High accuracy transcription, cleaning and fast term detection for spoken documents

2012 Fiscal Year Final Research Report

• Project/Area Number: 22300059
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Single-year Grants
• Section: 一般
• Research Field: Perception information processing/Intelligent robotics
• Research Institution: Toyohashi University of Technology
• Principal Investigator: NAKAGAWA Seiichi 豊橋技術科学大学, 大学院・工学研究科, 教授 (20115893)
• Co-Investigator(Kenkyū-buntansha): AKIBA Tomoyoshi 豊橋技術科学大学, 大学院・工学研究科, 准教授 (00356346) ; YAMAMOTO Kazumasa 豊橋技術科学大学, 大学院・工学研究科, 准教授 (40324230) ; TSUCHIYA Masatoshi 豊橋技術科学大学, 情報メディア基盤センター, 助教 (70378256)
• Project Period (FY): 2010 – 2012
• Keywords: 音声情報処理 / 音声認識 / 話し言葉 / 書き言葉 / 整形 / 情報検索 / 音声ドキュメント

Research Abstract

We studied on automatic speech recognition, written style transformation from spoken style transcription and fast term detection for spoken documents. For accurately automatic speech recognition, we proposed a novel speech recognition model beyond the conventional HMM, and showed the effectiveness. For cleaning speech recognition results, we proposed a probabilistic model of cleaning from spoken style with recognition errors to written style and showed the effectiveness. Finally, we proposed a fast term detection method based on tri-gram indexes from the transcription of spoken documents and showed the effectiveness.

