CUED Publications database

Speech recognition using hidden Markov model decomposition and a general background speech model

Wang, MQ and Young, SJ (1992) Speech recognition using hidden Markov model decomposition and a general background speech model. In: UNSPECIFIED pp. 253-256..

Full text not available from this repository.

Abstract

© 1992 IEEE. HMM Decomposition is used for recognising speech in the presence of an interfering background speaker. The foreground speech is modelled by a set of left-to-right isolated word HMM's trained on a small isolated word database, and the background speech is modelled by a parallel ergodic HMM trained on a subset of TIMIT. The standard Output Approximation (OA) method of estimating the output probability distributions is used, and compared with a simple Model Combination technique (MC). Recent work in this area has shown the effectiveness of vocabulary specific background speech models, and hence this is used as a baseline. The results show that the general ergodic background model is as effective as a vocabulary specific model. However, the MC technique is not effective.

Item Type: Conference or Workshop Item (UNSPECIFIED)
Subjects: UNSPECIFIED
Divisions: Div F > Machine Intelligence
Depositing User: Cron Job
Date Deposited: 25 Jul 2017 03:54
Last Modified: 24 Aug 2017 01:28
DOI: