作者: Zhao Lv , Xiao-pei Wu , Mi Li , Dexiang Zhang
DOI: 10.1016/J.PATREC.2009.12.017
关键词:
摘要: Bio-based human computer interface (HCI) has attracted more and attention of researchers all over the world in recent years. The paper is concerned with eye movement detection algorithm EOG-based interface. linear predictive coding cepstrum (LPCC) coefficients EOG pulses are extracted as feature vectors, which used for pattern matching. Dynamic time warping (DTW) adopted to solve discrepancy duration among different trials. Furthermore, spectral entropy detect endpoints pulses, can improve robustness increase recognition rate noisy background. Experimental simulation results based on real-life signals show that proposed stable performance be online controlling communication HCI system.