A Research on EOG Feature Parameters Extraction Based on Linear Predictive Coding Model

EOG is a very effective eye movement recording method, which is not only noninvasive, but also can record any eye movements´ information. In order to extract some features come form these information contained in EOG automatically, this paper presents a novel EOG feature parameters extraction algorithm. The proposed algorithm includes three main parts. The first is endpoint detection unit which detects the startpoint and endpoint of EOG signals. The second is a preprocessor which consists of band-pass filter, frame blocking procedure and windowing step. The third is feature parameters extracting part which extracts EOG feature by a linear predictive coding model and converts LPC parameters to LPC cepstral (LPCC) coefficients. Finally, the paper gives experimental results.