Powered by OpenAIRE graph
Found an issue? Give us feedback
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/ Estudo Geralarrow_drop_down
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
Estudo Geral
Master thesis . 2018
Data sources: Estudo Geral
addClaim

Classification of gestures with EMG

Authors: Silva, João Ricardo Ferreira dos Santos;

Classification of gestures with EMG

Abstract

Este trabalho tem como objetivo principal a deteção, reconhecimento e classificação de gestos através de dados recolhidos a partir de sensores de eletromiografia superficiais (sEMG). A performance dos sensores é avaliada através a recolha de dados com intenção da criação de uma base de dados offline. A implementação de features e o valor do threshold das mesmas é estudado tanto para a segmentação como para a classificação de gestos. Os melhores resultados obtidos na avaliação offline foram conseguidos com uma base de dados equilibrado entre gestos e não gestos. Para a segmentação os resultados apresentaram um erro de (3.7%) com recurso ao classificador ensemble Gentleboost. Para a classificação os resultados mostram um erro na ordem de (0.9%) e foi obtido com recurso a um classificador K-nn Spearman com 1 vizinho mais próximo e 7 features implementadas. A implementação online começa a ser explorada e é rapidamente apresentada, com erro de 15% na segmentação e 35% na classificação. (((A segmentação de dados de gestos tem sido estudada nos últimos anos, por exemplo (Simão et al. 2017) abordou esse problema com o uso de uma Cyber Data Glove para detectar os gestos de mão e braço do usuário. No entanto, este método tem as desvantagens de ser caro e não prático. Além disso, em (Lopes, João 2016) este problema é abordado usando uma braçadeira MYO que usa sensores EMG combinados com unidade de movimento inercial (IMU). Esta pulseira tinha a vantagem de ser acessível, mas nem sempre era a mais confiável em termos de dados, apresentando algum erro usando os sensores juntos ou separadamente.)))

This work has as its main objective the segmentation and classification of gestures through the data collected from surface electromyography (sEMG) sensors. The performance of the sensors is evaluated through the collection of data to create an offline dataset. The implementation of different features and their threshold values is studied both for the segmentation and classification of gestures. The best offline results were obtained using a balanced dataset and for a library of 7 gesture patterns. For segmentation, the result showed an error of 3.7% using a Gentle boost ensemble classifier. For classification, the results showed an error of around 0.9% and were obtained for a K - Nearest Neighbors (K-NN) Spearman classifier with one nearest neighbor and seven implemented features. The online implementation is briefly presented with the error for segmentation of 15% and an error for classification of 35%. (((The segmentation of gesture data has been studied in the last few years, for example (Simão et al. 2017) addressed this problem with the use of a Cyber Data Glove to detect the hand and arm gestures of the user. However, this method has the drawbacks of being expensive and not practical. Also, in (Lopes, João 2016) this problem is addressed using a MYO Armband that uses EMG sensors combined with inertial movement unit (IMU). This bracelet had the advantage of being affordable, but it was not always the most reliable in terms of data, presenting some error using both the sensors together or separately.)))

Dissertação de Mestrado Integrado em Engenharia Mecânica apresentada à Faculdade de Ciências e Tecnologia

Country
Portugal
Related Organizations
Keywords

EMG, Segmentation, Segmentação, Classificação, Classification, Features

  • BIP!
    Impact byBIP!
    selected citations
    These citations are derived from selected sources.
    This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    0
    popularity
    This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
    Average
    influence
    This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    Average
    impulse
    This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
    Average
Powered by OpenAIRE graph
Found an issue? Give us feedback
selected citations
These citations are derived from selected sources.
This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Citations provided by BIP!
popularity
This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
BIP!Popularity provided by BIP!
influence
This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Influence provided by BIP!
impulse
This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
BIP!Impulse provided by BIP!
0
Average
Average
Average
Green