Abstract | ||
---|---|---|
In this paper, a new noise front-end is proposed to improve the performance of Distributed Speech Recognition (DSR) system using a combination of conventional Mel-Cepstral Coefficients (MFCC) and Mel-Line Spectral Frequencies (MLSF). These features are adequately transformed and reduced in a multi-stream scheme using Karhunen-Loeve Transform (KLT). We investigate the performance of a new front-end DSR in terms of recognition accuracy in adverse conditions as well as in terms of dimensionality reduction. Our results showed that for highly noisy speech, using the proposed transformation scheme MLSF-KLT leads to a significant improvement in recognition accuracy on Aurora 2 task. |
Year | DOI | Venue |
---|---|---|
2011 | 10.1109/DeSE.2011.94 | Developments in E-systems Engineering |
Keywords | Field | DocType |
recognition accuracy,adverse condition,new noise front-end,proposed transformation scheme,new front-end dsr,conventional mel-cepstral coefficients,mel-line spectral frequencies,karhunen-loeve transform,mobile communications,speech recognition,multi-stream scheme,mfcc,front end,vectors,hidden markov model,karhunen loeve transform,hidden markov models,speech,acoustics,mobile communication,noise measurement,robustness | Front and back ends,Mobile radio,Mel-frequency cepstrum,Dimensionality reduction,Pattern recognition,Noise measurement,Computer science,Robustness (computer science),Speech recognition,Artificial intelligence,Hidden Markov model,Mobile telephony | Conference |
ISBN | Citations | PageRank |
978-1-4577-2186-1 | 0 | 0.34 |
References | Authors | |
4 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Addou Djamel | 1 | 7 | 1.30 |
bachir boudraa | 2 | 12 | 4.79 |