Title :
Object Decomposition based on Superellipses
Author_Institution :
Dept. of Comput. Sci. & Inf. Eng., Nat. Penghu Univ.
fDate :
Aug. 30 2006-Sept. 1 2006
Abstract :
Superellipse is a flexible primitive and it can represent a large variety of shapes. Hence, this paper presents a superellipse decomposition of object contours. In the proposed method, part segmentation algorithm based on breakpoint types is used to obtain the convex parts of object contours. These convex parts are isolated and fitted by using superellipses. Tests show that the proposed method has a number of interesting properties including being scale/rotation/translation invariant, threshold-free, and efficient
Keywords :
curve fitting; image segmentation; object detection; pattern classification; breakpoint classification; contour-based part segmentation algorithm; object decomposition; superellipse fitting; Application software; Computer science; Computer vision; Feature extraction; Fitting; Humans; Partitioning algorithms; Pattern recognition; Shape; Testing;
Conference_Titel :
Innovative Computing, Information and Control, 2006. ICICIC '06. First International Conference on
Conference_Location :
Beijing
Print_ISBN :
0-7695-2616-0
DOI :
10.1109/ICICIC.2006.484