Title | ||
---|---|---|
Feature extraction combining spectral noise reduction and cepstral histogram equalization for robust ASR |
Abstract | ||
---|---|---|
This work is mainly focused on showing experimental results us- ing a combination of two methods for noise compensation which are shown to be complementary: classical spectral subtraction al- gorithm and histogram equalization. While spectral subtraction is focused on the reduction of the additive noise in the spectral do- main, histogram equalization is applied in the cepstral domain to compensate the remaining non-linear effects associated to channel distortion and additive noise. The estimation of the noise spec- trum for the spectral subtraction method relies on a new algorithm for speech / non-speech detection (SND) based on order statis- tics. This SND classication is also used for dropping long speech pauses. Results on Aurora 2 and Aurora 3 are reported. |
Year | Venue | Keywords |
---|---|---|
2002 | INTERSPEECH | histogram equalization,noise reduction,speech detection,feature extraction |
DocType | Citations | PageRank |
Conference | 8 | 0.82 |
References | Authors | |
2 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
José C. Segura | 1 | 481 | 38.14 |
M. Carmen Benítez | 2 | 303 | 25.05 |
Ángel de la Torre | 3 | 482 | 34.91 |
Antonio J. Rubio | 4 | 466 | 36.14 |