Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
Invariance in Kernel Methods by Haar-Integration Kernels
 

Summary: Invariance in Kernel Methods by
Haar-Integration Kernels
B. Haasdonk1
, A. Vossen2
, and H. Burkhardt1
1
Computer Science Department,
Albert-Ludwigs-University Freiburg, 79110 Freiburg, Germany
{haasdonk, burkhardt}@informatik.uni-freiburg.de
2
Institute of Physics, Albert-Ludwigs-University Freiburg,
79104 Freiburg, Germany
vossen@physik.uni-freiburg.de
Abstract. We address the problem of incorporating transformation in-
variance in kernels for pattern analysis with kernel methods. We intro-
duce a new class of kernels by so called Haar-integration over trans-
formations. This results in kernel functions, which are positive definite,
have adjustable invariance, can capture simultaneously various contin-
uous or discrete transformations and are applicable in various kernel
methods. We demonstrate these properties on toy examples and experi-

  

Source: Albert-Ludwigs-Universität Freiburg, Institut für Informatik,, Lehrstuhls für Mustererkennung und Bildverarbeitung
Haasdonk, Bernard - Institut für Numerische und Angewandte Mathematik, Universität Münster

 

Collections: Computer Technologies and Information Sciences; Mathematics