Session-Independent Array-Based EMG-to-Speech Conversion using Convolutional Neural Networks

Konferenz: Speech Communication - 13. ITG-Fachtagung Sprachkommunikation
10.10.2018 - 12.10.2018 in Oldenburg, Deutschland

Tagungsband: ITG-Fb. 282: Speech Communication

Seiten: 5Sprache: EnglischTyp: PDF

Persönliche VDE-Mitglieder erhalten auf diesen Artikel 10% Rabatt

Autoren:
Diener, Lorenz; Felsch, Gerrit; Angrick, Miguel; Schultz, Tanja (Cognitive Systems Lab, University of Bremen, Bremen, Germany)

Inhalt:
This paper presents an evaluation of the performance of EMG-to-Speech conversion based on convolutional neural networks. We present an analysis of two different architectures and network design considerations and evaluate CNN-based systems for their within-session and cross-session performance. We find that they are able to perform on par with feedforward neural networks when trained and evaluated on a single session and outperform them in cross session evaluations.