SciELO - Scientific Electronic Library Online

 
vol.14 número4Algoritmos para estimar la frecuencia instantánea de una secuencia respiratoria variable en el tiempoIntegración de rasgos y aprendizaje semi-supervisado para la clasificación funcional de enzimas utilizando K-medias de Spark índice de autoresíndice de assuntospesquisa de artigos
Home Pagelista alfabética de periódicos  

Serviços Personalizados

Journal

Artigo

Indicadores

  • Não possue artigos citadosCitado por SciELO

Links relacionados

  • Não possue artigos similaresSimilares em SciELO

Compartilhar


Revista Cubana de Ciencias Informáticas

versão On-line ISSN 2227-1899

Resumo

DIAZ-AMADOR, Roberto; A.MENDOZA-REYES, Miguel  e  FERRER-RIESGO, Carlos A.. Image-based representation to myoelectric pattern recognition in the presence of inter-sessions variability. Rev cuba cienc informat [online]. 2020, vol.14, n.4, pp.123-133.  Epub 01-Dez-2020. ISSN 2227-1899.

The myoelectric control systems based on pattern recognition are able to adequately classify the movement intention from the surface EMG signal. However, these systems present intersession variability, reporting a drop in performance in the test sessions compared to the training session. The objective of this work is to investigate an alternative representation of the HD-EMG signal based on imaging features to improve inter-session recognition. In this work, we implement features calculated from an instantaneous 2D representation that is formed by considering each sample of the HD-EMG signal as a pixel of an image. This feature set is compared with the use of temporal domain features calculated from each channel of HD-EMG. The time-domain features considered are the absolute mean value, the number of zero crossings, the length of the waveform, and the sign change of the slope. The features from the 2D representation that are considered are based on the pixel value and based on the texture. The proposed features improve by 15% (p <0.05) the use of the time-domain features when each session is used as a training and the other as a test. The results suggest that using features from the 2D representation proposed in this work show greater robustness to inter-session variability.

Palavras-chave : myoelectric control; image processing; inter-sesion variability.

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )