The application of median fuzzy clustering and robust weighted averaging for electronystagmography signal processing

Abstract

The purpose of this paper is to present a robust approach to processing of electronystagmography (ENG) signal which contain saccades. The accurate saccade detection and localization allows for recognition of the optokinetic nystagmus (OKN), which is regarded as one of the important phenomena to be evaluated by ophthalmologists. The robust ENG processing consists in robust filtering of the signal to suppress outliers, robust fuzzy clustering to determine the amplitude threshold level for automatic saccade detection and robust averaging of similar optokinetic nystagmus cycles to estimate the nystagmus parameters. The proposed method was tested using the artificial as well as the real signals. The results show the usefulness of the proposed solution in the process of the OKN modeling.