Recognition of Mentally Pronounced Russian Phonemes Using Convolutional Neural Networks and Electroencephalography Data

We analyze a classification problem of mentally pronounced Russian phonemes based on data obtained by means of an electroencephalography device. We describe the data collection method as well as the methods of the obtained data processing. To solve the small sample size problem we present the augmentation techniques that use the time stretching and the white noise adding. Our approach uses an algorithm based on the convolutional neural networks and it is applicable to solving the binary and multiclass classification problems. The conducted experiments allow us to estimate the accuracy of our algorithms and to compare them to the existing algorithms based on the support vector machine.