Analysis of Fetal Heart Sound for Fetal Heart Rate Monitoring

Abstract

Perinatal mortality is the number of stillbirths and deaths during the first seven days of life per 1000 live births. Most (98%) of this death occurs in low- and middle-income countries (LMIC) and is known as the 'silent' epidemic. Perinatal mortality reflects the quality of healthcare for mothers and new-borns. Asphyxia, one of the causes of perinatal mortality, is oxygen deprivation due to placental obstruction. Fetal monitoring helps identify asphyxia, wherein timely detection and intervention may be lifesaving for the new-born. Monitoring the signs of fetal compromise is essential to reduce perinatal mortality, and fetal heart rate trace can provide information about fetal compromise. Fetal heart rate (FHR) monitoring detects the change in the heart rate patterns by measuring the heart rate and rhythm of the fetus. The normal FHR range is between 120-160 bpm. Any deviation from this range indicates abnormality and pathological conditions. Therefore, correct, and reliable identification of FHR is necessary for prenatal clinical investigations. Fetal phonocardiography (fPCG), due to its non-invasive and passive nature, has become a focus of several studies for FHR determination and its variability. fPCG is recording of fetal heart sounds (FHS) by an acoustic sensor placed on the mother's abdomen. The fPCG recordings generally have a low signal-to-noise ratio (SNR) due to internal and external noises. Despite its potential, the real diagnostic value of the fPCG is not demonstrated to date, resulting in restricted clinical practice. The doctoral work aims to develop robust algorithms for extracting accurate and reliable FHR measurements from the fPCG recordings with different data quality and multiple sensors for the fPCG acquisition in the presence of noise.