Abstract
In this paper, we consider a multi-class classifier for problems where unknown classes are included in testing phase. Previous classifiers consider the “closed-set” case where the classes used for training and the classes used for testing are the same. A more realistic case is the “open-set” recognition, in which only a limited number of classes appear at training time, and unknown classes appear during testing. To handle such problems, we need classifiers that accurately classify data belonging to not only known classes but also unknown classes. In this paper, We introduce a Support Vector Machine (SVM) based Extreme Value Machine (EVM) to determine a compact class region. Any data outside of such class regions is rejected as being in unknown classes. To construct a class region, we approach the class decision boundary found by SVM towards the samples, by removing some support vectors close to the boundary. This SVM based EVM resolves the three problems that EVM possesses: unfair size of class regions, excessive sensibility to certain points and fragmentation of a class region.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsReferences
Bartlett, P.L., Wegkamp, M.H.: Classification with a reject option using a hinge loss. J. Mach. Learn. Res. 9(59), 1823–1840 (2008)
Bishop, C.M.: Pattern Recognition and Machine Learning (Information Science and Statistics). Springer, Heidelberg (2006)
Chow, C.: On optimum recognition error and reject tradeoff. IEEE Inf. Theory 16(1), 41–46 (1970)
Fischer, L., Hammer, B., Wersing, H.: Optimal local rejection for classifiers. Neurocomputing 214, 445–457 (2016)
Frey, P.W.: Letter recognition using Holland-style adaptive classifiers. Mach. Learn. 6, 161–182 (1991). https://doi.org/10.1007/BF00114162
Kotz, S., Nadarajah, S.: Extreme Value Distributions: Theory and Applications. Imperial College Press, London (2000)
Mendes Júnior, P.R., et al.: Nearest neighbors distance ratio open-set classifier. Mach. Learn. 106(3), 359–386 (2017)
Rudd, E.M., Jain, L.P., Scheirer, W.J., Boult, T.E.: The extreme value machine. IEEE Pattern Anal. Mach. Intell. 40(3), 762–768 (2018)
Scheirer, W.J., Jain, L.P., Boult, T.E.: Probability models for open set recognition. IEEE Pattern Anal. Mach. Intell. 36(11), 2317–2324 (2014)
Scheirer, W.J., de Rezende Rocha, A., Sapkota, A., Boult, T.E.: Toward open set recognition. IEEE Pattern Anal. Mach. Intell. 35(7), 1757–1772 (2012)
Schölkopf, B., Platt, J.C., Shawe-Taylor, J.C., Smola, A.J., Williamson, R.C.: Estimating the support of a high-dimensional distribution. Neural Comput. 13(7), 1443–1471 (2001)
Sokolova, M., Lapalme, G.: A systematic analysis of performance measures for classification tasks. Inf. Process. Manag. 45(4), 427–437 (2009)
Tax, D., Duin, R.: Growing a multi-class classifier with a reject option. Pattern Recogn. Lett. 29(10), 1565–1570 (2008)
Virtanen, P., Gommers, R., et al.: SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat. Methods 17, 261–272 (2020)
Acknowledgment
This work was partially supported by JSPS KAKENHI Grant Number 19H04128.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 Springer Nature Switzerland AG
About this paper
Cite this paper
Kaneko, Y., Kudo, M. (2021). SVM Based EVM for Open Space Problems. In: Hernández Heredia, Y., Milián Núñez, V., Ruiz Shulcloper, J. (eds) Progress in Artificial Intelligence and Pattern Recognition. IWAIPR 2021. Lecture Notes in Computer Science(), vol 13055. Springer, Cham. https://doi.org/10.1007/978-3-030-89691-1_24
Download citation
DOI: https://doi.org/10.1007/978-3-030-89691-1_24
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-89690-4
Online ISBN: 978-3-030-89691-1
eBook Packages: Computer ScienceComputer Science (R0)