AbstractInclusion of local color information in generic object recognition is ignored by almost all approaches, although it is important and can improve the recognition performance. In this paper, we present a generic object recognition approach using boosting as a learning technique. Simple local color descriptors combined with the SIFT descriptors are used. Experiments using benchmark and complex generic object datasets are performed, and good performance is obtained. © 2008 Pleiades Publishing, Ltd.
|Issue Date||1-Jun-2008||Journal||Pattern Recognition and Image Analysis||URI||http://research.asu.edu.eg/123456789/1034||DOI||2
Recommend this item
Items in Ain Shams Scholar are protected by copyright, with all rights reserved, unless otherwise indicated.