Image segmentation has long been an important problem in the computer vision community. In our recent work we have addressed the problem of texture segmentation, where we combined top-down and bottom-up views of the image into a unified procedure. In this paper we extend our work by proposing a modified procedure which makes use of graphs of image regions. In the top-down procedure a quadtree of image region descriptors is obtained in which a novel affine contractive transformation based on neighboring regions is used to update descriptors and determine stable segments. In the bottom-up procedure we form a planar graph on the resulting stable segments, where edges are present between vertices representing neighboring image regions. We then use a vertex merging technique to obtain the final segmentation. We verify the effectiveness of this procedure by demonstrating results which compare well to other recent techniques.
|Title of host publication||Lecture Notes in Computer Science|
|Place of Publication||Berlin / Heidelberg|
|Publication status||Published - 2009|
|Event||7th IAPR -TC-15 Workshop on Graph-based Representations in Pattern Recognition May 26–28 2009 Venice (Italy) - |
Duration: 1 Jan 2009 → …
|Conference||7th IAPR -TC-15 Workshop on Graph-based Representations in Pattern Recognition May 26–28 2009 Venice (Italy)|
|Period||01/01/2009 → …|