Please use this identifier to cite or link to this item:
https://hdl.handle.net/20.500.14279/3947
Title: | Quantitative evaluation of the effects of aging on biometric templates |
Authors: | Tsapatsoulis, Nicolas Lanitis, Andreas |
Major Field of Science: | Humanities |
Field Category: | Arts |
Keywords: | Estimation;Face recognition;Age progression |
Issue Date: | Nov-2011 |
Source: | IET computer vision, 2011, vol. 5, no. 6, pp. 338-347 |
Volume: | 5 |
Issue: | 6 |
Start page: | 338 |
End page: | 347 |
Journal: | IET computer vision |
Abstract: | The long-term performance of biometric authentication systems is highly depended on the permanence of biometric features stored in biometric templates. Aging variation causes modifications on biometric features that affect the matching between stored and captured biometric templates causing in that way deterioration in the performance of biometric authentication systems. In this study the authors attempt to quantify the effects of aging for different biometric modalities, so that it is possible to draw conclusions related to the effect of aging on different types of biometric templates. In this context variations between distributions containing biometric features from different age groups are quantified, allowing in that way the definition of age-sensitive and age-invariant biometric features. An important aspect of the proposed approach is the standardised and generic nature of the approach that allows the derivation of comparative results between different modalities and different feature vectors. The work presented in this study provides a valuable tool for selecting, either age-invariant features for use in identity authentication applications, or for selecting age-sensitive features for age-estimation-related applications. |
URI: | https://hdl.handle.net/20.500.14279/3947 |
ISSN: | 17519640 |
DOI: | 10.1049/iet-cvi.2010.0197 |
Rights: | © The Institution of Engineering and Technology |
Type: | Article |
Affiliation : | Cyprus University of Technology |
Publication Type: | Peer Reviewed |
Appears in Collections: | Άρθρα/Articles |
CORE Recommender
Items in KTISIS are protected by copyright, with all rights reserved, unless otherwise indicated.