Use of Electro Hysterogram (EHG) Signal to Diagnose Preterm Birth

The purpose of the present study is to embark on the uterine electromyogram (EMG) or electro hysterogram (EHG) signals analysis with an aim to efficient diagnosis of preterm birth i.e. birth prior to a thorough 37 weeks of gestation, and one can prevent the preterm birth of 15 million babies per year and reduce the death rate of these babies. These signals provide useful information regarding the uterine contractions for physicians. The algorithm used for the present probe was initially developed by implication of empirical mode decomposition (EMD) on the uterine EMG signals which yield smaller components named intrinsic mode functions (IMFs) followed by calculating of wavelet coefficients for each IMF using wavelet packet decomposition (WPD). AR model was used to extract the assumed features and then the important features are chosen using PSO which is known as a feature selection method. Support vector machine (SVM) was implemented to classify the features, and it is worth noting that by using 10 most superior features, the accuracy rate, sensitivity, and specificity were obtained as 97.1%, 95%, and 99%, respectively.