Invention Grant
US07729909B2 Block-diagonal covariance joint subspace tying and model compensation for noise robust automatic speech recognition
有权
块对角协方差联合子空间绑定和噪声鲁棒自动语音识别的模型补偿
- Patent Title: Block-diagonal covariance joint subspace tying and model compensation for noise robust automatic speech recognition
- Patent Title (中): 块对角协方差联合子空间绑定和噪声鲁棒自动语音识别的模型补偿
-
Application No.: US11369938Application Date: 2006-03-06
-
Publication No.: US07729909B2Publication Date: 2010-06-01
- Inventor: Luca Rigazio , David Kryze , Keiko Morii , Nobuyuki Kunieda , Jean-Claude Junqua
- Applicant: Luca Rigazio , David Kryze , Keiko Morii , Nobuyuki Kunieda , Jean-Claude Junqua
- Applicant Address: JP Osaka
- Assignee: Panasonic Corporation
- Current Assignee: Panasonic Corporation
- Current Assignee Address: JP Osaka
- Agency: Harness, Dickey & Pierce, P.L.C.
- Main IPC: G10L15/20
- IPC: G10L15/20 ; G10L15/06
Abstract:
Model compression is combined with model compensation. Model compression is needed in embedded ASR to reduce the size and the computational complexity of compressed models. Model-compensation is used to adapt in real-time to changing noise environments. The present invention allows for the design of smaller ASR engines (memory consumption reduced to up to one-sixth) with reduced impact on recognition accuracy and/or robustness to noises.
Public/Granted literature
Information query
