A New Approach for Electrooculogram Recognition Algorithms

In this work we analyze electrooculogram (EOG) signals’ properties, and propose two recognition algorithms for translating signal patterns into actual eye movements. The ob- jective is to provide a reliable and low cost classification method for Human Machine Interface (HMI) applications. An EOG signal database is generated through an acquisition system. This database is later used to validate the proposed pattern recognition methods, in which the discrete wavelet transform (DWT) was applied to represent the signal with less coefficients. Finally, the results are presented and compared with other extraction methods to distinguish patient’s intents through EOG.