Enhancement of explosive bowel sounds using Kurtosis-based filtering

An efficient method for the enhancement of bowel sounds (BS), based on kurtosis is presented in this paper. The proposed method employs an iterative procedure to compose a Kurtosis-based Detector (KD), for enhanced separation of explosive BS (EBS) from the background noise. Analysis of the experimental results, when applying the KD to EBS recorded from patients with gastrointestinal dysfunction, justify an efficient performance of the KD to outwit the noise presence by accurately extracting the authentic structure of EBS. Since the KD removes the undesired signal from the EBS recordings, it can easily be used for data-volume reduction when long-term recordings are employed, e.g., in sleep laboratories, intensive care units, homecare telemedical devices. Due to its simplicity, the KD can be implemented in a real-time context and can be used for continuous EBS screening.