Improving Core Speech Recognition Technology - CORETEX

Despite the progress in speech technologies over the last decade and the emergence of some commercially available voice recognition products, there are many outstanding research issues that still need to be addressed. The aim of CORETEX is to improve core speech recognition technologies.

The motivation behind the work is that although current commercial speech recognition systems work reasonably well for very specific tasks and languages, they are not easily adaptable to new domains, different acoustic environments and other languages. Porting a speech application to another language or environment still requires a substantial investment of time and money, and requires the knowledge of speech recognition experts.

OBJECTIVES

The objective of the CORETEX project is to develop generic speech recognition technologies that work well for a wide range of tasks and which is essentially 'application neutral'.

The specific objectives of the CORETEX can be summarised as:

• to develop generic speech recognition technology for a wide range of tasks with minimum domain dependancies
• to develop methods for rapid porting to new domains and languages with limited training data
• to investigate techniques for producing enriched symbolic speech transcription for higher level (symbolic) processing
• to explore methods to improve language models and for automatic pronunciation generation for vocabulary extension
• to integrate the methods into showcases in order to validate them in relevant applications
• to establish an evaluation framework, collect test suites, and define objective measures to assess improvements
• to disseminate information about the research carried out in CORETEX and to facilitate contact with interested user groups so as to widely exploit the results.

Go to the CORETEX web site

Project factsheet optimized for printing

Last updated: 11.12.2002

[  Projects Home  |  Alphabetical List  |  Programme List  |  Group List  |  Search Projects  ]