Artificial intelligence

Abstract: In this paper is considered the problem of finding correspondences between key points of stereo images. Stereo images are presented by graphs each vertex of which corresponds to the key point, and the edges represent the relationships between them. The correspondence between the key points is a result of isomorphic intersection of graphs. Search graphs intersections by using the method of constructing an additional graph pyramid. It is shown that the use of the relative position of key points and the method based on the use of an additional graph pyramid can efficiently solve the problem under consideration without involving complex descriptors of key points that take into account their individual characteristics.

Keywords: pattern recognition, key points, graphs matching, stereo vision, key points matching

References:

1. David G. Lowe Distinctive Image Features from Scale-Invariant Keypoints / David G. Lowe // International Journal of Computer Vision. – 2004. – Vol. 2, No 60. – P.91-110.
2. Mikolajczyk K. Detection of local features invariant to affine transformations / Mikolajczyk K. // Ph.D. thesis. – Institut National Polytechnique de Grenoble, France. – 2002. – 171 p.
3. Krystian Mikolajczyk Scale & Affine Invariant Interest Point Detectors / Krystian Mikolajczyk аnd Cordelia Schmid // International Journal of Computer Vision. – 2004. – Vol. 60, No 1. – P. 63-86.
4. Krystian Mikolajczyk and Cordelia Schmid. A performance evaluation of local descriptors // IEEE Transactions on Pattern Analysis & Machine Intelligence. – 2005. – Vol. 27, No 10. – P. 1615-1630.
5. Gyuri Dork'o Maximally Stable Local Description for Scale Selection / Gyuri Dork'o and Cordelia Schmid // European Conference on Computer Vision. – Graz, Austria. – 2006. – Vol. 4. – P. 504-516.
6. Winter M. Maximally Stable Corner Clusters: A novel distinguished region detector and descriptor / Winter M., Bischof H. and Fraundorfer F // 1st Austrian Cognitive Vision Workshop. – Zell an der Pram, Austria. – 2005. – P. 59-66.
7. Martin A. Fischer Random Consensus: A Paradigm for Model Fitting with Application to Image Analysis and Automated Cartography / Martin A. Fischer and Robert C. Bolles // Commun. Assoc. Comp. Mach. – 1981. – Vol. 24. – P. 381-395.
8. Gorokhovatsky V.O. Evaluation of the structural similarity of objects as sets of components / V.O. Gorokhovatsky // System Research & Information Technologies. – 2011. – No 1. – P. 57-70.
9. Gorokhovatsky V.O. Systems of features based on spatially-attributive relations of image structural elements / V.O. Gorokhovatsky // Bionics of Intelligence: Sci. Mag. – 2011. – No 1 (75). – P. 48-51.
10. Gorokhovatsky V.O. Filtering false correspondences descriptor points based on the analysis of geometrical data / V.O. Gorokhovatsky // Registration, storage and processing. – 2009. – V. 11, No 1. – P. 11-19. – ISSN 1560-9189.
11. Optimal Feature Matching Method using Bayesian Graph Theory / Wan Hyun Cho, In Seop Na, Sun Worl Kim and Soo Hyung Kim // International Journal of Multimedia and Ubiquitous Engineering. – 2012. – Vol. 7, No 3. – P. 123-132.
12. Torresani L. Feature Correspondence Via Graph Matching: Models and Global Optimization. / Torresani L., Kolmogorov V., Rother C. // Proceedings of the 10th European Conference on Computer Vision : Part II. – 2008. – P. 596-609.
13. Agarkov A.V. The method of two graphs matching in polynomial time / A.V. Agarkov // Artificial Intelligence. – 2003. – No 4. – P. 172-184.
14. Agarkov A.V. Isomorphic Intersection of Two Graphs Searching in Polynomial Time / A.V. Agarkov // Artificial Intelligence. – 2007. – No 2. – P. 62-74.
15. Agarkov A.V. Search of the Set of Maximal Cliques Based on the Method for Constructing Complementary Graph / A.V. Agarkov // Artificial Intelligence. – 2011. – No 3. – P. 190-199.