Computational Linguistics Colloquium

Thursday, 10 January, 16:15
Conference Room, Building C7 4

Analysis, Synthesis and Evaluation of Visual Speech

Methods for the analysis of visible speech movements and various approaches for articulation modeling are presented. The results are evaluated in terms of intelligibility by comparing natural visual and audiovisual speech to re-synthesized speech movements rendered by talking heads. One talking head is created using the speaker cloning methodology of the Institut de la Communication Parlée in Grenoble (now department for speech and cognition in GIPSA-Lab) where a German speaker with colored markers on the face was recorded audiovisually using multiple cameras. A second talking head was created from scratch to visualize German speech movements driven by combined singe-view video data and EMA data. Advantages and weaknesses of both methods are presented.

The slides can be viewed here

If you would like to meet with the speaker, please contact Jürgen Trouvain.