DocumentCode :
457145
Title :
Matching 2D Shapes using their Symmetry Sets
Author :
Kuijper, Arjan ; Olsen, Ole Fogh ; Bille, Philip ; Giblin, Peter
Author_Institution :
RICAM, Linz
Volume :
2
fYear :
0
fDate :
0-0 0
Firstpage :
179
Lastpage :
182
Abstract :
We introduce a shape descriptor that is based on the symmetry set. This set represents pairwise symmetric points and consists of several branches. The begin and end points of the branches relate to extrema of the curvature along the shape. Consequently, extrema of the curvature are pairwise connected via a symmetry set branch with a certain finite length. The novel shape descriptor is given by a string representing these extrema, together with the pair wise connections and a length measure. Next, an algorithm is given to match strings. This algorithm is based on a modified shortest path algorithm, taking into account the allowed changes of the symmetry set. Examples show the usability of the presented theory, applied to different types of shapes, including noise and occlusions
Keywords :
graph theory; image matching; optimisation; set theory; 2D shape matching; pairwise symmetric points; shape descriptor; shortest path algorithm; symmetry sets; Data structures; Decision support systems; Electric shock; Indexing; Length measurement; Marine animals; Noise shaping; Shape measurement; Skeleton; Usability;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Pattern Recognition, 2006. ICPR 2006. 18th International Conference on
Conference_Location :
Hong Kong
ISSN :
1051-4651
Print_ISBN :
0-7695-2521-0
Type :
conf
DOI :
10.1109/ICPR.2006.782
Filename :
1699176
Link To Document :
بازگشت