Title :
A new indexing approach using outlines primitives
Author_Institution :
Comput. Sci. Dept., Univ. of Sci. & Technol., Houari Boumediene, Algeria
Abstract :
In this paper, a new approach is proposed for modeling and indexing outline shapes by using simple primitives extracted from the contours of objects. We also propose a new and an efficient way to index the outline shapes. Indexes are issued from characteristic polygons determined on the outline shapes of the objects. Despite the recognition process is not the aim of our method, the proposed indexing process only is sufficient to retrieve a few most similar models for a given query. Experiments give very encouraging results in order to apply our approach to the recognition process which will be the next step after this work.
Keywords :
image retrieval; indexing; object recognition; shape recognition; characteristic polygons; model retrieval; object contour; object outline shape indexing; object outline shape modeling; outlines primitive; primitive extraction; query processing; Computational modeling; Computer vision; Equations; Indexing; Mathematical model; Shape; Indexing; Matching; outline shape; parametric equations; primitives;
Conference_Titel :
Computer Systems and Applications (AICCSA), 2013 ACS International Conference on
Conference_Location :
Ifrane
DOI :
10.1109/AICCSA.2013.6616424