Digital image processing (2nd ed.)
Digital image processing (2nd ed.)
Computer and Robot Vision
Spatial Size Distributions: Applications to Shape and Texture Analysis
IEEE Transactions on Pattern Analysis and Machine Intelligence
Shape description from generalized support functions
Pattern Recognition Letters
| Hi-index | 0.00 |
The subject of this paper is to propose and test a set ofnumerical descriptors of 2D binary planar shapes. Given a shape, A, the transformations of A with a given mathematical morphological operation and different structuring elements are considered. The measures of this family of transformed setsprovide a numerical description of the original set A.These descriptors are very robust against noise and maintain areasonable discriminatory power. The robustness against different levels of contour degradation is tested bysimulation. Starting with a clean (without noise) set, &Lgr;, it isassumed that the observed set, A, is a noisy version (with contourdegradation) of &Lgr;.The performance of the descriptors, when they are used to comparedifferent shapes or shapes from a scene with models, is studied andcompared with related descriptors based on thegranulometric analysis of the original set, which are the closest previous alternative to our approach in the literature.