Proximity graphs based multi-scale image segmentation
- Los Alamos National Laboratory
We present a novel multi-scale image segmentation approach based on irregular triangular and polygonal tessellations produced by proximity graphs. Our approach consists of two separate stages: polygonal seeds generation followed by an iterative bottom-up polygon agglomeration into larger chunks. We employ constrained Delaunay triangulation combined with the principles known from the visual perception to extract an initial ,irregular polygonal tessellation of the image. These initial polygons are built upon a triangular mesh composed of irregular sized triangles and their shapes are ad'apted to the image content. We then represent the image as a graph with vertices corresponding to the polygons and edges reflecting polygon relations. The segmentation problem is then formulated as Minimum Spanning Tree extraction. We build a successive fine-to-coarse hierarchy of irregular polygonal grids by an iterative graph contraction constructing Minimum Spanning Tree. The contraction uses local information and merges the polygons bottom-up based on local region-and edge-based characteristics.
- Research Organization:
- Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
- Sponsoring Organization:
- USDOE
- DOE Contract Number:
- AC52-06NA25396
- OSTI ID:
- 957800
- Report Number(s):
- LA-UR-08-04684; LA-UR-08-4684; TRN: US201016%%198
- Resource Relation:
- Journal Volume: 5358; Conference: Intern Symposium on Visual Computing ; December 1, 2008 ; Las Vegas
- Country of Publication:
- United States
- Language:
- English
Similar Records
Patch-based image segmentation of satellite imagery using minimum spanning tree construction
Man-made objects cuing in satellite imagery