Analysis of normal and abnormal electrogastrograms using machine learning and internet of things

Abstract

Electrogastrography is a non invasive technique to record electrical activity of the human digestive system The electrical signals generated by the stomach muscles are called Electrogastrograms EGG The features obtained from the EGG signals are highly useful for diagnosis and staging of several digestive diseases In this work a cost effective three electrode measurement system is developed for the acquisition of normal and abnormal EGG signals The work reported in this thesis has been ethically approved by Institutional Ethics Committee Global Hospitals and Health City Chennai The surface electrodes with two different contact diameters 16mm and 19 mm are used to analyze the effect of contact area of the surface electrodes on the quality of measured EGG signals, using four different Information measures such as Renyi entropy Hjorth Parameters Teager Kaiser energy and Fractal Dimensions The EGG signals from normal individuals and diabetic patients N=30+30 are recorded for a period of 10 minutes The experimental procedures are clearly explained and informed consent has been obtained from normal individuals and Type II diabetic patients The recorded normal and diabetic EGG signals are preprocessed using Empirical Mode Decomposition EMD to filter unwanted frequency components and drift effects newline

Description

Keywords

Citation

item.page.endorsement

item.page.review

item.page.supplemented

item.page.referenced