FACE (Future Adaptive Communication Environment) - SMC part

FACE is an ongoing project in CPK, involving SMC, WING, CSys and A&P groups. With the deployment of always-on communication terminals and the growing availability of information retrieval services, the challenge of this project is to provide users with adaptive and flexible access to a given service via a variety of networks via user-friendly interfaces. Speech is obviously the most natural interface for communication. Within this project Speech and Multimedia Communications is targeted towards speech processing in the presence of networks and the integration of conversational interfaces in network environments. The emphasis is on the establishment of speech technology modules distributed in networks and adapted to the qualities of involved networks, aimed at providing end-users an overall optimising of the 'Perceived' QoS. The research has been done so far are two aspects of distributed speech recognition (DSR) in the presence of channel transmission errors in wireless networks. The first is on investigating channel error protection scheme. A frame-based channel error protection scheme is used instead of the frame-pair based scheme standardised by the ETSI-DSR Group. The recognition experiments showed a significant increase in recognition accuracy for different channel error distributions, at the cost of a marginal increase in bit-rate. The second is on exploiting the knowledge about channel transmission errors for the purpose of optimising the out-of-vocabulary (OOV) detection. Transmission errors influence the acoustic likelihood, and therefore affect the optimal threshold setting for discrimination between in-vocabulary words and OOV words. An OOV-detection method is proposed in which the estimated frame-error-rate is used to adjust the discrimination threshold. This method proved successful in maintaining a constant false rejection rate across a range of error rates. The principle of exploiting information about the channel (e.g. fading channels, overloading services, congested networks and degraded acoustic environments) can be exploited to adapt the dialogue and the applied grammars and vocabularies. This will enable a graceful modification of the behaviour of a given dialogue application according to the current quality of the channel. (Zheng-Hua Tan, Paul Dalsgaard , Børge Lindberg)