DocumentCode :
2204408
Title :
Image Indexing Using Weighted Color Co-occurrence Matrix and Feature Selection
Author :
Liang, Dong ; Lam, Edmund Y.
Author_Institution :
Inst. of Image Process. & Patter Recognition, Shanghai Jiao Tong Univ.
fYear :
2006
fDate :
14-17 Nov. 2006
Firstpage :
1
Lastpage :
4
Abstract :
In this paper, image indexing based on weighted color co-occurrence matrix (WCCM) feature and isolation parameter-based feature selection is introduced. In this method, isolation parameter (IP) is used to indicate the visual perception complexity and conduct feature selection for each query image. When indexing images from database in the reduced feature space, the similarities of diagonal elements and non-diagonal elements of CCM feature are weighted separately with different values based on the isolation parameters of query image and images from database. The experiments show that the proposed method provides better results than modified color co-occurrence matrix (MCCM) based method and sub-range cumulative histogram (SCH) based method
Keywords :
feature extraction; image colour analysis; image retrieval; indexing; matrix algebra; visual databases; visual perception; diagonal elements; feature space; image database; image indexing; image retrieval; isolation parameter-based feature selection; nondiagonal elements; query images; visual perception complexity; weighted color co-occurrence matrix; Color; Frequency; Image databases; Image processing; Image retrieval; Indexing; Pattern recognition; Spatial databases; Visual databases; Visual perception;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
TENCON 2006. 2006 IEEE Region 10 Conference
Conference_Location :
Hong Kong
Print_ISBN :
1-4244-0548-3
Electronic_ISBN :
1-4244-0549-1
Type :
conf
DOI :
10.1109/TENCON.2006.343708
Filename :
4142398
Link To Document :
بازگشت