Study on method of filtering noises from electroencephalography signals and its application for identification of several electroencephalography signals

Abstract

Electroencephalographic (EEG) signals have usually been affected by different types of noise as 50 Hz noise, mechanical noise caused by body movements, heart disturbance, eye noise... In this paper, methods such as: independent component analysis (independent component analysis-ICA), discrete wavelet transform and design of digital filters, were used to filter the noises, to classify the basic components for EEG signals. Then the mean of energy value was calculated to identify the status of the EEG signals such as blink, thoughts, emotion, smoking and blood pressure. The results of calculations and simulations of signals EEG could demonstrate the efficiency of the method.