Similar Matching for Images with Complex Spatial Relations

Complex spatial relations are important in content-based image retrieval (CBIR) and other applications. However state of the art topological relation models can not deal with complex spatial relations. A complex topological relation model (CTR) is proposed. The similar matching algorithm with CTR is given and applied to CBIR. Experiment shows our method is superior to previous methods in search similar images with complex topological relations.

[1]  Naif Alajlan,et al.  Geometry-Based Image Retrieval in Binary Image Databases , 2008, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[2]  MAX J. EGENHOFER,et al.  Point Set Topological Relations , 1991, Int. J. Geogr. Inf. Sci..

[3]  Yohei Kurata,et al.  The 9+-Intersection: A Universal Framework for Modeling Topological Relations , 2008, GIScience.

[4]  Anthony G. Cohn,et al.  A Spatial Logic based on Regions and Connection , 1992, KR.

[5]  Sanjiang Li,et al.  A complete classification of topological relations using the 9‐intersection method , 2006, Int. J. Geogr. Inf. Sci..

[6]  Sanjiang Li,et al.  Topological Relations between Convex Regions , 2010, AAAI.

[7]  Ulrich Eckhardt,et al.  Shape descriptors for non-rigid shapes with a single closed contour , 2000, Proceedings IEEE Conference on Computer Vision and Pattern Recognition. CVPR 2000 (Cat. No.PR00662).

[8]  Luis Fariñas del Cerro,et al.  A New Tractable Subclass of the Rectangle Algebra , 1999, IJCAI.