Robust Analytical Speech Recognition System


ROARS - 5516

Keywords speech recognition, feature-based system, noise robustness, user interfaces


Start Date: 28-DEC-90 / Duration: 36 months

[ contact / participants ]


Objectives and Approach

The goal of the ROARS project is to increase the robustness of an existing analytical speech recognition system (ie one using statistical knowledge about syllables, phonemes and phonetic features), and to use it as part of a speech understanding system with connected words and dialogue capability for two languages: French and Spanish.

The work started from an existing system implemented for the French language. This system has been shown to operate in real time, to be speaker-independent, and has had satisfactory results with continuously uttered connected words.

The aim of the first phase of the project was to develop and implement the corresponding knowledge-bases for the Spanish language and to enhance for both languages the robustness of this system against:

All these tasks are run in parallel for both languages, French and Spanish. In order to study, implement and test improvements, two identical hardware prototypes were built (one for the French application and one for the Spanish).

The aim of the second phase was the implementation of two demonstrations of speech understanding for air traffic control (one in French, one in Spanish) and will be the integration of voice input with other devices (such as keyboards, tracker-balls and screens).

Progress and Results

Two identical hardware prototypes were built for speech analysis, the statistical knowledge necessary for the feature-based approach has been established for Spanish and improved for French and, finally, the corresponding speech recognition systems were implemented.

In this way acceptable recognition rates in a multi-speaker environment were successfully demonstrated for the French recognition system and for a hundred words size vocabulary (work is still in progress for the Spanish system).

Simultaneously, a progressive and slow adaptation for vowel recognition was investigated and the Lombard effect was studied both in French and Spanish

Concerning the improvement of voice input robustness by use of dialogue and understanding, a set of tools was developed (editors, compilers,....) in the domain of "finalised dialogue", that is to say task oriented and cooperative. The implementation of the two demonstrations is in progress.

Exploitation

Thomson-CSF is applying the French ROARS system for developing visual aids for deaf people and for voice input for a SONAR console. It is investigating voice dealing and information request in PABX.

ENA Telecommunications will introduce the Spanish ROARS recognition system into its product line for SOS highway service operator control and is also investigating voice dialing in a PABX environment.

CRIN will deepen the noise effects further for both language prototypes.

Universidad Politecnica de Valencia will establish a Spanish community reference through the Spanish prototype.


CONTACT POINT

Mr Pierre Alinat
THOMSON-CSF/SINTRA-ASM
525 Route des Dolines
Parc de Sophia Antipolis, BP 157
F - 06903 SOPHIA ANTIPOLIS CEDEX
tel: + 33/ 92-964211
fax: + 33/ 92-963950
telex: THOMSI 470111 F
email: Pierre.Alinat@spl.y-net.fr

Participants

THOMSON-CSF/SINTRA-ASM - F - C
CRIN - F - P
UNIVERSIDAD POLITECNICA
DE VALENCIA - E - P
ENA TELECOMUNICACIONES SA - E - P


ST synopses home page ST acronym index ST number index
All synopses home page all acronyms index all numbers index

ROARS - 5516, December 1993


please address enquiries to the ESPRIT Information Desk

html version of synopsis by Nick Cook