Redes Neurais Profundas Triplet Aplicadas à Classificação de Sinais em Interfaces Cérebro-Computador

In this work, proposes an approach to EEG signal classification in brain-computer interfaces based on triplet deep neural networks. The network is tested over a database with ten users, in a setup that excludes the evaluated user's data from the network training process. The results are promising, although further investigation is necessary to verify if there are additional nonlinear relationships in the data to be explored.