Preterm-term birth classification using EMD-based time-domain features of single-channel electrohysterogram data

Abstract

The preterm birth anticipation is a crucial task that can reduce the rate of preterm birth and also the complications of preterm birth. Electrohysterogram (EHG) or uterine electromyogram (EMG) data have been evidenced that they can provide an information useful for preterm birth anticipation. Four distinct time-domain features, i.e., mean absolute value, average amplitude change, difference absolute standard deviation value, and log detector, commonly applied to EMG signal processing are applied and investigated in this study. A single-channel of EHG data is decomposed into its constituent components, i.e., intrinsic mode functions, using empirical mode decomposition (EMD) before their time-domain features are extracted. The time-domain features of intrinsic mode functions of EHG data associated with preterm and term births are applied for preterm-term

Comparison of wavelet‐based decomposition and empirical mode decomposition of electrohysterogram signals for preterm birth classification

“…In addition, in Acharya and others [27], the adaptive synthetic sampling approach (ADASYN) [60] is employed to balance between the number of positive class, that is, preterm birth, and the number of negative class, that is, term birth. Four distinct time-domain features, namely, MAV, AAC, difference in absolute standard deviation value, and log detector extracted from four IMFs of EHG epochs [28] were shown to provide the best F 1 -score, that is, 0.9366. The best accuracy on preterm birth classification that is 0.9776 is however achieved using the selected time-domain features of EHG subbands as reported in this study.…”

“…Quantitative features, which portray underlying characteristics and behaviors associated with physiological and clinical states, are subsequently obtained. A variety of quantitative features including time-domain features [25][26][27][28][29][30][31][32][33], frequency-domain features [13,26,29,[30][31][32][34][35][36][37], and nonlinear features [13,26,[29][30][31][34][35][36][37] are extracted from EHG data and applied for EHG data processing. Quantitative features are extracted from EHG signals through wavelet transforms, that is, DWT and wavelet packet, and applied for preterm birth classification [26,38,39] and labor classification [13,34,40].…”

“…Quantitative features are extracted from EHG signals through wavelet transforms, that is, DWT and wavelet packet, and applied for preterm birth classification [26,38,39] and labor classification [13,34,40]. Likewise, EMD is applied for preterm birth classification [27,28,41] and labor classification [42]. It is shown in literature that the promising performance on preterm birth classification and labor classification can be obtained from several quantitative features [13,27].…”