DocumentCode :
576972
Title :
Efficient real-time contour matching
Author :
Robert, Varga ; Arthur, Costea ; Istvan, Szakats ; Sergiu, Nedevschi
Author_Institution :
Tech. Univ. of Cluj Napoca, Cluj-Napoca, Romania
fYear :
2012
fDate :
Aug. 30 2012-Sept. 1 2012
Firstpage :
193
Lastpage :
199
Abstract :
In this paper we present certain shape descriptors from the literature that enable real-time contour matching. We propose slight modifications to these descriptors and more importantly provide fast and efficient matching techniques that return distances between shapes in the order of milliseconds. We show that applying these instead of more time consuming matching algorithms matching accuracy remains the same but the speed is at least 10 times faster. We show results on the MPEG7 benchmark. Using our matching we improve some scores from the literature, this is mainly due to the aligning of the shapes and the usage of the L1 metric.
Keywords :
computer vision; image matching; real-time systems; shape recognition; MPEG7 benchmark; computer vision; efficient real-time contour matching; image matching techniques; shape descriptors; Complexity theory; Context; Histograms; Measurement; Real-time systems; Shape; Vectors; Image processing; MPEG7 benchmark; contour matching; object classification; shape context descriptor; shape descriptors;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Intelligent Computer Communication and Processing (ICCP), 2012 IEEE International Conference on
Conference_Location :
Cluj-Napoca
Print_ISBN :
978-1-4673-2953-8
Type :
conf
DOI :
10.1109/ICCP.2012.6356185
Filename :
6356185
Link To Document :
بازگشت