Full Text

Turn on search term navigation

© 2023. This work is published under https://creativecommons.org/licenses/by-nc-nd/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.

Abstract

Palabras-clave: Aprendizaje automático; biometría; emociones; señales EEG Abstract: Biometric identification is in constant development as well as the systems that violate it, so it is an open field of research that requires new analysis and application of techniques to identify its vulnerabilities and improve its reliability levels. In this work we propose a biometric identification system based on EEG signals in multiple emotional states considering that the underlying dynamics of EEG signals vary according to the emotional state, which may affect the accuracy of classification models. El algoritmo de bosques aleatorios demostró la mayor exactitud (94%) superando los demás algoritmos. 2. Diferentes medidas fueron obtenidas para cada cada una de las señales EEG, a partir usando características derivadas de moving average, linear dynamic systems (LDS), rational asymmetry (RASM), computed differential symmetry (DASM), differential entropy (DE), differential caudality (DCAU), power spectral density (PSD), asymmetry (ASM).

Details

Title
Metodología para la identificación biométrica a partir de señales EEG en múltiples estados emocionales
Author
Duque-Mejía, Carolina 1 ; Castro, Andrés 2 ; Duque, Eduardo 3 ; Serna-Guarín, Leonardo 4 ; Lorente-Leyva, Leandro L 5 ; Peluffo-Ordóñez, Diego; Becerra, Miguel A

 Instituto Tecnológico Metropolitano, 050042, Medellín, Colombia 
 Institución Universitaria Pascual Bravo, 050042, Medellín, Colombia 
 Institución Universitaria ESUMER, 050042, Medellín, Colombia 
 SDAS Research Group (https://sdas-group.com/), Ben Guerir 43150, Morocco 
 College of Computing, Mohammed VI Polytechnic University, Ben Guerir 43150, Morocco 
Pages
281-288
Publication year
2023
Publication date
Oct 2023
Publisher
Associação Ibérica de Sistemas e Tecnologias de Informacao
ISSN
16469895
Source type
Scholarly Journal
Language of publication
Spanish
ProQuest document ID
2880949468
Copyright
© 2023. This work is published under https://creativecommons.org/licenses/by-nc-nd/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.