eduzhai > Applied Sciences > Engineering >

Classification of Imagined Speech Using Siamese Neural Network

  • king
  • (0) Download
  • 20210507
  • Save

... pages left unread,continue reading

Document pages: 6 pages

Abstract: Imagined speech is spotlighted as a new trend in the brain-machine interfacedue to its application as an intuitive communication tool. However, previousstudies have shown low classification performance, therefore its use inreal-life is not feasible. In addition, no suitable method to analyze it hasbeen found. Recently, deep learning algorithms have been applied to thisparadigm. However, due to the small amount of data, the increase inclassification performance is limited. To tackle these issues, in this study,we proposed an end-to-end framework using Siamese neural network encoder, whichlearns the discriminant features by considering the distance between classes.The imagined words (e.g., arriba (up), abajo (down), derecha (right), izquierda(left), adelante (forward), and atrás (backward)) were classified using theraw electroencephalography (EEG) signals. We obtained a 6-class classificationaccuracy of 31.40 for imagined speech, which significantly outperformed othermethods. This was possible because the Siamese neural network, which increasesthe distance between dissimilar samples while decreasing the distance betweensimilar samples, was used. In this regard, our method can learn discriminantfeatures from a small dataset. The proposed framework would help to increasethe classification performance of imagined speech for a small amount of dataand implement an intuitive communication system.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×