Title :
Feature selection in categorizing activities by eye movements using electrooculograph signals
Author :
Mala, S. ; Latha, K.
Author_Institution :
Dept. of CSE, Anna Univ., Tiruchirappalli, India
Abstract :
Eyes are the windows to the brain and the eye movements are a rich source of information in information processing. The aim of this paper is to select the features with CBFS Feature selection algorithm using eye movements by ElectroOculoGraph (EOG) signals during reading and writing task. The objective is to impart the fundamental functionality to get an extensive understanding of how EOG signals can be applied in human computer interaction(HCI) and what can be inferred from those signals using feature selection and data mining classification techniques. This paper first identifies the importance of eye movements and EOG signals then analyze EOG signals by CBFS (clearness based feature selection), mRMR (minimum redundancy maximum relevance) feature Selection methods and the third section demonstrates the performance of data mining classification in EOG signals.
Keywords :
brain; data mining; electro-oculography; feature selection; human computer interaction; medical signal processing; CBFS feature selection algorithm; EOG signals; HCI; brain; clearness based feature selection; data mining classification techniques; electrooculograph signals; eye movements; fundamental functionality; human computer interaction; information processing; mRMR; minimum redundancy maximum relevance; reading; writing task; Classification algorithms; Electrodes; Electrooculography; Feature extraction; Human computer interaction; Tracking; Writing; CBFS; Classification; Eye Movements; Feature Selection; mRMR;
Conference_Titel :
Science Engineering and Management Research (ICSEMR), 2014 International Conference on
Print_ISBN :
978-1-4799-7614-0
DOI :
10.1109/ICSEMR.2014.7043559