DocumentCode :
276106
Title :
The matching problem: heuristic approach considering both split and merge competition
Author :
De Vriendt, J.A.
Author_Institution :
Ghent Univ.
fYear :
1992
fDate :
7-9 Apr 1992
Firstpage :
53
Lastpage :
56
Abstract :
Differences of images in image sequences can be introduced by the motion of objects in the scene, or the motion of the camera. These differences contain important information about the scene. An initial problem in the interpretation of the images is the matching problem. The matching problem is to find matches between sets of feature points in each image. Earlier Barnard and Thompson (1980) and Dreschler and Nagel (1982) proposed matching algorithms which apply the technique of relaxation labeling. Their main assumption is the continuity of the displacement field. Though this assumption is violated when two or more objects, which are near each other, have different velocities. Therefore, the author only assumes the continuity of the displacement field on contours (edges). Indeed, different displacement fields are separated by these edges, but the motion of the edges themselves is continuous
Keywords :
computerised pattern recognition; computerised picture processing; displacement fields; feature points; heuristic approach; image sequences; matching problem; merge; split;
fLanguage :
English
Publisher :
iet
Conference_Titel :
Image Processing and its Applications, 1992., International Conference on
Conference_Location :
Maastricht
Print_ISBN :
0-85296-543-5
Type :
conf
Filename :
146737
Link To Document :
بازگشت