Summary: Invariant Feature Histograms for Texture Classification
S. Siggelkow, H. Burkhardt
Institut f¨ ur Informatik, Albert­Ludwigs­Universit¨ at Freiburg
79085 Freiburg i. Br., Germany
sven.siggelkow@informatik.uni­freiburg.de \Lambda
Abstract. In this paper nonlinear invariant
feature histograms are introduced. By inte­
grating nonlinear functions over the group of
Euclidean motion we extract features that are
invariant with respect to translation and rota­
tion and from which a unique representation
can be formed. One can show, that the inte­
gration can be split into two parts: First for
every pixel of the image a nonlinear local func­
tion is evaluated, and then these results are
averaged. Instead of the second step we cal­
culate a fuzzy histogram of the local computa­
tions which preserves the invariance property
and is more robust with respect to texture
variations or defects. Furthermore in a mul­

Source: Albert-Ludwigs-Universität Freiburg, Institut für Informatik,, Lehrstuhls für Mustererkennung und Bildverarbeitung

Collections: Computer Technologies and Information Sciences