| Summary: | 碩士 === 國立成功大學 === 醫學工程研究所碩博士班 === 90 === Abstract
The auscultation of the abdomen is a noninvasive and simple way to provide the clinical diagnosis reference for gastrointestinal disease. However, auscultation is a time consuming and subjective work. Therefore, it is needful to develop a computer-aided bowel sound monitoring system. In this study, an adaptive filter was designed to reduce the environmental interference, then the CoBliSS ( Convolutive Blind Signal Separation) algorithm was employed to extract the pure bowel sound.
The bowel sound mornitoring system includes analog hardware and digital signal processing algorithm. The analog hardware is composed of microphone, pre-amplifier, band-pass filter and post-amplifier circuit. The bowel sound signal passed through an analog circuit, then through an Analog/Digital converter stored in PC for off-line processing. An adaptive filter algorithm based on Labview environment was used to reduce environmental interference. The CoBliSS algorithm developed on Matlab Signal tool box was used to extract the bowel sound from two different sound sources.
The results show that the heart sound (50~150 Hz) is easily eliminated through the band-pass filter, the environmental interference is effectively reduced by the adaptive filter algorithm, and the CoBliSS algorithm extracts the bowel sound efficiently. We hope that this system can be used in clinical applications.
|
|---|
