Tangent Distance Kernels for Support Vector Machines
B. Haasdonk, and D. Keysers. Proceedings / 16th International Conference on Pattern Recognition, 2, page 864-868. Los Alamitos, Calif., IEEE Computer Society, (2002)
Abstract
When dealing with pattern recognition problems one encounters different types of a-priori knowledge. It is important to incorporate such knowledge into the classification method at hand. A very common type of a-priori knowledge is transformation invariance of the input data, e.g. geometric transformations of image-data like shifts, scaling etc. Distance based classification methods can make use of this by a modified distance measure called tangent distance 13, 14. We introduce a new class of kernels for support vector machines which incorporate tangent distance and therefore are applicable in cases where such transformation invariances are known. We report experimental results which show that the performance of our method is comparable to other state-of-the-art methods, while problems of existing ones are avoided.
%0 Conference Paper
%1 haasdonk2002tangent
%A Haasdonk, Bernard
%A Keysers, Daniel
%B Proceedings / 16th International Conference on Pattern Recognition
%C Los Alamitos, Calif.
%D 2002
%E Kasturi, Rangachar
%I IEEE Computer Society
%K fis liste ubs_30123
%P 864-868
%T Tangent Distance Kernels for Support Vector Machines
%V 2
%X When dealing with pattern recognition problems one encounters different types of a-priori knowledge. It is important to incorporate such knowledge into the classification method at hand. A very common type of a-priori knowledge is transformation invariance of the input data, e.g. geometric transformations of image-data like shifts, scaling etc. Distance based classification methods can make use of this by a modified distance measure called tangent distance 13, 14. We introduce a new class of kernels for support vector machines which incorporate tangent distance and therefore are applicable in cases where such transformation invariances are known. We report experimental results which show that the performance of our method is comparable to other state-of-the-art methods, while problems of existing ones are avoided.
@inproceedings{haasdonk2002tangent,
abstract = {When dealing with pattern recognition problems one encounters different types of a-priori knowledge. It is important to incorporate such knowledge into the classification method at hand. A very common type of a-priori knowledge is transformation invariance of the input data, e.g. geometric transformations of image-data like shifts, scaling etc. Distance based classification methods can make use of this by a modified distance measure called tangent distance [13, 14]. We introduce a new class of kernels for support vector machines which incorporate tangent distance and therefore are applicable in cases where such transformation invariances are known. We report experimental results which show that the performance of our method is comparable to other state-of-the-art methods, while problems of existing ones are avoided.},
added-at = {2023-08-21T15:14:23.000+0200},
address = {Los Alamitos, Calif.},
author = {Haasdonk, Bernard and Keysers, Daniel},
biburl = {https://puma.ub.uni-stuttgart.de/bibtex/28f8c8a215981b084dcb06be43860738a/unibiblio-4},
booktitle = {Proceedings / 16th International Conference on Pattern Recognition},
editor = {Kasturi, Rangachar},
eventdate = {2002-08-11/2002-08-15},
eventtitle = {16. International Conference on Pattern Recognition, ICPR 2002},
interhash = {5a4cee72935946baf4afe8ca0f669f30},
intrahash = {8f8c8a215981b084dcb06be43860738a},
keywords = {fis liste ubs_30123},
language = {eng},
pages = {864-868},
publisher = {IEEE Computer Society},
timestamp = {2023-09-01T10:39:54.000+0200},
title = {Tangent Distance Kernels for Support Vector Machines},
venue = {Québec},
volume = 2,
year = 2002
}