Algorithms for clustering data
Algorithms for clustering data
Image Analysis and Mathematical Morphology
Image Analysis and Mathematical Morphology
Hierarchical clustering based on ordinal consistency
Pattern Recognition
| Hi-index | 0.10 |
This paper introduces a new hierarchical clustering method based on a morphological transformation of a threshold graph. A description of a complex large graph as a union of simple connected subgraphs is obtained by linking nodes between progressively eroded versions of the original graph. The resulting hierarchical clustering (i) agrees with the intuitive notion of perceptual grouping and (ii) is an efficient tool for qualitative and quantitative cluster analysis.