5th International Conference on Spoken Language Processing

Sydney, Australia
November 30 - December 4, 1998

A Comparative Study Of Hybrid Modelling Techniques For Improved Telephone Speech Recognition

Rathinavelu Chengalvarayan

Lucent Technologies, USA

This paper presents a new technique for modelling heterogeneous data sources such as speech signals received via distinctly different channels which arises when an automatic speech recognition is deployed in wireless telephony in which highly heterogenous channels coexist and interoperate. The key problem is that a simple model may become inadequate to describe accurately the diversity of the signal, resulting in an unsatisfactory recognition performance. To cope up with this problem, different hybrid modelling techniques have been proposed and investigated in this paper by intelligently combining models from two different wireline and wireless environments.

Full Paper

Bibliographic reference.  Chengalvarayan, Rathinavelu (1998): "A comparative study of hybrid modelling techniques for improved telephone speech recognition", In ICSLP-1998, paper 0022.