EEG artefacts Benchmark Dataset for the Evaluation of Artefact Detection and Correction Pipelines

EEG signals are inherently characterized by significant noise. Traditional methods of denoising, primarily involving manual correction of visual artifacts, are both time-consuming and subject to variability influenced by the expertise and educational background of the professional involved. Conseque...

Full description

Bibliographic Details
Author: Egaña Elosua, Paula
Format: master thesis
Publication Date:2024
Country:España
Institution:Universitat Politècnica de Catalunya (UPC)
Repository:UPCommons. Portal del coneixement obert de la UPC
Language:English
OAI Identifier:oai:upcommons.upc.edu:2117/415200
Online Access:https://hdl.handle.net/2117/415200
Access Level:Open access
Keyword:Fire extinction--Equipment and supplies
Incendis--Extinció--Equip i accessoris
Àrees temàtiques de la UPC::Enginyeria química
Description
Summary:EEG signals are inherently characterized by significant noise. Traditional methods of denoising, primarily involving manual correction of visual artifacts, are both time-consuming and subject to variability influenced by the expertise and educational background of the professional involved. Consequently, there has been a growing interest in automating the detection and correction of artifacts in EEG signals. Despite ongoing efforts to enhance the automation of artifact detection, two primary challenges persist. Firstly, the absence of a comprehensive database inhibits the assessment of model performance across various scenarios with diverse artifact types. Secondly, the lack of standardized metrics hampers the objective evaluation of the quality of the processed signals. This project aims to address these issues by first recording a dataset wherein participants deliberately introduce artifacts. Subsequently, five artifact correction pipelines will be analyzed and compared to assess their effectiveness in artifact correction. Additionally, the project will establish objective metrics facilitating a standardized comparison of the various correction models.