Abstract
We address a one-class classification (OCC) problem aiming at detection of objects that come from a pre-defined target class. Since the non-target class is ill-defined, an effective set of features discriminating between the targets and non-targets is hard to obtain. Alternatively, when raw data are available, dissimilarity representations describing an object by its dissimilarities to a set of target examples can be used.
A complex problem can be approached by fusing information from a number of such dissimilarity representations. Therefore, we study both the combined dissimilarity representations (on which a single OCC is trained) as well as fixed and trained combiners applied to the outputs of the base OCCs, trained on each representation separately. An experiment focusing on the detection of diseased mucosa in oral cavity is conducted for this purpose. Our results show that both approaches allow for a significant improvement in performance over the best results achieved by the OCCs trained on single representations, however, concerning the computational cost, the use of combined representations might be more advantageous.
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 subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Bradley, A.P.: The use of the area under the ROC curve in the evaluation of machine learning algorithms. Pattern Recognition 30(7), 1145–1159 (1997)
Bradley, P.S., Mangasarian, O.L., Street, W.N.: Feature selection via mathematical programming. INFORMS Journal on Computing 10, 209–217 (1998)
Duin, R.P.W., Juszczak, de, P., Ridder, D., Paclík, P., Pękalska, E., Tax, D.: PR-Tools, a Matlab toolbox for pattern recognition (2004)
Esposito, F., Malerba, D., Tamma, V., Bock, H.H., Lisi, F.A.: Analysis of Symbolic Data. chapter Similarity and Dissimilarity. Springer, Heidelberg (2000)
Goldfarb, L.: A new approach to pattern recognition. In: Progress in Pattern Recognition, vol. 2, pp. 241–402. Elsevier Science Publishers B.V., Amsterdam (1985)
Jacobs, D.W., Weinshall, D., Gdalyahu, Y.: Classification with non-metric distances: Image retrieval and class representation. IEEE Trans. on Pattern Analysis and Machine Intelligence 22(6), 583–600 (2000)
Pękalska, E.: working title: Dissimilarity-based pattern recognition. PhD thesis, Delft University of Technology, The Netherlands (2004) (expected in)
Pękalska, E., Paclík, P., Duin, R.P.W.: A generalized kernel approach to dissimilarity- based classification. Journal of Machine Learning Research 2(2), 175–211 (2001)
Pękalska, E., Tax, D.M.J., Duin, R.P.W.: One-class LP classifier for dissimilarity representations. In: NIPS, pp. 761–768. MIT Press, Cambridge (2003)
Schölkopf, B., Platt, J.C., Smola, A.J., Williamson, R.C.: Estimating the support of a high-dimensional distribution. Neural Computation 13, 1443–1471 (2001)
Skurichina, M., Duin, R.P.W.: Combining different normalizations in lesion diagnostics. In: Supplementary Proc. ICANN/ICONIP, Turkey, pp. 227–230 (2003)
Stadler, B.M.R., Stadler, P.F., Wagner, G.P., Fontana, W.: The topology of the possible: Formal spaces underlying patterns of evolutionary change. Journal of Theoretical Biology 213(2), 241–274 (2001)
Tax, D.M.J.: DD-Tools, a Matlab toolbox for data description, outlier and novelty detection (2003)
Tax, D.M.J., Duin, R.P.W.: Combining one-class classifiers. In: Kittler, J., Roli, F. (eds.) MCS 2001. LNCS, vol. 2096, pp. 299–308. Springer, Heidelberg (2001)
Tax, D.M.J., Duin, R.P.W.: Support vector data description. Machine Learning 54(1), 45–56 (2002)
de Veld, D.C.G., Skurichina, M., Witjes, M.J.H., et al.: Autofluorescence characteristics of healthy oral mucosa at different anatomical sites. Lasers in Surgery and Medicine (2003) (submitted)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2004 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Pękalska, E., Skurichina, M., Duin, R.P.W. (2004). Combining Dissimilarity-Based One-Class Classifiers. In: Roli, F., Kittler, J., Windeatt, T. (eds) Multiple Classifier Systems. MCS 2004. Lecture Notes in Computer Science, vol 3077. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-25966-4_12
Download citation
DOI: https://doi.org/10.1007/978-3-540-25966-4_12
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-22144-9
Online ISBN: 978-3-540-25966-4
eBook Packages: Springer Book Archive