Análisis de reducción de ruido en señales EEG orientado al reconocimiento de patrones

Abstract

A study on background noise reduction (denoising) on EEG signals using wavelet transform is presented, assuming that extracted features are susceptible to common noise within classes; besides, the feature space separability is compared using a linear Bayesian classifier. An increment of 1% in the average recognition rate is reached performing noise reduction in the identification of two functional states.