Hidden Markov model with noise-adaptive codebook for noisy speech recognition

Author

Wang Shuying ; Wu Shanpei

Author_Institution

Beijing Univ. of Posts & Telecommun., China

Volume

3

fYear

1993

fDate

19-21 Oct. 1993

Firstpage

325

Abstract

At present, many researchers turn their attention to automatic speech recognition in noisy environments. The main reason is that speech recognizers trained in quiet conditions but operated in a noisy environment usually have poor performance. We use discrete the hidden Markov model with noise-adaptive codebook for noisy speech recognition. The goal is to improve the recognition accuracy of recognizer in a noisy environment. When testing with noise-contaminated utterances at an SNR of 20 dB, the system has a recognition accuracy of 35%, by using the noise-adaptive codebook, the system has an accuracy of up to 90%.<>

Keywords

hidden Markov models; noise; speech coding; speech recognition; 20 dB; SNR; automatic speech recognition; hidden Markov model; noise environment; noise-adaptive codebook; noise-contaminated utterances; noisy speech recognition; recognition accuracy; Acoustic noise; Acoustic testing; Autocorrelation; Automatic speech recognition; Degradation; Hidden Markov models; Speech enhancement; Speech recognition; Vectors; Working environment noise;

fLanguage

English

Publisher

ieee

Conference_Titel

TENCON '93. Proceedings. Computer, Communication, Control and Power Engineering.1993 IEEE Region 10 Conference on

Conference_Location

Beijing, China

Print_ISBN

0-7803-1233-3

Type

conf

DOI

10.1109/TENCON.1993.327988

Filename

327988