Presentation | 2018-05-24 A probabilistic verification method using multi-feature for environment-robust biometrics Hidetsugu Uchida, Narishige Abe, Shigefumi Yamada, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | This paper reports a probabilistic veri?cation method using multi-feature for environment-robust biometrics. In biometrics, veri?cation is performed by comparing features which are extracted to contain individuality from biometric information such as face images. However, as well as individuality, those features contain undesired information derived from environmental variations where biometric information is captured. The undesired information is mixed the feature as distortions and degrades veri?cation performance. Toward to environment-robust veri?cation, probabilistic linear discriminative analysis (PLDA)-based veri?cation has been studied. PLDA models individuality space and environment space with large learning data and can perform veri?cation referring both of twospaceinwhole. Inthispaper, PLDAisextendedtodealwithtwokindsoffeaturesextractedfromonebiometric information simultaneously. This proposed model can model individuality space and environment space with the two kinds of features. In results of experiments using face images, the proposed model is superior to conventional PLDA. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Biometrics / veri?cation score / generative model / probabilistic linear discriminative analysis / probabilistic canonical correlation analysis |
Paper # | BioX2018-3 |
Date of Issue | 2018-05-17 (BioX) |
Conference Information | |
Committee | BioX / ITE-ME / ITE-IST |
---|---|
Conference Date | 2018/5/24(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | Kanazawa Univ. Nishimachi Satelite Plaza |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | Biometrics, etc. |
Chair | Kazuhiko Sumi(AGU) / 田川 憲男(首都大) / Takayuki Hamamoto(TUS) |
Vice Chair | Hiroshi Takano(Toyama Pref. Univ.) / Hitoshi Imaoka(NEC) / / Junichi Akita(Kanazawa Univ.) |
Secretary | Hiroshi Takano(Shizuoka Univ.) / Hitoshi Imaoka(Fujitsu Labs.) / (日本工業大) / Junichi Akita |
Assistant | Masatsugu Ichino(Univ. of Electro-Comm.) / Naoyuki Takada(Secom) / Norihiro Okui(KDDI Research) / 村松 正吾(新潟大) / 広明 敏彦(NEC) / 望月 貴裕(NHK) |
Paper Information | |
Registration To | Technical Committee on Biometrics / Technical Group on Media Engineering / Technical Group on Information Sensing Technologies |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | A probabilistic verification method using multi-feature for environment-robust biometrics |
Sub Title (in English) | |
Keyword(1) | Biometrics |
Keyword(2) | veri?cation score |
Keyword(3) | generative model |
Keyword(4) | probabilistic linear discriminative analysis |
Keyword(5) | probabilistic canonical correlation analysis |
1st Author's Name | Hidetsugu Uchida |
1st Author's Affiliation | FUJITSU LABORATORIES LTD.(FUJITSU LAB.) |
2nd Author's Name | Narishige Abe |
2nd Author's Affiliation | FUJITSU LABORATORIES LTD.(FUJITSU LAB.) |
3rd Author's Name | Shigefumi Yamada |
3rd Author's Affiliation | FUJITSU LABORATORIES LTD.(FUJITSU LAB.) |
Date | 2018-05-24 |
Paper # | BioX2018-3 |
Volume (vol) | vol.118 |
Number (no) | BioX-55 |
Page | pp.pp.21-26(BioX), |
#Pages | 6 |
Date of Issue | 2018-05-17 (BioX) |