Title :
Visually Significant Dynamics for Watershed Segmentation
Author :
Climent, Joan ; Sanfeliu, Alberto
Author_Institution :
Dept. of Comput. Eng. & Autom. Control, Tech. Univ. of Catalonia
Abstract :
The watershed transform is a powerful tool for segmentation once we can deal with oversegmentation. To solve the oversegmentation problem, hierarchical approaches are considered in order to retain the most significant regions of the image at different scales. The dynamics of the regional minima have been used to build this hierarchy. In this paper we present a new measure for computing the dynamics of the minima based on human perception of shapes. The described technique solves the major drawbacks of the hierarchical segmentations based on contrast dynamics or volume dynamics
Keywords :
image segmentation; contrast dynamics; hierarchical segmentations; volume dynamics; watershed segmentation; watershed transform; Area measurement; Automatic control; Computer vision; Humans; Image segmentation; Object recognition; Shape measurement; Topology; Visual perception; Volume measurement;
Conference_Titel :
Pattern Recognition, 2006. ICPR 2006. 18th International Conference on
Conference_Location :
Hong Kong
Print_ISBN :
0-7695-2521-0
DOI :
10.1109/ICPR.2006.1182