Classification of Brain Signals with OpenViBE
Laurent Bougrain, Guillaume Serrière · 2016
This chapter examines the software OpenViBE, the classification phase in the processing chain of a brain–computer interface (BCI), that is how to recognize the brain activity of the subject, source of information and commands within BCIs. It follows the phases of data acquisition, preprocessing and feature extraction. It is followed by the translation into commands of the mental state recognized by the system. The chapter describes how to train the parameters of a classifier from learning examples, how to evaluate the performances of the resulting classifier and how to use it to automatically classify new signals. OpenViBE consists of processing modules called boxes. All processing boxes are available in the classification category. The main boxes to use for training are: for classification, classifier trainer and classifier processor, and for evaluation, accuracy, confusion matrix, ROC curve and kappa factor.