| Presentation | 2014-01-28 Study on speech recognition performance prediction using unsupervised measurement for speech quality evaluation based on higher-order statistics Yuka HIRANO, Ryoichi MIYAZAKI, Hiroshi SARUWATARI, Satoshi NAKAMURA, Tomoya TAKATANI, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | In this paper, a speech kurtosis ratio is introduced as a new unsupervised measurement of speech distortion in speech recognition performance prediction. The speech kurtosis ratio as a explanatory variable of speech recognition performance prediction needs to estimate the speech kurtosis of before and after noise reduction processing. Before noise reduction processing, we have proposed the speech kurtosis estimation using the noise signal blindly estimated by ICA. However, speech kurtosis estimation after noise reduction processing has not been discussed so far. In this paper, we address speech kurtosis estimation after noise reduction processing and evaluate speech recognition performance prediction by the proposed method. The results of the experiments using the proposed method show that the speech kurtosis ratio can be used as a explanatory variable of speech recognition performance prediction. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | higher-order statistics / speech recognition performance / speech kurtosis |
| Paper # | EA2013-119 |
| Date of Issue |
| Conference Information | |
| Committee | EA |
|---|---|
| Conference Date | 2014/1/20(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 | Engineering Acoustics (EA) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Study on speech recognition performance prediction using unsupervised measurement for speech quality evaluation based on higher-order statistics |
| Sub Title (in English) | |
| Keyword(1) | higher-order statistics |
| Keyword(2) | speech recognition performance |
| Keyword(3) | speech kurtosis |
| 1st Author's Name | Yuka HIRANO |
| 1st Author's Affiliation | Nara Institute of Science and Technology() |
| 2nd Author's Name | Ryoichi MIYAZAKI |
| 2nd Author's Affiliation | Nara Institute of Science and Technology |
| 3rd Author's Name | Hiroshi SARUWATARI |
| 3rd Author's Affiliation | Nara Institute of Science and Technology |
| 4th Author's Name | Satoshi NAKAMURA |
| 4th Author's Affiliation | Nara Institute of Science and Technology |
| 5th Author's Name | Tomoya TAKATANI |
| 5th Author's Affiliation | Toyota Motor Corporation |
| Date | 2014-01-28 |
| Paper # | EA2013-119 |
| Volume (vol) | vol.113 |
| Number (no) | 413 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |