Image registration using wavelet-based motion model

被引:57
作者
Wu, YT
Kanade, T
Li, CC
Cohn, J
机构
[1] Natl Chi Nan Univ, Dept Comp Sci & Informat Engn, Puli 545, Taiwan
[2] Carnegie Mellon Univ, Inst Robot, Pittsburgh, PA 15213 USA
[3] Univ Pittsburgh, Dept Elect Engn, Pittsburgh, PA 15261 USA
[4] Univ Pittsburgh, Dept Psychiat & Psychol, Pittsburgh, PA 15261 USA
关键词
image registration; coarse-to-fine motion pyramid; wavelet-based motion model; Cai-Wang wavelet; sum squared difference (SSD); warping;
D O I
10.1023/A:1008101718719
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
An image registration algorithm is developed to estimate dense motion vectors between two images using the coarse-to-fine wavelet-based motion model. This motion model is described by a linear combination of hierarchical basis functions proposed by Cai and Wang (SIAM Numer. Anal., 33(3):937-970, 1996). The coarser-scale basis function has larger support while the finer-scale basis function has smaller support. With these variable supports in full resolution, the basis functions serve as large-to-small windows so that the global and local information can be incorporated concurrently for image matching, especially for recovering motion vectors containing large displacements. To evaluate the accuracy of the wavelet-based method, two sets of test images were experimented using both the wavelet-based method and a leading pyramid spline-based method by Szeliski et al. (International Journal of Computer Vision, 22(3):199-218, 1996). One set of test images, taken from Barron et al. (International Journal of Computer Vision, 12:43-77, 1994), contains small displacements. The other set exhibits low texture or spatial aliasing after image blurring and contains large displacements. The experimental results showed that our wavelet-based method produced better motion estimates with error distributions having a smaller mean and smaller standard deviation.
引用
收藏
页码:129 / 152
页数:24
相关论文
共 46 条
[1]   SPATIOTEMPORAL ENERGY MODELS FOR THE PERCEPTION OF MOTION [J].
ADELSON, EH ;
BERGEN, JR .
JOURNAL OF THE OPTICAL SOCIETY OF AMERICA A-OPTICS IMAGE SCIENCE AND VISION, 1985, 2 (02) :284-299
[2]   A COMPUTATIONAL FRAMEWORK AND AN ALGORITHM FOR THE MEASUREMENT OF VISUAL-MOTION [J].
ANANDAN, P .
INTERNATIONAL JOURNAL OF COMPUTER VISION, 1989, 2 (03) :283-310
[3]  
[Anonymous], ECCV 94
[4]   Optic flow calculation using robust statistics [J].
BabHadiashar, A ;
Suter, D .
1997 IEEE COMPUTER SOCIETY CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, PROCEEDINGS, 1997, :988-993
[5]   PERFORMANCE OF OPTICAL-FLOW TECHNIQUES [J].
BARRON, JL ;
FLEET, DJ ;
BEAUCHEMIN, SS .
INTERNATIONAL JOURNAL OF COMPUTER VISION, 1994, 12 (01) :43-77
[6]  
BERGEN JR, 1992, LECT NOTES COMPUT SC, V588, P237
[7]   Learning parameterized models of image motion [J].
Black, MJ ;
Yacoob, Y ;
Jepson, AD ;
Fleet, DJ .
1997 IEEE COMPUTER SOCIETY CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, PROCEEDINGS, 1997, :561-567
[8]   The robust estimation of multiple motions: Parametric and piecewise-smooth flow fields [J].
Black, MJ ;
Anandan, P .
COMPUTER VISION AND IMAGE UNDERSTANDING, 1996, 63 (01) :75-104
[9]   On the unification of line processes, outlier rejection, and robust statistics with applications in early vision [J].
Black, MJ ;
Rangarajan, A .
INTERNATIONAL JOURNAL OF COMPUTER VISION, 1996, 19 (01) :57-91
[10]   Tracking people with twists and exponential maps [J].
Bregler, C ;
Malik, J .
1998 IEEE COMPUTER SOCIETY CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, PROCEEDINGS, 1998, :8-15