Electrooculogram (EOG) signals as a part of human-controlled interface (HCI) is proposed for detecting the relevant information in EOG with and without delay in movement of eyes. The performance of eye movements is studied with the accuracy in identification of information along with single and double blink. The algorithm consists of a simple first order derivative, threshold windowing technique, and pattern recognition. The EOG pattern recognition was studied with time domain features mean value (MV) and ensemble of MV and zero crossing (ZC). The highest average classification accuracy of 85% and 84.4% is obtained from continuous movement of eyes for three classes (L, R, DB and L, R, SB) with two time-domain features. Further, the accuracy of 90% and 88% from two eye movement detection is obtained.
TopIntroduction
Electric signal are obtained due to the polarization of human eye (Geethanjali et al., 2013). EOG when acquired from two channel provide ten different control signals including single blink and double blink (Manuel et al., 2010). Human eye acts as a dipole and can generate amplitude up to 3mV and the signal is referred as electrooculogram (EOG) (Manuel et al., 2010, Rajan et al.,2006). This EOG signal find potential applications in rehabilitation engineering for physical as well as neuromuscular impaired people to aid in communication and or control of devices. In the last couple of decades, research based on electroencephalogram (EEG) in Brain computer interface (BCI) (Hashimoto et al., 2009) are attempted to help the neuromuscular impaired subjects in communication and control of devices in the environment. EEG signals are also bio electric signals acquired in µV on human scalp (Hashimoto et al., 2009). But, EOG is easy to acquire and has given better efficiency than EEG signals (Hashimoto et al., 2009). EOG signals based interface find application in paralyzed people, capable of moving eyes. These signals may be used to control device such as a wheelchair or computer either moving mouse of a computer as an input or a keyboard input as non verbal information [Manuel et al., 2010, Hashimoto et al., 2009). The purpose with EOG and EEG based interface is to improve the quality of life of individual with impairments and reduce the burden of the care taker. EOG is preferred due to features of low cost, ease in acquisition and processing of signals to identify the intention of the user Hashimoto et al., 2009).
Several devices and machines have been proposed to assist disabled people using EOG signals [Rajan et al., 2006, Hashimoto et al., 2009, Geethanjali and Ray, 2015). Researchers have used several techniques such as threshold windowing, maximum amplitude of peak, maximum amplitude of the valley, position value of maximum peak amplitude and maximum amplitude of the valley, value of the curved area, threshold value crossing, variance of the signal (Wijesoma et al., 2005, Aungsakul et al., 2012) Researchers attempted on filtering and peak amplitude characterization according to eye saccades and blinks (Postelnicu et al., 2012). Researchers also attempted with continuous wavelet transform (CWT) technique to eliminate the fixation problems in detecting eye movements when the patient is tired (Barea et al., 2012). .Authors have designed wheelchair to control it using EOG signals with different techniques and strategies (Barea et al., 2003, Barea et al., 2002). Electroretinography (ERG) is another kind of electrical signal produced due to the photonic accumulation and chemical stimulation in the receptor of the eye (Creel, 2012). Researchers have used spectral analysis of EOG signals to increase the accuracy of recognition (Lv et al., 2010). Authors of this paper have attempted to study the effect of transition on eye movement in identifying the intention of the user. The transition of eye movement had been studied with single channel EOG signals in different experiments and found that that eye movement with delay in transition performs better.
TopElectrooculography Signal Acquisition
EOG is bio electric signals, generated between the retina and cornea with eye saccades, where retina as negative and cornea as positive (Gandhi et al., 2010). Due to this an electric field is generated across the eye. As the eye moves, negative ions are formed in the back side of the eye (retina) and make the front (cornea) positive. This electric dipole field is positioned with the line of electrodes when the eye rotates towards it. Three different experiments were conducted to study the effect of transition in eye movements in decoding the information as explained below.
The subjects were trained to refrain from muscle movements due to any involuntary reflexes to get minimum noise. Acquisition of signal was done across one channel (Horizontal axis).Two AgCl electrodes were placed a cm away across the lateral canthus. TruScan system was used to acquire EOG, using three electrodes. Reference electrode is placed at the forehead. Figure. 1 shows the electrode placement for EOG signals acquisition.