Pattern Recognition, International Conference on (2006)

Hong Kong

Aug. 20, 2006 to Aug. 24, 2006

ISSN: 1051-4651

ISBN: 0-7695-2521-0

pp: 666-669

DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/ICPR.2006.947

HongFang Wang , University of York, York, UK

Edwin R. Hancock , University of York, York, UK

ABSTRACT

In this paper a new formulation of probabilistic relaxation labeling is developed using the theory of diffusion processes on graphs. According to this picture, the label probabilities are given by the state-vector of a continuous time random walk on a support graph. The state-vector is the solution of the heat equation on the support-graph. The nodes of the support graph are the Cartesian product of the object-set and label-set of the relaxation process. The compatibility functions are combined in the weight matrix of the support graph. The solution of the heat-equation is found by exponentiating the eigensystem of the Laplacian matrix for the weighted support graph with time. We demonstrate the new relaxation process on a feature correspondence matching problem abstracted in terms of relational graphs.

INDEX TERMS

CITATION

H. Wang and E. R. Hancock, "Probabilistic Relaxation using the Heat Equation,"

*2006 18th International Conference on Pattern Recognition(ICPR)*, Hong Kong, 2006, pp. 666-669.

doi:10.1109/ICPR.2006.947

CITATIONS