Title:	A SIMILARITY MEASURE BASED ON GEOMETRIC CONSISTENCY FOR IMAGE RETRIEVAL USING LOCAL DESCRIPTORS
Author(s):	A. Najjar, E. Zagrouba
ISBN:	978-972-8939-48-9
Editors:	Yingcai Xiao
Year:	2011
Edition:	Single
Keywords:	CBIR, Geometric consistency, RANSAC, Similarity measures, SIFT.
Type:	Short Paper
First Page:	223
Last Page:	227
Language:	English
Cover:
Full Contents:	click to dowload
Paper Abstract:	The content based image retrieval (CBIR) issue can be formulated as a problem of research in the primitive space. These primitives generally describe the similarity at the photometric level. In this work, we focus on the study of geometric consistency to measure the similarity between images in order to build an image retrieval system. In fact, we compute similarity by considering the geometric positions of the points of interest (SIFT). Then, we compare our method to a RANSAC algorithm based method as an alternative to compute geometric consistency. The evaluation of results using two subsets of image (MPEG7 and COIL) shows that the proposed method allows obtaining a better retrieval result.
	Go Back