Variational optical flow computation assisted by robust point matching

被引:1
作者
Shi, Jinlong [1 ]
Bai, Suqin [1 ]
Shu, Xin [1 ]
机构
[1] Jiangsu Univ Sci & Technol, Sch Comp Sci & Engn, Zhenjiang 212003, Peoples R China
来源
OPTIK | 2015年 / 126卷 / 24期
关键词
Motion estimation; Optical flow; Variational; Point matching; TRAJECTORIES; VIDEO;
D O I
10.1016/j.ijleo.2015.09.115
中图分类号
O43 [光学];
学科分类号
070207 ; 0803 ;
摘要
Lots of optical flow methods are based on coarse-to-fine strategy. However, most of coarse-to-fine optical flow techniques face two important problems: namely (1) large displacements of small structures are usually estimated in an incorrect manner and (2) large affine motions of non-rigid objects cannot be computed accurately. In this paper, we present an optical flow computation methodology that integrates robust point matching into a variational optical flow model to deal with the above-mentioned difficulties. Experimental results are demonstrated for a variety of image sequences to validate the effectiveness. (C) 2015 Elsevier GmbH. All rights reserved.
引用
收藏
页码:5092 / 5099
页数:8
相关论文
共 26 条
[1]  
[Anonymous], 1981, P 7 INT JOINT C ARTI
[2]  
[Anonymous], 2007, P IEEE COMP VIS PATT
[3]   A Database and Evaluation Methodology for Optical Flow [J].
Baker, Simon ;
Scharstein, Daniel ;
Lewis, J. P. ;
Roth, Stefan ;
Black, Michael J. ;
Szeliski, Richard .
INTERNATIONAL JOURNAL OF COMPUTER VISION, 2011, 92 (01) :1-31
[4]   High accuracy optical flow estimation based on a theory for warping [J].
Brox, T ;
Bruhn, A ;
Papenberg, N ;
Weickert, J .
COMPUTER VISION - ECCV 2004, PT 4, 2004, 2034 :25-36
[5]   Large Displacement Optical Flow: Descriptor Matching in Variational Motion Estimation [J].
Brox, Thomas ;
Malik, Jitendra .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2011, 33 (03) :500-513
[6]  
Brox T, 2009, PROC CVPR IEEE, P41, DOI 10.1109/CVPRW.2009.5206697
[7]  
Bruhn A, 2005, IEEE I CONF COMP VIS, P749
[8]   Lucas/Kanade meets Horn/Schunck: Combining local and global optic flow methods [J].
Bruhn A. ;
Weickert J. ;
Schnörr C. .
International Journal of Computer Vision, 2005, 61 (3) :1-21
[10]  
Furukawa Y, 2010, GEOM COMPUT, V5, P193, DOI 10.1007/978-3-642-12392-4_9