Title :
Static and Dynamic Spectral Features: Their Noise Robustness and Optimal Weights for ASR
Author :
Yang, Chen ; Soong, Frank K. ; Lee, Tan
Author_Institution :
Dept. of Electron. Eng., Chinese Univ. of Hong Kong, China
fDate :
March 18-23, 2005
Keywords :
cepstral analysis; hidden Markov models; learning (artificial intelligence); optimisation; speech recognition; ASR noise robustness; HMM based ASR system; SNR; additive noise; discriminative weight training; dynamic cepstral features; dynamic spectral features; exponential weight optimization; log likelihood weighting; relative word error rate reduction; speaker independent continuous digit databases; static spectral features; unequal noise robustness; Acoustic noise; Automatic speech recognition; Cepstral analysis; Decoding; Hidden Markov models; Noise robustness; Spatial databases; Speech recognition; Testing; Working environment noise;
Conference_Titel :
Acoustics, Speech, and Signal Processing, 2005. Proceedings. (ICASSP '05). IEEE International Conference on
Print_ISBN :
0-7803-8874-7
DOI :
10.1109/ICASSP.2005.1415095