Research Products

• [Journal Article] 距離つき音節n-gram 索引による音声検索語検出の距離尺度の厳密化 (2013)
  • Author(s): 岩見圭祐,坂本渚,中川聖一
  • Journal Title: 情報処理学会論文誌 Volume: Vol.54, No.2 Pages: 495-505
  • Peer Reviewed
• [Journal Article] 音声中の検索語検出のためのテストコレクションの構築と分析 (2013)
  • Author(s): 伊藤慶明,西崎博光,中川聖一,秋葉友良,河原達也,胡新輝,南篠浩光,松井知子,山下洋一,相川清明
  • Journal Title: 情報処理学会論文誌 Volume: Vol.54, No.2 Pages: 471-483
  • Peer Reviewed
• [Journal Article] 音声検索語検出を前処理に用いた未知語や認識誤りに頑健な音声ドキュメント検索 (2013)
  • Author(s): 瀧上智子, 秋葉 友良
  • Journal Title: 情報処理学会論文誌 Volume: Vol.54, No.2 Pages: 506-517
  • Peer Reviewed
• [Journal Article] 音節継続時間を利用した直線検出に基づく音声検索語検出 (2013)
  • Author(s): 大野哲平, 秋葉 友良
  • Journal Title: 情報処理学会論文誌 Volume: Vol.54, No.2 Pages: 484-494
  • Peer Reviewed
• [Journal Article] A robust/fast spoken term detection method based on a syllable n-gram index with a distance metric (2012)
  • Author(s): S. Nakagawa, K. Iwami, Y. Fujii, K.Yamamoto
  • Journal Title: Speech Communication Volume: Vol.35 Pages: 470-485
  • DOI: DOI:10.1016/j.specom.2012.12.001
  • Peer Reviewed
• [Journal Article] Evaluation Framework Design of Spoken Term Detection Study at the NTCIR-9 IR for Spoken Documents Task (2012)
  • Author(s): H. Nishizaki, T. Akiba, K.Aikawa, T.Kawahara and T. Matsui
  • Journal Title: Journal of Natural Language Processing Volume: Vol.19, No.4 Pages: 330-350
  • DOI: DOI:10.5715/jnlp.19.329
  • Peer Reviewed
• [Journal Article] Risk-based semi-supervised discriminative language modeling for broadcast transcription (2012)
  • Author(s): A. Kobayashi, T. Oku, T. Imai, S.Nakagawa
  • Journal Title: IEICE Trans. Inf. & Syst Volume: Vol.E95-D,No.11 Pages: 2674-2681
  • DOI: DOI:10.1587/transinf.E95.D.2674
  • Peer Reviewed
• [Journal Article] Class-based n-gram language Model for new words using out-of-vocabulary to in-vocabulary similarity (2012)
  • Author(s): W. Naptali, M. Tsuchiya, S. Nakagawa
  • Journal Title: IEICE Trans. Inf. & Syst Volume: Vol.E95-D,No.9 Pages: 2308-2316
  • DOI: DOI:10.1587/transinf.E95.D.2308
  • Peer Reviewed
• [Journal Article] Hidden conditional neural fields for continuous phoneme recognition (2012)
  • Author(s): Y. Fujii, K. Yamamoto, S. Nakagawa
  • Journal Title: IEICE Trans. Inf. & Syst. Volume: Vol.E95-D, No.8 Pages: 2094-2104
  • DOI: DOI:10.1587/transinf.E95.D.2094
  • Peer Reviewed
• [Journal Article] Improving the readability of ASR results for lectures using multiple hypotheses and sentence-level knowledge (2012)
  • Author(s): Y. Fujii, K. Yamamoto, S. Nakagawa
  • Journal Title: IEICE Trans. Information and Systems Volume: Vol.E-95, No.4 Pages: 1101-1111
  • DOI: DOI:10.1587/transinf.E95.D.1101
  • Peer Reviewed
• [Journal Article] Topic-dependent-class-based n-gram language model (2012)
  • Author(s): W. Naptali, M. Tsuchiya, S. Nakagawa
  • Journal Title: IEEE Trans. Audio,Speech and Language Processing Volume: Vol.20, No.5 Pages: 1513-1525
  • DOI: DOI:10.1109/TASL.2012.2183870
  • Peer Reviewed
• [Journal Article] ポーズを考慮した話し言葉言語モデルの構築 (2012)
  • Author(s): 太田健吾,土屋雅稔,中川聖一
  • Journal Title: 情報処理学会論文誌 Volume: Vol.53, No.2 Pages: 889-900
  • Peer Reviewed
• [Journal Article] 部分距離空間上の索引付けに基づく音声中の高速検索語検出法 (2012)
  • Author(s): 金子 泰輔、秋葉 友良
  • Journal Title: 電子情報通信学会論文誌 Volume: Vol.J95-D, No.3 Pages: 608-617
  • Peer Reviewed
• [Journal Article] Topic-dependent language model with voting on noun history (2010)
  • Author(s): W. Naptali, M. Tsuchiya, S. Nakagawa
  • Journal Title: ACM Trans.Asian Language Information Processing Volume: Vol.9, No.2 Pages: 7:1-7:31
  • DOI: DOI:10.1145/1781134.1781137
  • Peer Reviewed
• [Presentation] Incorporating Syllable Duration into Line-Detection-Based Spoken Term Detection (2012)
  • Author(s): T. Ohno and T. Akiba
  • Organizer: 2012 IEEE Workshop on Spoken Language Technology, Paper No.TU-AM.8
  • Place of Presentation: Miami, USA
  • Year and Date: 2012-12-04
