Speech Synthesis using EEG
This work addresses speech synthesis for individuals with communication impairments, but it is incremental as it builds on existing EEG feature sets and methods.
The paper tackled the problem of synthesizing speech directly from EEG features, demonstrating feasibility with results from four subjects using both spoken and listening EEG data.
In this paper we demonstrate speech synthesis using different electroencephalography (EEG) feature sets recently introduced in [1]. We make use of a recurrent neural network (RNN) regression model to predict acoustic features directly from EEG features. We demonstrate our results using EEG features recorded in parallel with spoken speech as well as using EEG recorded in parallel with listening utterances. We provide EEG based speech synthesis results for four subjects in this paper and our results demonstrate the feasibility of synthesizing speech directly from EEG features.