Linear Decoding of Eye Gazing Target Continuous Motion Information via Electrooculography
Zhang Jin-hu · Xi'an Jiaotong Daxue xuebao · 2013
To obtain continuous motion information of eye gazing target,an electrooculography(EOG)-based eye movement decoding method is proposed.Adopting the wavelet transform and double threshold,a new blink detection algorithm is suggested,which is able to distinguish and separate conscious single and double blink from raw vertical EOG signal.Following the baseline and noise removal,the eye gazing target motion information decoding model is established on the basis of EOG,and the parameters of the model are evaluated by recursive least squares algorithm.The experimental results show that the correct rate of blink detection achieves94.7%,the model is endowed with quite fine decoding property,and the offline statistical decoding accuracy approaches to 11.32mm in horizontal and 16.9mm in vertical.