DocumentCode
3086165
Title
Multiple motion object segmentation based on homogenous region merging
Author
Li, H. ; Tye, B.J. ; Ong, E.P. ; Lin, W.S. ; Ko, C.
Author_Institution
Dept. of Electr. Eng., Nat. Univ. of Singapore, Singapore
Volume
5
fYear
2001
fDate
2001
Firstpage
175
Abstract
In this paper, a novel multiple motion object segmentation scheme based on homogenous region merging is presented. This scheme is based on utilization of a robust multiresolution motion estimator that can provide good motion estimates from irregularly-shaped regions. In addition, a novel region-merging criterion and an iterative region-merging process have been proposed. Experimental results demonstrate the effectiveness of the proposed method
Keywords
image segmentation; iterative methods; motion estimation; object recognition; homogenous region merging; irregularly-shaped regions; iterative region-merging process; multiple motion object segmentation; region-merging criterion; robust multiresolution motion estimator; Computer vision; Data mining; Image segmentation; Merging; Motion estimation; Motion measurement; Motion segmentation; Object segmentation; Robustness; Testing;
fLanguage
English
Publisher
ieee
Conference_Titel
Circuits and Systems, 2001. ISCAS 2001. The 2001 IEEE International Symposium on
Conference_Location
Sydney, NSW
Print_ISBN
0-7803-6685-9
Type
conf
DOI
10.1109/ISCAS.2001.922013
Filename
922013
Link To Document