Analysis of medical signals by an automatic method of segmentation and classification without any a-priori knowledge

Abstract

An automatic system for the segmentation and analysis of biomedical signals is developed. These signals must be of an impulsive type, and they can be highly nonstationary. Two main parts exist in the proposed method: first, segmentation of the several tracks constituting the signal, and second, classification of the results of this segmentation. All these processings are automatic and do not need any prior statistical knowledge of the signal. The first field of application for this method is the study of esophageal peristalsis.<>