Classifications Of The Vertical And Horizontal Movements Of Electrooculography (EOG) Signals Using Neural Network

Siti Aishah, Zainal (2015) Classifications Of The Vertical And Horizontal Movements Of Electrooculography (EOG) Signals Using Neural Network. Project Report. UTeM, Melaka, Malaysia. (Submitted)

Text (24 pages) Classifications Of The Vertical And Horizontal Movements Of Electrooculography (EOG) Signals Using Neural Network 24 Pages.pdf - Submitted Version Download (181kB)

Abstract

The study of Artificial Neural Network (ANN) in classifying electrooculography (EOG) signals for the vertical and horizontal movement purpose is mainly to validate that ANN is good in classifying data. Hoping that this study can contribute in helping other researchers in choosing the best classifier for their experiment or project in helping the community of an extreme disable society’s life. Also, maybe that someday that this study can help other developer’s project to function smoothly and comfortably in order to give a great service not just for disable users, but also to people that have the interest in trying something new. The two main objectives of this study are to classify the features of EOG specifically on cornea-retinal potential and also to evaluate and validate the features extracted using Neural Network. There are a few steps needed to be completed throughout this study before the ANN can be validated as the best classifier. These steps included the study on ANN, extracting the EOG signals on a few subjects, extracting the features of the signals, testing and training the network, and lastly analyze the results. There will be some hardware used in this study, which are NI MyRIO 1900, Muscle Sensor V3 kit and disposable electrodes. Mostly this study focuses more on analysis research on how the feature signal will be classified and analyzed. There are many methods that can be used in classifying EOG signals, but ANN is the simplest yet can give an excellent classification result.

Actions (login required)

View Item