Find the most recent information on EU Funding activities in the field of Information and Communication Technologies (ICT) by visiting our ICT in FP7 website, which covers ICT in the 7th Framework Programme (FP7) 2007 - 2013.
Divines - Diagnostic and Intrinsic Variability?s in Natural Speech
DIVINES will be presented at the International Conference of Communicating European Research (CER) 2005 in Brussels, 14-15 November 2005
Divines Summary
DIVINES is concerned with more accurate and adaptable recognition of spoken language. When one considers human performances as the target, universal automatic recognition of speech is far from a solved problem. This is to be related by a large amount to feature extraction, modelling and adaptability weaknesses. Strikingly, these weaknesses remain fully in the case of clean speech in known conditions, emphasizing deficiencies in dealing with :
- speech intrinsic variability?s (speaker, gender, speech rate, vocal effort, regional accents, speaking style, non stationary...) as opposed to extrinsic variability?s (channel and noise),
- and extracting information form the signal itself, as opposed to using prior restrictive knowledge (for instance in the form of very specific grammars).
These deficiencies, together with weak adaptability, are major bottlenecks to the take-up of natural interactivity systems.
Divines Goals
The goal of DIVINES is to provide alternatives to the current state-of-the-art feature extraction, acoustic modelling and lexical modelling techniques, and to contribute to the long-term goal of bridging the gap between human and machine speech recognition performance.
DIVINES is a STREP funded under the Sixth Framework Programmes (FP6) Information Society Technologies (IST) priority (PDF, 144KB). It covers part of the multimodal interfaces strategic objective.
Further information:
Have a look at the Project website or the CORDIS Factsheet