Title :
Shape matching using curvature processes
Author :
Milios, Evangelos E.
Author_Institution :
Dept. of Comput. Sci., Toronto Univ., Ont., Canada
Abstract :
The problem of matching two smooth closed contours which are related by deformations that alter their curvature structure is discussed. Deformations typically include protrusions, a result of internal resistance, and indentations, a result of squashing. The author´s algorithm establishes correspondences between segments that have not drastically deformed (in the sense of introducing or eliminating protrusions and indentations), and recovers productions that explain how the remaining segments have deformed. Criteria involving metric information, such as the correlation between extended circular images of segments of constant curvature sign, are imposed to select the preferred interpretation
Keywords :
correlation methods; pattern recognition; picture processing; constant curvature sign; correlation; curvature processes; curvature structure; deformations; extended circular images; indentations; internal resistance; metric information; protrusions; shape matching; smooth closed contours; squashing; Clouds; Computer graphics; Dynamic programming; Image analysis; Image segmentation; Piecewise linear approximation; Piecewise linear techniques; Satellites; Shape measurement; Spline;
Conference_Titel :
Acoustics, Speech, and Signal Processing, 1988. ICASSP-88., 1988 International Conference on
Conference_Location :
New York, NY
DOI :
10.1109/ICASSP.1988.196750