppler Ultrasound Signals

“…Our results revealed that if the number of features in SVM increases, the specificity incearses by 16.5% and the precision of the classification increases by 11.5% as compared to the reference results obtained by [CostaEn, SampEn] features [5], [6], [7], however the sensitivity decreases. This decrase in sensitivity supports the findings of Al-Angari et al [13] which showed that such types of kurtosis-based features are not only sensitive to detect IUGR but also sensitive to cardiac diseases. Kurtosis results were precise (55-100 %) which is in agreement with the work done by Fan and Zuo [17].…”

“…Kurtosis is the measure of the thickness or heaviness of the tails of a distribution of the signal, and the measure of the "tailedness" of the probability distribution [13]. Kurtosis is a descriptor of the shape of a probability distribution.…”

“…As FHRs have not only been explored using the theory of complexity [6], [11], [12] or through machine learning and classification. Recently, a crucial study extracted kurtosis feature from cardiac signals and used classification approaches to detect pathologies [13]. However, as per our thorough review, kurtosis is not yet applied to FHRs.…”

Independent Analysis of Decelerations and Resting Periods through CEEMDAN and Spectral-Based Feature Extraction Improves Cardiotocographic Assessment

“…The empirical mode decomposition (EMD) [48], which corresponds to a previous version of this technique, has been used in different biomedical engineering approaches [49]. Particularly, in FHR signal analysis it has been used for FHR estimation from Doppler Ultrasound signals [57], analysis of the FHR signal components in order test the reliability of the EMD performance [58], FHR baseline estimation with analysis of fetal movements [59], assessment of the high frequency FHR information [60], enhancement of the CTG signal quality by reducing signal artifacts [50], and FHR signal feature extraction and classification [49,61]. These works have concluded that EMD could be appropriate for the analysis of the FHR signal components in the time-domain.…”

A Novel Technique for Fetal Heart Rate Estimation Based on Ensemble Learning

“…Marzbanrad et al (2014) proposed an automated estimation of fetal cardiac intervals from a DUS signal based upon a unique combination of empirical mode decomposition (EMD) and hybrid support vector machine-hidden Markov model(SVM/HMM). Al-Angari et al (2017) proposed a new method based on EMD and kurtosis statistics to estimate fetal heart rate and variability of fetal Doppler ultrasound signal, showing excellent performance in estimating fetal beat to beat interval. Recently, Seong (2018) has disclosed a method for extracting fetal heart rate in, in which average magnitude difference function (AMDF) and ACF are used.…”