Interactive image segmentation by matching attributed relational graphs

  • Authors:

  • Alexandre Noma;Ana B. V. Graciano;Roberto M. Cesar Jr;Luis A. Consularo;Isabelle Bloch

  • Affiliations:

  • Institute of Mathematics and Statistics, University of São Paulo Rua do Matão, 1010, CEP 05508-090 São Paulo, Brazil;Institute of Mathematics and Statistics, University of São Paulo Rua do Matão, 1010, CEP 05508-090 São Paulo, Brazil;Institute of Mathematics and Statistics, University of São Paulo Rua do Matão, 1010, CEP 05508-090 São Paulo, Brazil;Tribunal Superior Eleitoral, Praça dos Tribunais Superiores-Bloco C-SAS-CEP 70096-900, Brasília-DF, Brazil;Telecom ParisTech, CNRS LTCI, 46 rue Barrault, 75013 Paris, France

  • Venue:
  • Pattern Recognition
  • Year:
  • 2012

Quantified Score

Hi-index 0.01

Visualization

Abstract

A model-based graph matching approach is proposed for interactive image segmentation. It starts from an over-segmentation of the input image, exploiting color and spatial information among regions to propagate the labels from the regions marked by the user-provided seeds to the entire image. The region merging procedure is performed by matching two graphs: the input graph, representing the entire image; and the model graph, representing only the marked regions. The optimization is based on discrete search using deformed graphs to efficiently evaluate the spatial information. Note that by using a model-based approach, different interactive segmentation problems can be tackled: binary and multi-label segmentation of single images as well as of multiple similar images. Successful results for all these cases are presented, in addition to a comparison between our binary segmentation results and those obtained with state-of-the-art approaches. An implementation is available at http://structuralsegm.sourceforge.net/.