知識を用いた音声認識による野球実況中継の構造化  [in Japanese] Structuring Baseball Live Game Base on Knowledge Dependent Speech Recognition  [in Japanese]

Access this Article

Search this Article

Author(s)

Abstract

本研究では,野球の実況中継に対して構造化を行うことを目的としている.構造化を行うために,大語彙連続音声認識を利用する.音声認識では,発音は似ているが意味に照らして考えればありえないような単語を認識することがある.このような単語に対しては,野球の知識を用いることで認識誤りを低減できる.また,状況依存の情報として,アナウンサーの感情も利用する.音声認識の結果を用いて構造化を行いながら,推定された情報を音声認識に利用する手法として「状態推定音声認識」の枠組みを提案する.実験により,構造化を行うために重要と考えられるキーワード正解精度が2.3%向上し,また,約73.3%の構造化正解率が得られた.The purpose of this study is to automatically structure sports live speech, especially baseball live speech using Large Vocabulary Continuous Speech Recognition System. Since it is a difficult problem to recognize baseball live speech. We propose in this paper a speech recognition method of incorporating the baseball game knowledge such as counting of inning, out, strike, ball and announcer's emotion. This method is formalized in the framework of probability theory and implemented in the conventional speech decoding (Viterbi) algorithm. This method enables to seek a word sequence as well as a structure sequence. The experimental results showed that the proposed approach improved the structuring and segmentation accuracy to 73.3% as well as keywords accuracy by 2.3%.

The purpose of this study is to automatically structure sports live speech, especially baseball live speech using Large Vocabulary Continuous Speech Recognition system. Since it is a difficult problem to recognize baseball live speech. We propose in this paper a speech recognition method of incorporating the baseball game knowledge such as counting of inning, out, strike, ball and announcer's emotion. This method is formalized in the framework of probability theory and implemented in the conventional speech decoding (Viterbi) algorithm. This method enables to seek a word sequence as well as a structure sequence. The experimental results showed that the proposed approach improved the structuring and segmentation accuracy to 73.3% as well as keywords accuracy by 2.3%.

Journal

  • IPSJ SIG Notes

    IPSJ SIG Notes 2004(131(2004-SLP-054)), 331-336, 2004-12-22

    Information Processing Society of Japan (IPSJ)

References:  11

Codes

  • NII Article ID (NAID)
    110002950629
  • NII NACSIS-CAT ID (NCID)
    AN10442647
  • Text Lang
    JPN
  • Article Type
    Technical Report
  • ISSN
    09196072
  • NDL Article ID
    7214605
  • NDL Source Classification
    ZM13(科学技術--科学技術一般--データ処理・計算機)
  • NDL Call No.
    Z14-1121
  • Data Source
    CJP  NDL  NII-ELS  IPSJ 
Page Top