Imagined Speech Classification using EEG based Brain-Computer Interface

Imagined speech classification in Brain-Computer Interface (BCI) has acquired recognition in a variety of fields including cognitive biometric, silent speech communication, synthetic telepathy etc. The major objective of this paper is to develop an imagined speech classification system based on Electroencephalography (EEG). In the proposed framework features are extracted using discrete wavelet transform (DWT) and maximum linear cross-correlation (MaxLCor). These combined features are fed to the multiclass support vector machine (SVM). In comparison to some of the most common classification algorithms in BCI, the suggested research produces promising results in classification accuracy, indicating the great potential to create BCI-based speech prosthesis controllers for various applications.