• [Presentation] Developing partially-transcribed speech corpus from edited transcriptions (2012)
  • Author(s): K. Ohta, M. Tsuchiya, S. Nakagawa
  • Organizer: Proc. LREC
  • Place of Presentation: ISTANBUL,TURKEY
  • Year and Date: 2012-05-25
• [Presentation] Deep-hidden conditional neural fields for continuous phoneme speech recognition (2012)
  • Author(s): Y. Fujii, K. Yamamoto, S. Nakagawa
  • Organizer: Proc. Int. Workshop on Statistical Machine Learning for Speech Recognition
  • Place of Presentation: 国立京都国際会館, 京都市
  • Year and Date: 2012-03-31
• [Presentation] Detection of precisely transcribed parts from in exact transcribed corpus (2011)
  • Author(s): K. Ohta, M. Tsuchiya, S. Nakagawa
  • Organizer: Proc. ASRU
  • Place of Presentation: Hawaii, USA
  • Year and Date: 2011-12-15
• [Presentation] Hidden boosted MMI and hierarchical state posterior feature for automatic speech recognition based on hidden conditional neural fields (2011)
  • Author(s): Y. Fujii, K. Yamamoto, S. Nakagawa
  • Organizer: Proc.Interspeech
  • Place of Presentation: Florence, Italy
  • Year and Date: 2011-08-31
• [Presentation] Efficient out-of-vocabulary term detection by N-gram array in deices with distance from a syllable lattices (2011)
  • Author(s): K. Iwami, Y. Fujii, K. Yamamoto, S. Nakagawa
  • Organizer: Proc.ICASSP
  • Place of Presentation: Prague, Czech Republic
  • Year and Date: 2011-05-22
• [Presentation] Automatic speech recognition using hidden conditional neural fields (2011)
  • Author(s): Y. Fujii, K. Yamamoto, S. Nakagawa
  • Organizer: Proc.ICASSP
  • Place of Presentation: Prague, Czech Republic
  • Year and Date: 2011-05-22
• [Presentation] Large vocabulary speech recognition system:SPOJUS++ (2011)
  • Author(s): Y. Fujii, K. Yamamoto, S. nakagawa
  • Organizer: Proc. 11th Wseas International Conf. MUSP-11
  • Place of Presentation: Venice, Italy
  • Year and Date: 2011-03-07
• [Presentation] Out of vocabulary term detection by N-gram array with distance from continuous syllable recognition results (2010)
  • Author(s): K. Iwami, Y. Fujii, K. Yamamoto, S.Nakagawa
  • Organizer: Proc. IEEE Workshop on SLT
  • Place of Presentation: Berkeley, California
  • Year and Date: 20101514
• [Presentation] Topic dependent class based language model evaluation on automatic speech recogniton (2010)
  • Author(s): W. Naptali, M. Tsuchiya, S. Nakagawa
  • Organizer: Proc.IEEE Workshop on SLT
  • Place of Presentation: Berkeley, California
  • Year and Date: 2010-12-15
• [Presentation] Improving the readability of class lecture ASR results using a confusion network (2010)
  • Author(s): Y. Fujii, K. Yamamoto, S. Nakagawa
  • Organizer: Proc. Interspeech
  • Place of Presentation: 幕張メッセ国際会議場, 千葉市
  • Year and Date: 2010-09-30
• [Presentation] Integration of a cache-based model and topic dependent class model with soft clustering and soft voting (2010)
  • Author(s): W. Naptali, M. Tsuchiya, S. Nakagawa
  • Organizer: Proc. Interspeech
  • Place of Presentation: 幕張メッセ国際会議場,千葉市
  • Year and Date: 2010-09-29
• [Presentation] Metric Subspace Indexing for Fast Spoken Term Detection (2010)
  • Author(s): T. Kaneko, T. Akiba
  • Organizer: Proc. Interspeech
  • Place of Presentation: 幕張メッセ国際会議場, 千葉市
  • Year and Date: 2010-09-28
• [Presentation] Constructing Japanese Test Collections for Spoken Term Detection (2010)
  • Author(s): Y. Itoh, H. Nishizaki, X. Hu, H. Nanjo,T. Akiba, T. Kawahara, S. Nakagawa, T.Matsui, Y. Yamashita, K. Aikawa
  • Organizer: Proc. Interspeech
  • Place of Presentation: 幕張メッセ国際会議場, 千葉市
  • Year and Date: 2010-09-28
• [Book] 音声言語処理と自然言語処理 (2013)
  • Author(s): 中川聖一,小林聡,峯松信明,宇津呂武仁,秋葉友良,北岡教英,山本幹雄,甲斐充彦,山本一公,土屋雅稔
  • Total Pages: 264
  • Publisher: コロナ社
• [Book] 近代科学社 (2010)
  • Author(s): 中川聖一
  • Total Pages: 242
  • Publisher: 情報理論-基礎から応用まで-
• [Remarks]
  • URL: http://www.slp.cs.tut.ac.jp