On the choice of the first level on graph pyramids
| Autor: | Isabelle E. Magnin, Christophe Mathieu |
|---|---|
| Rok vydání: | 1996 |
| Předmět: |
Statistics and Probability
business.industry Applied Mathematics ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Pattern recognition Image processing Minimum spanning tree Condensed Matter Physics Minimum spanning tree-based segmentation Modeling and Simulation Graph (abstract data type) Segmentation Geometry and Topology Computer Vision and Pattern Recognition Artificial intelligence business Mathematics |
| Zdroj: | Journal of Mathematical Imaging and Vision. 6:85-96 |
| ISSN: | 1573-7683 0924-9907 |
| DOI: | 10.1007/bf00127376 |
| Popis: | After a brief review and description of the existing graph pyramids used for image processing, particularly stochastic (SP) and adaptive (AP) pyramids, we propose a new strategy to improve the final segmentation provided by such methods. The idea is to translate the image to be segmented into a minimal spanning tree (MST) before building the pyramid. It is shown that the use of an MST in this process improves the results of the pyramidal segmentation. Some experimental results are presented on synthetic and actual images. Finally, a filtering pyramidal algorithm is proposed, using the properties of the minimal spanning tree. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink