Semi-automatic Syllable Alignment to Improve Efficiency of Labelling to Speech-imagery EEG Data

Abstract

EEG (Electroencephalogram) is an electrical signal representing activity of the brain and have been used for healthcare and brain-machine interface. Recently, researches to estimate imagined linguistic information from EEG signals were launched. These researches need to make labeled EEG dataset that are given boundaries of imagined syllables. In this paper, we propose a semi-automatic syllable alignment method to improve efficiency of the manual labeling to EEG data.