Progress in Facial Expression based Affective Speech Translation

Zeeshan Ahmed, Ingmar Steiner, Éva Székely, Julie Carson-Berndsen

In: Petra Wagner (Hrsg.). Elektronische Sprachsignalverarbeitung 2013. Seiten 189-196 Studientexte zur Sprachkommunikation 65 ISBN 978-3-944331-03-4 TUDpress Dresden, Germany 3/2013.


Speech-to-speech translation is an emerging field for applications in spoken language technology. Translation systems currently focus on the processing of linguistic content, without taking into account the significance of paralinguistic information conveyed by visual gestures in human face-to-face communication. In the project presented in this paper, we have implemented a speech-to-speech translation system which preserves information about the user's affective state by transmitting it through the processing pipeline to the output component, which renders the translated content in the appropriate speaking style using expressive speech synthesis.

feast-essv.pdf (pdf, 261 KB )

Deutsches Forschungszentrum für Künstliche Intelligenz
German Research Center for Artificial Intelligence