Section: Contracts and Grants with Industry
ASPI–IST FET STREP
The ASPI (Audiovisual to Articulatory Speech Inversion) project is funded by the European Commission in the framework of the 6th PCRD. The ASPI project, started on November 2005, aims at recovering the vocal-tract shape (from vocal folds to lips) dynamics from the acoustical speech signal, supplemented by image analysis of the speaker's face. The partners of this project are KTH (Stockholm), ULB (Brussels), ENST LTCI (Paris), and NTUA-ICCS (Athens). Together with the Magrit project we are involved in this project.
The final review of the project hold successfully in January 2009.
The main contributions of Parole were about inversion algorithm, especially inversion from standard spectral data (MFCC for instance), the inversion of fricatives, the incorporation of constraints and the development of software to analyze and exploit X-ray moving pictures.
CMCU - Tunis University
This cooperation involves the LSTS (Laboratoire des systèmes et Traitement du Signal) of Tunis University headed by Prof. Noureddine Ellouze and Kais Ouni. This new project involves the investigation of automatic formant tracking, the modelling of peripheral auditory system and more generally speech analysis and parameterization that could be exploited in automatic speech recognition.
The Oesovox Project 2009-2011: 4 international groups associated...
It is possible for laryngectomees to learn a substitution voice: the esophageal voice. This voice is far from being natural. It is characterized by a weak intensity, a background noise that bothers listening, and a low pitch frequency. A device that would convert an esophageal voice to a natural voice would be very useful for laryngectomees because it would be possible for them to communicate more easily. Such natural voice restitution techniques would ideally be implemented in a portable device. In order to answer the INRIA Euromed 3+3 Mediterranean 2006 call, the INRIA Parole group (Joseph Di Martino, LORIA senior researcher, Laurent Pierron, INRIA engineer and Pierre Tricot, Associated Professor at INPL-ENSEM) associated with the following partners:
Spain: Begoña Garcia Zapirain, Deusto University (Bilbao-Spain), Telecommunication Department, PAS-"ESOIMPROVE" research group.
Tunisia: Sofia Ben Jebara, TECHTRA research group, SUP'COM, Tunis.
Morocco: El Hassane Ibn-Elhaj, SIGNAL research group, INPT, Rabat.
This project named LARYNX has been subsidized by the INRIA Euromed program during the years 2006-2008. Our results have been presented during the INRIA 2008 Euromed colloquium (Sophia Antipolis, 9-10 October 2008). During this international meeting, The French INRIA institute decided to renew our project with the new name "OESOVOX". This new project will be subsidized during the years 2009-2011.
In the framework of the European COADVISE-FP7 program, two PhD students have assigned to the Euromed 3+3 Oesovox project. These students are, Miss Fadoua Bahja from INPT-Rabat (Morocco) whose PhD thesis title is "Detection of F0 in real-time for audio: application to pathological voices" and Mr. Ammar Werghi from SUP'COM-Tunis (Tunisia) whose PhD thesis title is "Voice conversion techniques applied to pathological voice repair". The activity reports of these two students for the year 2009 is described in 6.1.6 .