Premium
An efficient reflection invariance region‐based image retrieval framework
Author(s) -
Kao ChiChou,
Lai YenTai,
Lin ChiaHui
Publication year - 2010
Publication title -
international journal of imaging systems and technology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.359
H-Index - 47
eISSN - 1098-1098
pISSN - 0899-9457
DOI - 10.1002/ima.20228
Subject(s) - image retrieval , computer science , artificial intelligence , image (mathematics) , shape context , histogram , reflection (computer programming) , context (archaeology) , pattern recognition (psychology) , feature (linguistics) , computer vision , visual word , geography , linguistics , philosophy , archaeology , programming language
The shape context is used as the shape feature descriptor in a region‐based image retrieval system. However, a given image is uneasy to be retrieved if the images are reflected. To retrieve efficiently different images, an efficient reflection invariance region‐based image retrieval framework is presented in this article. In the proposed reflection invariance algorithm, the reflecting shape contexts together are first clustered and then label each cluster so that the shape contexts in each cluster have the same label. Using the histogram of label frequencies can quickly and efficiently search for similar or reflecting shapes. The experimental results have shown that our region‐based image retrieval system is effective and has better retrieval performance than the existing systems. © 2010 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 20, 155–161, 2010