Title :
Localized spectro-temporal features for noise-robust speech recognition
Author :
Kovács, Gy ; Tóth, L.
Author_Institution :
Reseach Group on Artificial Intell., Univ. of Szeged & Hungarian Acad. of Sci., Szeged, Hungary
Abstract :
In speech recognition there has been a trend to incorporate more and more knowledge about human hearing into the feature extraction step. One such approach is the application of localized spectro-temporal analysis, which is inspired by neurophysiological studies. Here we experiment with extracting features from the patches of the widely used criticial-band log-energy spectrum by applying the two-dimensional cosine transform. Compared to earlier similar studies with the spectrogram representation, we find that our method is not worse, and faster. In experiments with noisy speech the proposed representation proves more noise-robust than the conventional mel-frequency cepstral features.
Keywords :
Auditory system; Cepstral analysis; Feature extraction; Humans; Mel frequency cepstral coefficient; Noise robustness; Signal processing algorithms; Spectral analysis; Spectrogram; Speech recognition;
Conference_Titel :
Computational Cybernetics and Technical Informatics (ICCC-CONTI), 2010 International Joint Conference on
Conference_Location :
Timisoara, Romania
Print_ISBN :
978-1-4244-7432-5
DOI :
10.1109/ICCCYB.2010.5491225