Summary: Parallelization of Structured Invariant Neural Networks for Shift and
Rotation Invariant Pattern Recognition
S. Kr¨oner 1 , M. N¨olle 2 , G. Schreiber 3
1 Technologie­Zentrum Informatik, Abteilung Bildverarbeitung, Universit¨at Bremen,
Universit¨atsallee 21­23, D­28359 Bremen, Germany, e­mail: kroener@informatik.uni­bremen.de
currently on leave to: Image Processing Lab., DEEI, University of Trieste, Via Valerio 10, I­34127 Trieste, Italy
2 Institute for Software Technology and Parallel Systems, Liechtensteinstraße 22, A­1090 Vienna, Austria
now: Institut f¨ur Informatik, Lehrstuhl f¨ur Mustererkennung und Bildverarbeitung, Universit¨at Freiburg,
Am Flughafen 17, D­79085 Freiburg i.Br., Germany, e­mail: michael.noelle@informatik.uni­freiburg.de
3 bbcom Broadband Communications, Harburger Schloßstr. 6­12, D­21079 Hamburg, Germany,
e­mail: g.schreiber@bbcom­hh.de
Abstract
In this paper is presented an algorithm for the par­
allel implementation and data distribution of struc­
tured invariant neural networks (SINN) for invari­
ant pattern recognition.
Structured invariant neural networks have shown
convincing results for the task of shift and rotation
invariant pattern recognition. They perform the in­
variant feature extraction and adaptive classification

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

Collections: Computer Technologies and Information Sciences