Quantum mechanics

Hilbert–Huang transform

Biology

Support vector machine

Preterm labor

Labor pain

Fetus

Computer science

Filter (signal processing)

Genetics

Physics

Pattern recognition (psychology)

Artificial intelligence

Medicine

Term (time)

Preterm delivery

Pregnancy

Computer vision

Preterm birth is the leading cause defining the infant mortality and morbidity globally. Non-invasive surface uterine electromyogram (sEMG) also known as Electrohysterogram (EHG) is the most promising biophysical signature for the study of uterine contractions. Therefore, it can prove to be a marker for the detection of preterm birth, which might enable us to diagnose the preterm birth before the labor. In this study, we proposed a signal processing approach to predict Preterm birth using raw EHG signals with shorter time recording (1 min). The raw EHG recording is first preprocessed and segmented using Empirical Mode Decomposition by selecting only first intrinsic mode function. Only four features namely Shannon Energy, Log Energy, Median Frequency and Lyapunov Exponent, extracted from segmented EHG record are fed to Support Vector Machine classifier. The system achieves 95.5% accuracy on publicly available Term-Preterm EHG Database. Such an accurate system will help medical professionals to make effective decisions about the treatment. Hence the expectant mothers undergo minimal or no complications of preterm labor. On the other hand, it also helps to avoid unnecessary hospitalization and treatment for women who are having a false labor pain.

Characterization of Term and Preterm Deliveries using Electrohysterograms Signatures