Effective immediately, postdoctoral position for 1 year
Modeling from multi-modal dynamic images : application to the vocal tract
Currently, dynamic models of the vocal tract are computed from MRI images of sustained sounds using PCA techniques. This is insufficient to handle the dynamic properties of the vocal tract.
In nancy, we have designed a multimodal acquisition system which allows us to obtain ultrasound dynamic images of the vocal tract which are registered with the MRI images.
The goal of this post-doctoral position is to build automatically a model of the tongue that makes use of both the MRI and the ultrasound data.
The challenge is to introduce highly dynamic but possibly incomplete and inaccurate features in the learning stage while roughly preserving the shape learned from MRI.
Background: PhD in computer science, electrical engineering or applied mathematics with expertise in shape modeling and computer vision. Please contact Marie-Odile Berger (marie-odile.berger@loria.fr) for inquiries. More informations on the team can be found at http://magrit.loria.fr
-- Marie-Odile BERGER - INRIA & LORIA Post: INRIA Nancy grand est, Projet Magrit - Batiment C, 615 rue du jardin botanique, 54600 Villers CEDEX, France Phone: +33 3 54 95 85 01 WWW URL: http://magrit.loria.fr
Keine Kommentare:
Kommentar veröffentlichen