Relaxation labeling for non-rigid point matching under neighbor preserving

Abstract

Non-rigid point matching has received more and more attention. Recently, many works have been developed to discover global relationships in the point set which is treated as an instance of a joint distribution. However, the local relationship among neighboring points is more effective under non-rigid transformations. Thus, a new algorithm taking advantage of shape context and relaxation labeling technique, called SC-RL, is proposed for non-rigid point matching. It is a strategy that joints estimation for correspondences as well as the transformation. In this work, correspondence assignment is treated as a soft-assign process in which the matching probability is updated by relaxation labeling technique with a newly defined compatibility coefficient. The compatibility coefficient is one or zero depending on whether neighboring points preserving their relative position in a local coordinate system. The comparative analysis has been performed against four state-of-the-art algorithms including SC, ICP, TPS-RPM and RPM-LNS, and the results denote that SC-RL performs better in the presence of deformations, outliers and noise.

Keywords

non-rigid / point matching / shape context / relaxation labeling

References

[1] WangZ-z, LiangMin. Locally affine invariant descriptors for shape matching and retrieval [J]. IEEE Siginal Processing Letters, 2010, 17(9): 803-806

[2] WangW, XiongB-l, SunH, CaiH-p, KuangG-yao. An affine invaraint relative attitude relationship descriptor for shape matching based on ratio histograms [J]. EURASIP Journal on Advances in Signal Processing, 2012

[3] WangY, TeohE K. 2D affine invariant contour matching using B-Spline model [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2007, 29(10): 1853-1858

[4] ZhangG, MaZ-m, NiuL-q, ZhangC-ming. Modified Fourier descriptor for shape feature extraction [J]. Journal of Central South University of Technology, 2012, 19(2): 488-495

[5] WangW, JiangY-m, XiongB-l, ZhaoL-j, KuangG-yao. Contour matching using the affine invariant support point set [J]. IET Computer Vision, 2014

[6] CrossA D J, HancockE R. Graph matching with a dual-step EM algorithm [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1998, 20(11): 1236-1253

[7] LeeJ-h, WonC-hee. Topology preserving relaxation labeling for nonrigid point matching [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(2): 427-432

[8] BeslP J, MckayN D. A method for registration of 3-D shapes [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1992, 14(2): 239-256

[9] ChuiH, RangarajanA. A new point matching algorithm for non-rigid registartion [J]. Computer Vision and Image Understanding, 2003, 89(2/3): 114-141

[10] MyronenkoA, SongX-bo. Point set registration: Coherent point drift [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010, 32(12): 2262-2275

[11] LiuZ-x, AnJ-b, JingYu. A simple and robust feature point matching algorithm based on restricted spatial order constraints for aerial image registration [J]. IEEE Transactions on Geoscience and Remote Sensing, 2012, 50(2): 514-527

[12] IzadiM, SaeediP. Robust weighted graph transformation matching for rigid and nonrigid image registration [J]. IEEE Transactions on Image Processing, 2012, 21(10): 4369-4382

[13] ZhangH, MuY, YouY-h, LiJ-wei. Multi-scale sparse feature point correspondence by graph cuts [J]. Science China (Information Sciences), 2010, 53(6): 1224-1232

[14] AguilarW, FrauelY, EscolanoF, PerezM M, RomeroA E, LozanoM. A robust graph transformation matching for non-rigid registration [J]. Image and Vision Computing, 2009, 27(7): 897-910

[15] BelongieS, MalikJ, PuzichaJ. Shape matching and object recognition using shape contexts [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002, 24(24): 509-522

[16] ZhengY-f, DoermannD. Robust point matching for nonrigid shapes by preserving local neighborhood structures [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2006, 28(4): 643-649

[17] ZhengY-f, DoermannD. Robust point matching for two-dimensional nonrigid shapes [C]. 10th IEEE Int Conf Comput Vis. Beijing, 20051561-1566

[18] WangH-f, HancockE R. Probabilistic relaxation labelling using the Fokker-Planck equation [J]. Pattern Recognition, 2008, 41: 3393-3411

[19] RosenfeldA, HummelR A, ZuckerS W. Scene labeling by relaxation operations [J]. IEEE Trans System, Man and Cybernetics, 1976, 6(6): 420-433

[20] FaugerasO D, BerthodM. Improving consistency and reducing ambiguity in stochastic labeling: An optimization approach [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1981, 3(4): 412-424

[21] HummelR A, ZuckerS W. On the foundations of relaxation labeling processes [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1983, 5(3): 267-287

[22] BookstainF L. Principal warps: Thin-plate splines and the decomposition of deformations [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1989, 11(6): 567-585

[23] YuilleA L, GrzywacsN M. A mathematical analysis of the motion coherence theory [J]. Internation Journal of Computer Vision, 1989, 3(2): 155-175