Continuous Speech Recognition Using a Grammar Based on Subtrees and Preterminal Bigrams

6 Citations 7 References

Bibliographic Information

Other Title
  • 部分木を単位とする構文規則と前終端記号のバイグラムを利用した連続音声認識

Search this article

Description

We are studying a spontaneous speech recognition method for building speech translation and/or spoken dialogue systems. It is necessary and important for such kinds of speech recognizers to be portable for some domains and/or tasks. Thus we are examining a speech parser that outputs subtree sequences based on syntactic rules and acoustic scores as one approach. First we design and develop syntactic rules that represent pause units, that mean segments separated by pause information, as subtrees. In this paper, we report experimental results of continuous speech recognition using moderated spontaneous speech in our database for research. Finally, we try to introduce preterminal bigrams in the speech recognition process. This suggests that preterminal bigrams are effective to achieve good performance.

Journal

  • IPSJ SIG Notes

    IPSJ SIG Notes 9 55-62, 1995-12-14

    Information Processing Society of Japan (IPSJ)

Citations (6)*help

See more

References(7)*help

See more

Keywords

Details 詳細情報について

  • CRID
    1573387452004356864
  • NII Article ID
    110002917013
  • NII Book ID
    AN10442647
  • ISSN
    09196072
  • Text Lang
    ja
  • Data Source
    • CiNii Articles

Report a problem

Back to top