We compare two tree-based, hierarchical representations of volumetric gray-scale images for data-driven image filtering. One representation is the max-tree, in which tree nodes represent connected components of all level sets of a data set. The other representation is the watershed tree, consisting of nodes representing nested, homogeneous image regions. Region attribute-based filtering is achieved by pruning the trees. Visualization is used to compare both the filtered images and trees. In our comparison, we also consider flexibility, intuitiveness, and extendability of both tree representations.
|Title of host publication||Mathematical Morphology and Its Applications to Image and Signal Processing (10th International Symposium, ISMM 2011, Verbania-Intra, Italy, July 6-8, 2011. Proceedings)|
|Editors||P. Soille, M. Pesaresi, G.K. Ouzounis|
|Place of Publication||Berlin|
|Publication status||Published - 2011|
|Name||Lecture Notes in Computer Science|