PARAMETERIZED KERNELS FOR SUPPORT VECTOR MACHINE CLASSIFICATION

Fernando De la Torre, Oriol Vinyals

2007

Abstract

Kernel machines (e.g. SVM, KLDA) have shown state-of-the-art performance in several visual classification tasks. The classification performance of kernel machines greatly depends on the choice of kernels and its parameters. In this paper, we propose a method to search over the space of parameterized kernels using a gradient-based method. Our method effectively learns a non-linear representation of the data useful for classification and simultaneously performs dimensionality reduction. In addition, we introduce a new matrix formulation that simplifies and unifies previous approaches. The effectiveness and robustness of the proposed algorithm is demonstrated in both synthetic and real examples of pedestrian and mouth detection in images.

Paper Citation

in Harvard Style

De la Torre F. and Vinyals O. (2007). PARAMETERIZED KERNELS FOR SUPPORT VECTOR MACHINE CLASSIFICATION . In Proceedings of the Second International Conference on Computer Vision Theory and Applications - Volume 2: VISAPP, ISBN 978-972-8865-74-0, pages 116-121. DOI: 10.5220/0002049401160121

in Bibtex Style

@conference{visapp07,
author={Fernando De la Torre and Oriol Vinyals},
title={PARAMETERIZED KERNELS FOR SUPPORT VECTOR MACHINE CLASSIFICATION},
booktitle={Proceedings of the Second International Conference on Computer Vision Theory and Applications - Volume 2: VISAPP,},
year={2007},
pages={116-121},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0002049401160121},
isbn={978-972-8865-74-0},
}

in EndNote Style

TY - CONF
JO - Proceedings of the Second International Conference on Computer Vision Theory and Applications - Volume 2: VISAPP,
TI - PARAMETERIZED KERNELS FOR SUPPORT VECTOR MACHINE CLASSIFICATION
SN - 978-972-8865-74-0
AU - De la Torre F.
AU - Vinyals O.
PY - 2007
SP - 116
EP - 121
DO - 10.5220/0002049401160121