Semi-supervised Co-update of Multiple Matchers
Classification algorithms based on template matching are used in many applications (e.g., face recognition). Performances of template matching classifiers are obviously affected by the representativeness of available templates. In many real applications, such representativeness can substantially decrease over the time (e.g., due to “aging” effects in biometric applications). Among algorithms which have been recently proposed to deal with such issue, the template co-update algorithm uses the mutual help of two complementary template matchers to update the templates over the time in a semi-supervised way . However, it can be shown that the template co-update algorithm derives from a more general framework which supports the use of more than two template matching classifiers. The aim of this paper is to point out this fact and propose the co-update of multiple matchers. Preliminary experimental results are shown to validate the proposed model.
KeywordsFace Recognition Template Match Biometric System Fingerprint Image Initial Template
Unable to display preview. Download preview PDF.
- 4.Ross, A., Nandakumar, K., Jain, A.K.: Handbook of Multibiometrics. Springer, Heidelberg (2006)Google Scholar
- 5.Zhu, X.: Semi-supervised learning literature survey, Technical report, Computer Sciences TR 1530, Univ. Wisconsis, Madison, USA (2006)Google Scholar
- 7.Didaci, L., Marcialis, G.L., Roli, F.: Modelling FRR of Biometric Verification Systems using the Template Co-update Algorithm. In: 3rd IAPR/IEEE Int. Conference on Biometrics ICB 2009, Alghero (Italy), June, 2-5. LNCS, Springer, Heidelberg (2009) (in press)Google Scholar
- 9.Blum, A., Mitchell, T.: Combining labeled and unlabeled data with co-training. In: Proc. of the Workshop on Computational Learning Theory, pp. 92–100 (1998)Google Scholar