Improved SVD matching based on Delaunay triangulation and belief propagation

Finding correspondences between pair of images of the same scene is a key problem in computer vision. When matching images undergo viewpoint change, partial occlusion, clutters and illumination change, there will be a lot of mismatches due to the limited repeatability and discriminative power of features. In this paper, we propose a robust matching algorithm that can remove false matches and propagate the correct ones to obtain more matches, thus improve the matching accuracy. Firstly, extract SURF (Speeded Up Robust Features) descriptors from the input two images, which can be used to build the proximity matrix. Then perform SVD(Singular Value Decomposition) on the proximity matrix to obtain the initial matches. Thirdly, refine the initial matches by the unique property of Delaunay triangulation that can produce the maximum clique of the two Delaunay graph. Finally, recover the lost matches with the constraint of dual graph of Voronoi. Experimental results on Oxford datasets indicate that our algorithm can improve the match performance compared to the RANSAC-based method.

[1]  Jianfang Dou,et al.  Image matching based local Delaunay triangulation and affine invariant geometric constraint , 2014 .

[2]  G LoweDavid,et al.  Distinctive Image Features from Scale-Invariant Keypoints , 2004 .

[3]  Mariette Yvinec,et al.  Algorithmic geometry , 1998 .

[4]  Maurizio Pilu,et al.  A direct method for stereo correspondence based on singular value decomposition , 1997, Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[5]  H. C. Longuet-Higgins,et al.  An algorithm for associating the features of two images , 1991, Proceedings of the Royal Society of London. Series B: Biological Sciences.

[6]  Abbas Edalat,et al.  Computing Delaunay Triangulation with Imprecise Input Data , 2003, CCCG.

[7]  J. Boissonnat,et al.  Algorithmic Geometry: Frontmatter , 1998 .

[8]  Manuel Abellanas,et al.  Structural Tolerance and Delaunay Triangulation , 1999, Inf. Process. Lett..

[9]  Jianfang Dou,et al.  Robust image matching based on SIFT and delaunay triangulation , 2012 .

[10]  Xinyue Zhao,et al.  Improved keypoint descriptors based on Delaunay triangulation for image matching , 2014 .

[11]  Binoy Pinto,et al.  Speeded Up Robust Features , 2011 .

[12]  Robert C. Bolles,et al.  Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography , 1981, CACM.

[13]  Jan Flusser,et al.  Image registration methods: a survey , 2003, Image Vis. Comput..

[14]  Lisa M. Brown,et al.  A survey of image registration techniques , 1992, CSUR.

[15]  D. Hill,et al.  Non-rigid image registration: theory and practice. , 2004, The British journal of radiology.

[16]  Luc Van Gool,et al.  SURF: Speeded Up Robust Features , 2006, ECCV.

[17]  David P. Dobkin,et al.  The quickhull algorithm for convex hulls , 1996, TOMS.

[18]  Feng Zhao,et al.  Image Matching Based on Singular Value Decomposition , 2004, PCM.

[19]  Francesca Odone,et al.  SVD-matching using SIFT features , 2006, Graph. Model..