Título: Classificação Com Máquina De Vetores De Suporte E Coerência Na Imagética Motora Com Prática Mental
Autores: Erazo-Costa, F.; Santos, C.C.; Gomes, M.E.D; Tierra-Criollo, C.J.
Resumo: A brain computer interface (BCI) is a system that communicates a person with an electronic device by means of mental signals. Pattern recognition is an important module for BCIs because it translates brain activity into commands for the activation of a machine. This paper aims at presenting the relation between mental practice and the algorithm classifier support vector machine in combination with the magnitude squared coherence for the classification of motor imagery movements of the index finger (imagination and spontaneous). An accuracy of 98 % was found with one subject with the training of the machine with data for the four sessions and validation of session 5. Furthermore, an incremental accuracy of the classification around the different sessions was found for all subjects with mental practice. This approach could be used for the development of BCIs.
Palavras-chave: Mental Practice; coherence; SVM
Código DOI: 10.21528/CBIC2011-13.5
Artigo em pdf: st_13.5.pdf
Arquivo BibTex: st_13.5.bib