Study On Extraction And Classification Of Eye Movement Signals

Human computer interaction is a bridge between people and computer or other electronic equipment. In recent years, with the rapid development of science and technology, human-computer interaction technology has made greater progress, and constantly to more natural, more harmonious, more convenient direction of development. As the organism itself contains a variety of biological signals, such as EEG, EMG, ECG, eye movement signal etc.. People can be collected on these signals and effectively decode these signals and then gives specific meaning, so that it can be used as human beings. Use these rich vital signs information resources to develop more interactive with human-computer interaction system of human oriented ",gradually become researchers and experts continue to explore problems.The eye movement signal is a kind of biological electrical signal that can cause the change of the electric potential around the eye caused by the movement of the eyes. Along with the eye movement and eye movement signals will change, the signal waveform characteristics and eye movements have direct correspondence, and eye movement signals with higher amplitude, waveform is easy to check, easy to deal with and other advantages. Therefore, it is a new direction to solve the problem of human computer interaction based on the eye movement signal, which is based on the eye movement signal.In this paper, we mainly study the eye movement signal acquisition and preprocessing, endpoint detection, effective eye movement signal extraction, feature extraction, and respectively using BP algorithm and SVM algorithm and DTW algorithm of eye movement signal for pattern recognition research. Among them, the DTW algorithm effectively solve eye movement signal varies from person to person,of varying lengths, and improve the eye movement signal recognition rate, achieved good results, to lay a foundation for the study of design of human-computer interaction system based on eye movement signal. This paper mainly completed the following main work:1.Eye movement signal acquisition and preprocessing: the design of the eye movement signal acquisition experiment and the number of subjects for data collection, access to a large number of original eye movement signal data. Andthrough the hardware circuit to the original data of the initial filtering, noise reduction processing.2.Extraction of the effective signal: the signal is processed by software filtering,framing, adding windows, calculating the short-time energy and endpoint detection.3.Feature extraction and analysis of eye movement signal: according to the waveform characteristics of the effective signal, the analysis can be used for pattern recognition feature value, extract the characteristic value of the eye movement signal based on the waveform characteristics.4.Using BP algorithm, SVM algorithm and DTW algorithm to classify and identify the eye movement signals, the advantages and disadvantages of the three methods are discussed.