Presentation | 2015-07-17 A study on effectiveness of pop noise for speaker verification Shiori Nakano, Ryosuke Nakanishi, Sayaka Shiota, Hitoshi Kiya, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | This paper investigates an effectiveness of pop noise, which is unconsciously caused by human breath, for automatic speaker verification (ASV). Recently, spoofing techniques against biometric systems have also progressed drastically. Experimentation using advanced speech synthesis and voice conversion techniques has showed unacceptable false acceptance rates and several new countermeasure algorithms have been explored to detect spoofing materials accurately. However, the countermeasures proposed so far are based on the acoustic differences between natural speech signals and artificial speech signals, expected to become gradually smaller in the near future. Therefore, voice liveness detection has been proposed for ASV. The aim of voice liveness detection validates whether the presented speech signals originated from a live human. To realize this framework, pop noise detection algorithms are also proposed. This paper focuses that the pop noise, which has been treated as addictive noise, is able to affect to ASV system or not. In the ASV experiments, speaker dependent (SD) model which is estimated by speech data with pop noise is compared with SD model which is estimated by speech data without pop noise. Experimental results show that pop noise is positively affecting the model accuracy when it is recorded under an adequate recording condition. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | automatic speaker verification / pop noise / UBM-GMM / liveness detection |
Paper # | SP2015-47 |
Date of Issue | 2015-07-09 (SP) |
Conference Information | |
Committee | SP / IPSJ-SLP |
---|---|
Conference Date | 2015/7/16(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | Katakura Suwako Hotel |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | Speech recognition and understanding, dialog system, etc. |
Chair | Kazunori Mano(Shibaura Inst. of Tech.) / Koichi Shinoda(Tokyo Inst. of Tech.) |
Vice Chair | Norihide Kitaoka(Tokushima Univ.) |
Secretary | Norihide Kitaoka(Tokyo City Univ.) / (Kobe Univ.) |
Assistant | Takashi Nose(Tohoku Univ.) / Taichi Asami(NTT) |
Paper Information | |
Registration To | Technical Committee on Speech / Special Interest Group on Spoken Language Processing |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | A study on effectiveness of pop noise for speaker verification |
Sub Title (in English) | |
Keyword(1) | automatic speaker verification |
Keyword(2) | pop noise |
Keyword(3) | UBM-GMM |
Keyword(4) | liveness detection |
1st Author's Name | Shiori Nakano |
1st Author's Affiliation | Tokyo Metropolitan University(Tokyo Metro Univ.) |
2nd Author's Name | Ryosuke Nakanishi |
2nd Author's Affiliation | Tokyo Metropolitan University(Tokyo Metro Univ.) |
3rd Author's Name | Sayaka Shiota |
3rd Author's Affiliation | Tokyo Metropolitan University(Tokyo Metro Univ.) |
4th Author's Name | Hitoshi Kiya |
4th Author's Affiliation | Tokyo Metropolitan University(Tokyo Metro Univ.) |
Date | 2015-07-17 |
Paper # | SP2015-47 |
Volume (vol) | vol.115 |
Number (no) | SP-146 |
Page | pp.pp.67-72(SP), |
#Pages | 6 |
Date of Issue | 2015-07-09 (SP) |