Contribution
Analysing Bowel Sound Signals: Insights from SonicGuard Data Collection
* Presenting author
Abstract:
Auscultation by doctors is a standard diagnostic method to assess the health status of a patient. Although being performed by experts, this procedure is only a short-term analysis. Recent developments and progress in medical IoT and wearable sensor systems now also allow for long-term monitoring of bodily sounds via sound analysis of different organs such as the heart, lung and bowel.Being able to capture more complex continuous time-series over a longer period of time systematically potentially leads to high quality datasets as a basis for the development of automatic analysis systems in support healthcare applications. Although a few dataset exists, we specifically identified a lack of a high quality dataset for bowel sound signals and digestive cycles. This study introduces SonicGuard as a multichannel acoustic sensor system design for capturing body sound in general and bowel sound in particular from 20 healthy subjects in PIUS hospital Oldenburg. The study covers the measurement campaign so far, taken quality measures, a descriptive analysis of the bowel sound signals, and reveals statistical analysis and semantics on captured bowel sound types as a multiclass classification problem.