Robust Total least Squares Based Optic Flow Computation

This paper considers the problem of finding a robust solution to the optic flow problem. The optical flow field is recovered by solving a system of over-determined linear equations with all the data matrices containing both outliers and noise. Here, we present a novel and very effective solution for this problem called weighted total least squares. The weights for this method are computed using a new robust statistical method named least median of squares orthogonal distances. Unlike the total least squares which is only robust to noise, this method is extremely robust to both noise and outliers and can tolerate up to 50% of equations in the system to be contaminated by outliers. The proposed weighting method is fast and the total computation remains inexpensive. To demonstrate the performance of the proposed algorithm, we compare the accuracy of our algorithm for computing optic flow field for a number of synthetic and real image sequences and show that the proposed method, despite being very simple and straightforward, out performs all methods used for comparison.

[1]  E. Delp,et al.  Estimating displacement vectors from an image sequence , 1989 .

[2]  P. Anandan,et al.  Hierarchical Model-Based Motion Estimation , 1992, ECCV.

[3]  Kenji Mase,et al.  Simultaneous multiple optical flow estimation , 1990, [1990] Proceedings. 10th International Conference on Pattern Recognition.

[4]  Richard Szeliski,et al.  Hierarchical spline-based image registration , 1994, 1994 Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.

[5]  Jitendra Malik,et al.  Robust computation of optical flow in a multi-scale differential framework , 1993, 1993 (4th) International Conference on Computer Vision.

[6]  Vishal Markandey,et al.  Total least squares fitting spatiotemporal derivatives to smooth optical flow fields , 1992, Defense, Security, and Sensing.

[7]  Michael J. Black.  Recursive Non-Linear Estimation of Discontinuous Flow Fields , 1994, ECCV.

[8]  Charles V. Stewart,et al.  Bias in robust estimation caused by discontinuities and multiple structures , 1997, IEEE Trans. Pattern Anal. Mach. Intell..

[9]  David Suter,et al.  Robust optic flow estimation using least median of squares , 1996, Proceedings of 3rd IEEE International Conference on Image Processing.

[10]  P. Rousseeuw.  Least Median of Squares Regression , 1984 .

[11]  A. Jepson,et al.  Estimating multiple independent motions in segmented images using parametric models with local deformations , 1994, Proceedings of 1994 IEEE Workshop on Motion of Non-rigid and Articulated Objects.

[12]  Sabine Van Huffel,et al.  Total least squares problem - computational aspects and analysis , 1991, Frontiers in applied mathematics.

[13]  T. O. Kvålseth.  Cautionary Note about R 2 , 1985 .

[14]  Andrea Giachetti,et al.  Refinement of Optical Flow Estimation and Detection of Motion Edges , 1996, ECCV.

[15]  Peter J. Rousseeuw,et al.  Robust regression and outlier detection , 1987 .

[16]  Berthold K. P. Horn,et al.  Determining Optical Flow , 1981, Other Conferences.

[17]  Subhasis Chaudhuri,et al.  Performance analysis of total least squares methods in three-dimensional motion estimation , 1991, IEEE Trans. Robotics Autom..

[18]  Michael Spann,et al.  Robust Computation of Optical Flow , 1995, BMVC.

[19]  Hans-Hellmut Nagel,et al.  Optical Flow Estimation: Advances and Comparisons , 1994, ECCV.

[20]  Ricardo D. Fierro,et al.  The Total Least Squares Problem: Computational Aspects and Analysis (S. Van Huffel and J. Vandewalle) , 1993, SIAM Rev..

[21]  Sabine Van Huffel,et al.  The total least squares problem , 1993 .

[22]  Michael J. Black,et al.  Skin and bones: multi-layer, locally affine, optical flow and regularization with transparency , 1996, Proceedings CVPR IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[23]  Michael J. Black,et al.  The Robust Estimation of Multiple Motions: Parametric and Piecewise-Smooth Flow Fields , 1996, Comput. Vis. Image Underst..

[24]  David Suter,et al.  Optic flow calculation using robust statistics , 1997, Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition.