Presentation | 2007/7/19 On Noise Robustness of Feature Expressing Temporal Variation for Word Speech Recognition Tetsuo FUNADA, Megumi UMENO, Hideyuki NOMURA, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | In previous paper, we proposed a feature FTTSS (Fourier Transform of Ternarized Spectral Slope) based on power spectral derivatives with regard to frequency to develop a robust word recognition system under noisy environments. Generally, word recognition using HMM is improved by adding features that express temporal variations, such as ΔMFCC or ΔFTTSS, because HMM can deal with only piecewise stationary signals. Actually, we have examined effectiveness of using ΔFTTSS in word recognition. In computing Δ-quantity, average temporal slope of the parameter in several frames is used after transforming spectral pattern to spectral parameter. On the other hand, considering the frequency analysis in cochlear and existence of neurons fired by detecting peak frequency of spectrum going up or down in auditory inferior colliculus, it is supposed that features showing raw temporal variations of power spectrum are effective in speech recognition. In this research, we propose a new feature FTTTS (Fourier Transform of Ternarized Temporal Slope) instead of ΔFTTSS. The FTTTS is defined by Fourier Transform along frequency of smoothed ternarized temporal variations. Experimentally, we have confirmed noise robustness of the proposed feature FTTTS as compared with ΔFTTSS. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Speech Recognition / Noise Robustness / Temporal Feature |
Paper # | SP2007-33 |
Date of Issue |
Conference Information | |
Committee | SP |
---|---|
Conference Date | 2007/7/19(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | |
Vice Chair | |
Secretary | |
Assistant |
Paper Information | |
Registration To | Speech (SP) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | On Noise Robustness of Feature Expressing Temporal Variation for Word Speech Recognition |
Sub Title (in English) | |
Keyword(1) | Speech Recognition |
Keyword(2) | Noise Robustness |
Keyword(3) | Temporal Feature |
1st Author's Name | Tetsuo FUNADA |
1st Author's Affiliation | Graduate School of Natural Science and Technology, Kanazawa University() |
2nd Author's Name | Megumi UMENO |
2nd Author's Affiliation | Graduate School of Natural Science and Technology, Kanazawa University |
3rd Author's Name | Hideyuki NOMURA |
3rd Author's Affiliation | Graduate School of Natural Science and Technology, Kanazawa University |
Date | 2007/7/19 |
Paper # | SP2007-33 |
Volume (vol) | vol.107 |
Number (no) | 165 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |