Abstract | ||
---|---|---|
In this paper, we present an effective feature normalization algorithm to improve the robustness of automatic speech recognition systems. At front-end, minimum mean square error log-spectral amplitude estimation speech enhancement is adopted to suppress noise from noisy speech. Then, at back-end, the histogram equalization feature normalization is used to deal with the residual mismatch between enhanced speech and clean speech. We have evaluated recognition performance under noisy environments using NOISEX-92 database and recorded speech signals in continuous speech recognition task. Experimental results show that our approach exhibits considerable improvements in the degraded environment. |
Year | DOI | Venue |
---|---|---|
2009 | 10.1109/JCAI.2009.208 | JCAI |
Keywords | DocType | Citations |
robust feature normalization algorithm,histogram equalization feature normalization,noisy speech,automatic speech recognition,effective feature normalization algorithm,recorded speech signal,continuous speech recognition task,log-spectral amplitude estimation speech,recognition performance,enhanced speech,automatic speech recognition system,clean speech,histograms,noise measurement,noise,histogram equalization,speech recognition,robustness,front end,speech,minimum mean square error | Conference | 0 |
PageRank | References | Authors |
0.34 | 9 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Jianjun Lei | 1 | 713 | 52.69 |
Zhen Yang | 2 | 13 | 6.68 |
Jian Wang | 3 | 302 | 48.27 |