Abstract
Occlusion boundaries are considered either as outliers or as noise in most optical flow algorithms. In order to treat the boundary problem, many probabilistic algorithms like maximum likelihood [6] or expectation-maximization (EM) [17,3] decrease the weights of pixels in boundary regions gradually during estimation iterations. However, these approaches still include the outliers in the estimation. If the number of pixels in boundary regions is comparable to the number of pixels with single motion, we will not be able to estimate the motion parameters robustly since probabilistic methods are purely based on statistics. In this paper, we mark the outliers directly using a method based on eigenvalue analysis [9]. Then we eliminate these outliers in the multiple motion estimation. Comparisons show that this method can improve the precision of estimation results. We also use the “warp-and-subtract” technique to localize and to track occlusion boundaries. The closest work has been done by Fleet et al. [2] as well as by Yu et al. [1]. These are the only approaches with an explicit model of occlusion which, however, is not sufficient to deal with outliers.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
W. Yu, K. Daniilidis, S. Beauchemin, and G. Sommer. Detection and characterization of multiple motion points. In IEEE Conf. Computer Vision and Pattern Recognition, volume I, pages 171-177, Fort Collins, CO, June 23-25, 1999.
D.J. Fleet, M.J. Black, and A.D. Jepson. Motion feature detection using steerable flow fields. In IEEE Conf. Computer Vision and Pattern Recognition, pages 274-281, Santa Barbara, CA, June 23-25, 1998.
M.J. Black and P. Anandan. The robust estimation of multiple motions: parametric and piecewise-smooth flow fields. Computer Vision and Image Understanding, 63(1)75–104, 1996.
M.J. Black and D.J. Fleet. Probabilistic detection and tracking of motion discontinuities. In Proc. Int. Conf. on Computer Vision, volume I, pages 551-558, Kerkyra, Greece, Sep. 20-27, 1999.
M. Bober and J. Kittler. Estimation of complex multimodal motion: an approach based on robust statistics and hough transform. Image and Vision Computing, 12(10):661–668, 1994.
P. Bouthemy. A maximum likelihood framework for determining moving edges. IEEE Trans. Pattern Analysis and Machine Intelligence, 11:499–511, 1989.
J.R. Bergen, P.J. Burt, R. Hingorani, and S. Peleg. A three-frame algorithm for estimating two-component image motion. IEEE Trans. Pattern Analysis and Machine Intelligence, 14(9):886–895, 1992.
B.K.P. Horn. Robot Vision. MIT Press, 1986.
B. Jahne. Spatio-Temporal Image Processing. Springer-Verlag, 1993.
A. Jepson and M.J. Black. Mixture models for optical flow computation. In IEEE Conf. Computer Vision and Pattern Recognition, pages 760-761, New York, NY, June 15-17, 1993.
H.H. Nagel and W. Enkelmann. An investigation of smoothness constraints for the estimation of displacement vector fields from image sequences. IEEE Trans. Pattern Analysis and Machine Intelligence, 8:565–593, 1986.
J.M. Nash, J.N. Carter, M.S. Nixon. Dynamic feature extraction via the velocity hough transform. Pattern Recognition Letters, 18:1035–1047, 1997.
B.G. Schunck. Image flow segmentation and estimation by constraint line clustering. IEEE Trans. Pattern Analysis and Machine Intelligence, 11(10) 1010–1027, 1989.
J. Shi and J. Malik. Normalized cuts and image segmentation. In IEEE Conf Computer Vision and Pattern Recognition, 1997.
J. Shi and J. Malik. Motion segmentation using normalized cuts. In Proc. Int. Conf. on Computer Vision, pages 1154-1160, Bombay, India, Jan. 4-7, 1998.
J. Weickert and C. Schnorr. Räumlich-zeitliche Berechnung des optischen Flusses mit nichtlinearen flußabhängigen Glattheitstermen. In DAGM Symposium Mustererkennung, pages 317-324, Bonn, Germany, Sep. 15 - 17, 1999.
Y. Weiss and E.H. Adelson. A unified mixture framework for motion segmentation: Incorporating spatial coherence and estimating the number of models. In IEEE Conf. Computer Vision and Pattern Recognition, pages 321-326, San Fransisco, CA, June 18-20, 1996.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2000 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Yu, W., Daniilidis, K., Sommer, G. (2000). Eliminating Outliers in Motion Occlusion Analysis. In: Sommer, G., Krüger, N., Perwass, C. (eds) Mustererkennung 2000. Informatik aktuell. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-59802-9_47
Download citation
DOI: https://doi.org/10.1007/978-3-642-59802-9_47
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-67886-1
Online ISBN: 978-3-642-59802-9
eBook Packages: Springer Book Archive