Detection and Analysis of Fetal Movements Using an Acoustic Sensor-based Wearable Monitor

Abstract

Monitoring of fetal movements (FM) is considered an important part of fetal well-being assessment due to its association with several fetal health conditions, e.g. fetal distress, fetal growth restriction, hypoxia, etc. However, the current standard methods of FM quantification, e.g. ultrasonography, MRI, and cardiotocography, are limited to their use in clinical environments. In this paper, we evaluate the performance of an acoustic sensor-based, cheap, wearable FM monitor that can be used by pregnant women at home. For data analysis, we develop a thresholding-based signal processing algorithm that fuses outputs from all the sensors to detect FM automatically. Obtained results demonstrate the promising performance of the system with a sensitivity, specificity, and accuracy of 83.3%, 87.8%, and 87.1%, respectively, relative to the maternal sensation of FM. Finally, a spike-like morphology of acoustic signals corresponding to true detected movements is found in the time-frequency domain through spectrogram analysis, which is expected to be useful for developing a more advanced signal processing algorithm to further improve the accuracy of detection.