Skip to Main Content
The abdominal phonogram is a signal recorded by a sensitive acoustic sensor positioned on the maternal womb. The signal conveys information that is valuable for fetal surveillance (e.g., heart sounds), but hidden by maternal and environmental noises. To recover such information, earlier work successfully used single-channel independent component analysis (SCICA) to decompose the phonogram into independent components (ICs). After that, knowing that some ICs belong to the same process, similar ICs were grouped using K -means through similarities in their spectral content, a step that misclassified some fetal and maternal ICs, and consequently, distorted the recovered sources and made them virtually useless for studying fetal condition. Here, the rich time structure of the physiological components underlying the abdominal phonogram is exploited to automatically classify similar components and retrieve the independent sources corresponding to maternal activity (respiration and cardiovascular), fetal heart sounds, and noise. To do so, a periodicity-based analysis scheme is proposed and tested on a dataset composed of 750 ICs extracted by SCICA from segments of 25 single-channel phonograms recorded at fetal gestational ages between 29 and 40 weeks. Based on autocorrelation and power spectral density analysis, this scheme not only manages to quickly and automatically group similar ICs, but also correlates the recovered sources with specific physiological phenomena (either maternal or fetal), which is a further advantage. Further research will be conducted on more phonograms and explore alternatives for dimensional reduction and reconstruction of entire time series suitable for surveillance, not only for fetal well-being, but also for maternal condition.