Summary: Proceedings of the IASTED International Conference
Signal and Image Processing
October 27­31, 1998, Las Vegas, Nevada ­ USA
IMAGE RETRIEVAL BASED ON LOCAL INVARIANT FEATURES
SVEN SIGGELKOW, HANS BURKHARDT
Institute of Computer Science
Albert­Ludwigs­Universit˜at Freiburg
79085 Freiburg, Germany
fsiggelkowjburkhardtg@informatik.uni­freiburg.de \Lambda
Abstract
Among other techniques especially methods working
fully automatically are of interest for image retrieval from
large databases. Color histograms proved to be successful
in automatic image retrieval, however, a major drawback
is that all structural information is lost. Therefore, we ex­
tend the color histogram approach by features that take the
relations within a local pixel neighborhood into account.
By integrating nonlinear functions over the group of Eu­
clidean motion we extract features that are invariant with
respect to translation and rotation. Thus, we combine the

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

Collections: Computer Technologies and Information Sciences