A feasibility study on eye movements using electrooculogram based HCI

Bio signal based Human Computer Interaction has the potential to facilitate severely immobilized people to impel external devices straightly by bioelectricity slightly than by bodily. This paper presents a preliminary study on electrooculography signals for EOG based HCI. Nine different eye movements from six subjects were studied. Statistical method was used to pull out the features. These features were used to train and testing the Time Delay Neural Network. The accuracy of the algorithms have an average classification efficiency of 87.72% was achieved by using Time Delay Neural Network. From the result it is examined that Time Delay Neural Network classifier have better classification for the nine different tasks for each of the subjects compared to feed forward neural network.