Nonparametric image segmentation
| Autor: | R. Kober, T. Kämpke |
|---|---|
| Rok vydání: | 1998 |
| Předmět: |
Mathematical optimization
Nonparametric statistics Scale-space segmentation Image segmentation Image (mathematics) Scale space Function approximation Artificial Intelligence Robustness (computer science) Computer Science::Computer Vision and Pattern Recognition Computer Vision and Pattern Recognition Algorithm Mathematics Parametric statistics |
| Zdroj: | Pattern Analysis and Applications. 1:145-154 |
| ISSN: | 1433-755X 1433-7541 |
| DOI: | 10.1007/bf01259364 |
| Popis: | Image segmentation is reduced to quantisation which in tum is reduced to function approximation. The function approximation problem is formulared and solved as a global optimisation problem requiring neither any parametric assumptions nor parametric input, except the number of desired segment classes of the image. These are characterised by different colours or grey values. The quantisation approach is overlayed with an iteration scheme in accordance with the notion of so-called stable extrema of functions. This leads to segmentations of considerable robustness. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink