Vowel Classification from Imagined Speech Using Machine Learning

Abstract

Imagined speech is a relatively new EEG neuro-paradigm, which has seen little use in BCI applications. Imagined speech can be used to allow physically impaired patients to communicate and to use smart devices by imagining desired commands and then detecting and executing those commands in a smart device. The goal of this research is to verify previous classification attempts made and then design a new, more efficient neural network that is noticeably less complex (fewer number of layers) that still achieves a comparable classification accuracy. The classifiers are designed to distinguish between EEG signal patterns corresponding to imagined speech of different vowels and words. This research uses a dataset that consists of 15 subjects imagining saying the 5 main vowels (a, e, i, o, u) and 6 different words 2 previous researches on imagined speech classification done on this same dataset are replicated and the replication results are compared. The pre-processing of data is described and a new CNN classifier with 3 different Transfer Learning methods are described and used to classify EEG signals. Classification accuracy is used as the performance metric.