Investigating Wavelet Energy Vector for pre-term labor detection using EHG signals

Abstract

Neonatal death can be prevented by early prediction of pre-term labor. During the last decade, uterine electrohysterography (EHG) signal has been considered as a noninvasive method for this aim. There is a wide range of researches which investigated EHG signals for diagnosis of pre-term labor. In this article, features have been extracted by Discrete Wavelet Transform (DWT) from EHG signals then Support Vector Machine (SVM) is applied to classify. The novelty of this research is to introduce Wavelet Energy Vector (WEV) as a feature. The dataset of this research consists of 26 records from term delivery (duration of pregnancy ≥37 weeks) and 26 records from pre-term delivery (duration of pregnancy <37 weeks). The data are divided into train (75%) and test (25%) sets. The obtained results show the proposed feature can achieve the accuracy of 86.67%.