Speech Recognition under Noisy Environment

• Nita Tsuneo

  Multimedia Engineering Laboratory,Toshiba Corp.

• Minami Shigenobu

  Information & Communication Engineering Laboratory,Toshiba Corp.

• Nakayama Akira

  Toshiba Compute Engineering Corp.

• Onogi Tomohiro

  Toshiba Compute Engineering Corp.

Bibliographic Information

Other Title

• 高騒音下における音声認識方式の検討

Abstract

Because of the spectral-pattern deformetion Caused by low SN- ratio and Lombard-effect,precise recognition of speech under noisy environments is one the most difficult problem.This paper presents a solution for this problem by using a combination of Adaptive- Noise-Canceller(ANC)and Differential-Orthogonalization-Filter(DOF) method. Experimintal results show that the proposed method can achieve accurate recognition scores for noisy-Lombard speech with low SN ratio of less than O dB.

Journal

• IEICE technical report. Speech

  IEICE technical report. Speech 94 45-52, 1994

  The Institute of Electronics, Information and Communication Engineers

Keywords

• Speech Recognition
• Noisy Environment
• Adaptive Filter
• Noise Canceller
• Lonbard Effect
• Subspace Method

Details

• CRID: 1572261552246727040
• NII Article ID: 110003297209
• NII Book ID: AN10013221
• Text Lang: ja
• Data Source:
  • CiNii Articles