发明名称 |
Perceptual harmonic cepstral coefficients as the front-end for speech recognition |
摘要 |
Pitch estimation and classification into voiced, unvoiced and transitional speech were performed by a spectro-temporal auto-correlation technique. A peak picking formula was then employed. A weighting function was then applied to the power spectrum. The harmonics weighted power spectrum underwent mel-scaled band-pass filtering, and the log-energy of the filter's output was discrete cosine transformed to produce cepstral coefficients. A within-filter cubic-root amplitude compression was applied to reduce amplitude variation without compromise of the gain invariance properties.
|
申请公布号 |
US7337107(B2) |
申请公布日期 |
2008.02.26 |
申请号 |
US20030363523 |
申请日期 |
2003.05.19 |
申请人 |
THE REGENTS OF THE UNIVERSITY OF CALIFORNIA |
发明人 |
ROSE KENNETH;GU LIANG |
分类号 |
G10L11/06;G10L11/04;G10L15/02;G10L19/06 |
主分类号 |
G10L11/06 |
代理机构 |
|
代理人 |
|
主权项 |
|
地址 |
|