Efficient noise robust front-end based on target speech model likelihood for automatic speech recognition

• Project/Area Number: 21700188
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Single-year Grants
• Research Field: Perception information processing/Intelligent robotics
• Research Institution: Tokyo Institute of Technology
• Principal Investigator: SHINOZAKI Takahiro Tokyo Institute of Technology, 大学院・情報理工学研究科, 助教 (80447903)
• Project Period (FY): 2009 – 2010
• Project Status: Completed (Fiscal Year 2010)
• Budget Amount: ¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000); Fiscal Year 2010: ¥1,950,000 (Direct Cost: ¥1,500,000、Indirect Cost: ¥450,000); Fiscal Year 2009: ¥2,340,000 (Direct Cost: ¥1,800,000、Indirect Cost: ¥540,000)
• Keywords: 音声情報処理 / 音声認識 / 耐雑音処理 / 統計モデル / GMM / 耐雑音 / スペクトルサブトラクション

Research Abstract

To improve speech recognition performance in adverse conditions, a noise compensation method is proposed and investigated that applies a transformation in the spectral domain whose parameters are optimized based on likelihood of speech GMM modeled on the feature domain. Experimental results show that the proposed method is able to work in real-time and it is effective to reduce noise effects.

Research Products

• [Journal Article] Investigations of Features and Estimators for Speech-based Age Estimation (2010)
  • Author(s): Toshiya Wada, Takahiro Shinozaki, Sadaoki Furui
  • Journal Title: Proc.APSIPA Vol.1 Pages: 470-473
  • NAID: 120006702339
  • Peer Reviewed
• [Journal Article] Investigations of Features and Estimators for Speech-based Age Estimation (2010)
  • Author(s): Toshiya Wada, Takahiro Shinozaki, Sadaoki Furui
  • Journal Title: Proc.APSIPA Volume: 1 Pages: 470-473
  • NAID: 120006702339
  • Peer Reviewed
• [Journal Article] Target Speech GMM-based Spectral Compensation for Noise Robust Speech Recognition (2009)
  • Author(s): Takahiro Shinozaki, Sadaoki Furui
  • Journal Title: Proc.INTERSPEECH Vol.1 Pages: 1255-1258
  • NAID: 120007016511
  • Peer Reviewed
• [Journal Article] Target speech GMM-based spectral compensation for noise robust speech recognition (2009)
  • Author(s): Takahiro Shinozaki, Sadaoki Furui
  • Journal Title: Proc.INTERSPEECH 1 Pages: 1255-1258
  • NAID: 120007016511
  • Peer Reviewed
• [Presentation] 年齢推定のための音声特徴量および推定器の検討 (2010)
  • Author(s): 和田俊也, 篠崎隆宏, 古井貞煕
  • Organizer: 電子情報通信学会
  • Place of Presentation: 九州大学(福岡県春日市)
  • Year and Date: 2010-06-17
• [Presentation] 目的音GMM 尤度基準スペクトル補正法の諸評価 (2009)
  • Author(s): 篠崎隆宏, 古井貞煕
  • Organizer: 日本音響学会
  • Place of Presentation: 日本大学(福島県郡山市)
  • Year and Date: 2009-09-15
• [Presentation] 目的音GMM尤度基準スペクトル補正法の諸評価 (2009)
  • Author(s): 篠崎隆宏, 古井貞煕
  • Organizer: 日本音響学会
  • Place of Presentation: 日本大学(福島県郡山市)
  • Year and Date: 2009-09-15
• [Remarks] ホームページ等
  • URL: http://www.furui.cs.titech.ac.jp/~shinot/
• [Remarks]
  • URL: http://www.furui.cs.titech.ac.jp/~shinot/