Parameterized Kernels for Support Vector Machine Classification

Fernando De la Torre Frade and Oriol Vinyals
International Conference on Computer Vision Theory and Applications, , March, 2007, pp. 207-213.


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.

Keywords
support vector machine, kernels, classification

Notes
Associated Project(s): Component Analysis for Data Analysis
Note: associated project Component Analysis for Data Analysis

Text Reference
Fernando De la Torre Frade and Oriol Vinyals, "Parameterized Kernels for Support Vector Machine Classification," International Conference on Computer Vision Theory and Applications, , March, 2007, pp. 207-213.

BibTeX Reference
@article{De_la_Torre_Frade_2007_5848,
   author = "Fernando {De la Torre Frade} and Oriol Vinyals",
   title = "Parameterized Kernels for Support Vector Machine Classification",
   journal = "International Conference on Computer Vision Theory and Applications",
   pages = "207-213",
   month = "March",
   year = "2007",
   Notes = "associated project Component Analysis for Data Analysis"
}