A Study on HMM-LR Viterbi best-first Search for Phrase recognition with Stochastic Context Free Grammar

  • Monzen Seikou
    Department of Electrical and Information Engineering,Faculty of Engineering,Yamagata University
  • Ohseki Masakazu
    Department of Electrical and Information Engineering,Faculty of Engineering,Yamagata University
  • Kohda Masaki
    Department of Electrical and Information Engineering,Faculty of Engineering,Yamagata University

Bibliographic Information

Other Title
  • 確率文脈自由文法を用いたHMM-LR文節音声認識におけるViterbi best-firstサーチの検討

Search this article

Description

The study on Viterbi search strategy is an important part of HMM based speech recognition resaerch.In previous work,best-first search method for continuous speech recognition has been formulated on one graph from the viewpoint of graph search,and then verifying the reduction in computation amount,an efficient phrase recognition could be achieved.In this paper,we will employ stochastic context free grammar for phrase speech recognition using the HMM-LR Viterbi best-first search,and will report on the effectiveness of it.

Journal

Keywords

Details 詳細情報について

  • CRID
    1574231877084121472
  • NII Article ID
    110003299788
  • NII Book ID
    AN10013232
  • Text Lang
    ja
  • Data Source
    • CiNii Articles

Report a problem

Back to top