Title :
A method for subpart decomposition for overlapped object identification
Author :
Damergi, Tayeb ; Ionescu, Dan
Author_Institution :
Dept. of Electr. Eng., Ottawa Univ., Ont., Canada
Abstract :
A number of computer vision systems have been devised to deal with overlapped object identification. Objects can be considered as built by subparts. A special technique of decomposing objects in subparts is used for building object models as lists of above subparts. The subparts are obtained by applying a decomposition algorithm on the edge representation of the object subject to identification. Examples are used to illustrate the approach
Keywords :
computer vision; edge detection; computer vision systems; decomposition algorithm; edge representation; overlapped object identification; subpart decomposition; subparts; Integrated circuit noise; Shape;
Conference_Titel :
Electrical and Computer Engineering, 1993. Canadian Conference on
Conference_Location :
Vancouver, BC
Print_ISBN :
0-7803-2416-1
DOI :
10.1109/CCECE.1993.332462
