Please use this identifier to cite or link to this item:
|Title:||Artifacts characterization and removal in EEG brain-computer interface||Authors:||Ren, Shu Heng||Keywords:||DRNTU::Engineering||Issue Date:||2016||Abstract:||Neurophysiological activities of brain commonly recorded utilizing the Electroencephalography (EEG) is contaminated with various types of artifacts resulting from eye blinking, saccades and xations, muscle movement, cardiac signals, and power line interference. There are so far three major methodological techniques in removing these artifacts from EEG signals: Regression, Blind Source Separation (BSS), and Wavelet Transfomation. Each methodology has its merits, as well as some signi cant limitations. Numerous hybrid methods are proposed and developed in recent researches recommending the hybrid composition of conventional techniques are more e ective for detecting and demoising a wide variety of artifacts in EEG signals. This paper presents a comparative study on three state-of-the-art automatic robust hybrid artifact removal methods: Automatic Artifact Removal (AAR), Fully Automated Statistical Thresholding for EEG artifacts Rejection (FASTER), and Fully Online and automated artifact Removal for brain-Computer interface (FORCe). In comparison, the metrics are devised to evaluate the performance of each method according to root mean squared error, mean absolute error, correlation coe cient, power spectral density suppression and distortion as well as visual inspection on the EEG after statistical computations. And in compliance with the vigorous comparison metrics and evaluation results, the conclusion are made to yield an optimum approach among the proposed EEG artifacts removal methods.||URI:||http://hdl.handle.net/10356/68067||Rights:||Nanyang Technological University||Fulltext Permission:||restricted||Fulltext Availability:||With Fulltext|
|Appears in Collections:||EEE Student Reports (FYP/IA/PA/PI)|
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.