E.O.G classification based on fuzzified symbolic representation

Abstract

This work is part of a study for a 3D navigation task of disabled person in virtual environment. The hardware used is an ocular command device called Cyclope which is based on the measurement and analyze of biological signals that vary with eye movements (Electro Oculo-Graphic, EOG). This paper presents a method to represent and classify EOG signal. The goal is to distinguish Ocular Command Movement which can be detected from pattern of EOG signals. Then, an adapted representation and classification of these patterns allows identifying a particular class of OCM. The proposed method extracts semi-qualitative linear episodes from biological signals. Two membership functions are implemented to transform the quantitative information in the linear episodes to a purely symbolic representation. Finally, the symbolic data is classified with qualitative similarity index.