DocumentCode
1610662
Title
A simple texture descriptor for texture retrieval
Author
Liu, Ying ; Zhou, Xiaofang
Author_Institution
Sch. of Inf. Technol. & Electr. Eng., Queensland Univ., Australia
Volume
2
fYear
2003
Firstpage
1662
Abstract
Texture feature extraction is an important part in content-based image retrieval system. This paper describes a simple and effective texture feature extraction algorithm. Previous work based on image decomposition techniques usually use mean and variance of each subband in decomposition domain to form texture descriptor. The proposed texture descriptor uses the variance information and the luminance property of the texture to form the descriptor. Experimental results on the Brodatz texture database show that the proposed descriptor can provide comparable retrieval performance with approximately only half the number of features used compared with previous work.
Keywords
brightness; computer graphics; content-based retrieval; discrete cosine transforms; discrete wavelet transforms; feature extraction; image texture; Brodatz texture database; content-based image retrieval system; luminance property; simple texture descriptor; texture feature extraction; texture retrieval; variance information; Content based retrieval; Discrete cosine transforms; Discrete wavelet transforms; Feature extraction; Image databases; Image retrieval; Indexing; Information retrieval; Multimedia communication; Spatial databases;
fLanguage
English
Publisher
ieee
Conference_Titel
Communication Technology Proceedings, 2003. ICCT 2003. International Conference on
Print_ISBN
7-5635-0686-1
Type
conf
DOI
10.1109/ICCT.2003.1209846
Filename
1209846
Link To Document