Using morphological filters to extract spiky transients in EEG

Abstract

In this paper, we use four basic morphological operators to construct a bandpass filter, and also investigate the construct method of the filter in detail. Furthermore, we give the algorithm flowchart of this filter. The result of experiment shows that this morphological filter can be applied to eliminate noise from EEG signal by two different length structure elements and it has advanced property that can remove noise without damage signal useful information. Using this algorithm, we can filter noise and extract spiky transients in EEG effectively.