Merged palette histogram relevance feedback for MPEG-7 dominant color descriptor

Author

Wong, Ka-Man ; Po, Lai-Man ; Chan, Kin-Yeung

Author_Institution

Dept. of Electr. Eng., City Univ. of Hong Kong, China

fYear

2004

fDate

20-22 Oct. 2004

Firstpage

177

Lastpage

180

Abstract

A relevance feedback (RF) technique for improving the effectiveness of content based image retrieval using MPEG-7 dominant color descriptor (DCD) is proposed in this paper. It uses a merged palette histogram (MPH) technique for generating a new query DCD histogram from the relevant image set. The MPH approach can generate a natural query that can represent the selected relevance images. The new method has been implemented on a MPEG-7 XM platform with use of the MPEG-7 common color dataset. The effectiveness of this new MPH-RF technique using different similarity measures is demonstrated by experimental results in the terms of averaged normalized modified retrieval rate (ANMRR), computational analysis and visual retrieved images comparison.

Keywords

code standards; content-based retrieval; feature extraction; image colour analysis; image representation; image retrieval; relevance feedback; statistical analysis; MPEG-7; XM platform; averaged normalized modified retrieval rate; common color dataset; computational analysis; content based image retrieval; dominant color descriptor; histogram relevance feedback; image representation; merged palette histogram; similarity measures; visual retrieved images comparison; Ambient intelligence; Color; Feedback; Histograms; Image analysis; Image retrieval; MPEG 7 Standard; Radio frequency; Radiofrequency identification; Size measurement;

fLanguage

English

Publisher

ieee

Conference_Titel

Intelligent Multimedia, Video and Speech Processing, 2004. Proceedings of 2004 International Symposium on

Print_ISBN

0-7803-8687-6

Type

conf

DOI

10.1109/ISIMP.2004.1434029

Filename

1434029