Clasificación de señales EEG bajo paradigma de imaginación motora utilizando descomposición wavelet y máquinas de soporte vectorial

A BCI (Brain-Computer Interfaces) system acquires and interprets brain signals in order to allow the user interaction with the environment through specific actions. This systems have had great development in recent years due to their great applications as tools for people with severe motor disabilit...

Full description

Bibliographic Details
Author: César Eduardo Hernández González
Format: master thesis
Status:Versión aceptada para publicación
Publication Date:2017
Country:México
Institution:Instituto Nacional de Astrofísica, Óptica y Electrónica
Repository:Repositorio Institucional del INAOE
Language:Spanish
OAI Identifier:oai:inaoe.repositorioinstitucional.mx:1009/853
Online Access:http://inaoe.repositorioinstitucional.mx/jspui/handle/1009/853
Access Level:Open access
Keyword:info:eu-repo/classification/Wavelet/Wavelet
info:eu-repo/classification/SVM/SVM
info:eu-repo/classification/EEG/EEG
info:eu-repo/classification/Clasificación de señal/Signal classification
info:eu-repo/classification/cti/1
info:eu-repo/classification/cti/22
info:eu-repo/classification/cti/2203
info:eu-repo/classification/cti/330790
Description
Summary:A BCI (Brain-Computer Interfaces) system acquires and interprets brain signals in order to allow the user interaction with the environment through specific actions. This systems have had great development in recent years due to their great applications as tools for people with severe motor disabilities, so they can communicate or control devices to improve their daily life. BCI systems based on EEG (electroencephalography) is counted among the non-invasive techniques and it is possible to be used by low cost equipment. In this thesis electroencephalography signals were analyzed under motor imagery paradigm, specifically the intention of movement of tongue, right hand, left hand and feet. Those signals were take from BCI Competition IV database, available to all community for academic and research propose. The proposed algorithm is conformed for feature extraction and classification of brain signals. In the first stage two techniques decomposition were used : Discrete wavelet transform (DWT) and maximum overlap discrete wavelet transform (MODWT). After this stage, a mean, median, variance and entropy measures were taken using maximum overlap. In the classification stage three types of Support vector machines were used with lineal, quadratic and cubic polynomial kerne for each learning machine. Algorithms and results were fully implemented using ®Matlab2016b.