Guided Relinking of Graph Pyramids

  • Authors:
  • Roland Glantz;Walter G. Kropatsch

  • Affiliations:
  • -;-

  • Venue:
  • Proceedings of the Joint IAPR International Workshops on Advances in Pattern Recognition
  • Year:
  • 2000

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper we propose a new method to relink graph pyramids by local relinking operations in an iterated parallel way. By representing graph pyramids as bases of valuated matroids, the goal of the relinking is expressed by a valuation on the corresponding matroid. This valuation guides the local relinking operations. The valuation attains its global maximum if none of the local relinking operations yields higher values. The new method is used for an adaption of graph pyramids towards having a given receptive field.