Translator Disclaimer
24 June 2005 A stepwise similarity approximation of spatial constraints for image retrieval
Author Affiliations +
Proceedings Volume 5960, Visual Communications and Image Processing 2005; 59604K (2005) https://doi.org/10.1117/12.632728
Event: Visual Communications and Image Processing 2005, 2005, Beijing, China
Abstract
A real image is assumed to be associated with some content-based meta-data about that image (i.e., information about objects in the image and spatial relationships among them). Recently Zhang and Yau have addressed the approximate picture matching problem, and have presented a stepwise approximation of intractable spatial constraints in an image query. In particular, in contrast with very few cases done in earlier related works, Zhang-Yau's algorthmic analysis shows that there are all possible 16 cases for results of the object matching step of image retrieval, and 13 out of these 16 cases are valid for the stepwise approximation of spatial constraints while the only other 3 cases are identified impossible for finding an exact picture-matching between a query picture and a databse picture. In this paper, Zhang and Yau have successfully used the stepwise approximation method to work out a simliarity measure between a query image and a database image, for image retrieval. The proposed similiarity measure utlizes the simliarity measures previously developed, by Gudivada and Raghavan (1995) and El-Kwae and Kabuka (1999), for the scenario of the single occurrence of each object in both query and databse images, and extends to cover all 13 valid cases.
© (2005) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Qing-Long Zhang and Stephen S.-T. Yau "A stepwise similarity approximation of spatial constraints for image retrieval", Proc. SPIE 5960, Visual Communications and Image Processing 2005, 59604K (24 June 2005); https://doi.org/10.1117/12.632728
PROCEEDINGS
12 PAGES


SHARE
Advertisement
Advertisement
Back to Top