This Article 
 Bibliographic References 
 Add to: 
Occam Algorithms for Computing Visual Motion
November 1995 (vol. 17 no. 11)
pp. 1033-1042

Abstract—The standard approach to computing motion relies on pixel correspondence. Computational schemes impose additional constraints, such as smoothness and continuity of the motion vector field, even though these are not directly related to pixel correspondence. This paper proposes an alternative to the multiple constraints approach. By drawing analogy with machine learning, motion is computed as a function that accurately predicts frames. The Occam-Razor principle suggests that among all functions that accurately predict the second frame from the first frame, the best predictor is the “simplest,” and simplicity can be rigorously defined in terms of encoding length. An implementation of a practical algorithm is described. Experiments with real video sequences verify the algorithm assumptions by showing that motion in typical sequences can be accurately described in terms of a few parameters. Our particular choice of predictors produces results that compare very favorably with other image flow algorithms in terms of accuracy and compactness. It may, however, be too constrained to enable accurate recovery of 3D motion and structure.

[1] G. Adiv,“Determining three-dimensional motion and structure from optical flowgenerated by several moving objects,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 7, pp. 384-401, 1985.
[2] P. Anandan,“A computational framework and an algorithm for the measurement ofvisual motion,” Int’l J. Computer Vision, vol. 2, pp. 283-310, 1989.
[3] J.L. Barron, D.J. Fleet, and S.S. Beauchemin, “Performance of Optical Flow Techniques,” Int'l J. Computer Vision, vol. 12, no. 1, pp. 43–77, 1994.
[4] J.R. Bergen and E.H. Adelson,“Hierarchical, computationally efficient motion estimationalgorithm,” J. Optical Soc. Am. A, vol. 4, no. 13, p. 48, 1987.
[5] J.R. Bergen, P. Anandan, K.J. Hanna, and R. Hingorani, “Hiercharchical Model-Based Motion Estimation,” Proc. European Conf. Computer Vision, pp. 237-252, 1992.
[6] 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, vol. 14, pp. 886-895, Sept. 1992.
[7] A. Blumer, A. Ehrenfeucht, D. Haussler, and M. Warmuth, "Occam's Razor," Information Processing Letters, vol. 24, NorthHolland, pp. 377-380, 1987.
[8] A. Blumer, A. Ehrenfeucht, D. Haussler, and M. Warmuth, "Learnability and the Vapnik-Chervonenkis Dimension," J. ACM, vol. 36, pp. 929-965, 1989.
[9] M. Hestenes,Conjugate Direction Methods in Optimization.New York: Springer-Verlag, 1980.
[10] T. Hidaka and K. Ozawa,“Subjective assesment of redundancy-reduced moving images forinteractive applications: Test methodology and report,” Signal Processing, Image Comm., vol. 2, no. 2, Aug. 1990.
[11] B.K. Horn, Robot Vision. Cambridge, Mass.: MIT Press, 1986.
[12] B.K. Horn and B.G. Schunck,“Determining optical flow,” Artificial Intelligence, vol. 17, pp. 185-203, 1981.
[13] S.L. Horowitz and T. Pavlidis, “Picture Segmentation by a Tree Traversal Algorithm,” J. ACM, vol. 23, no. 2, pp. 368-388, 1976.
[14] M. Irani, B. Rousso, and S. Peleg, “Detecting and Tracking Multiple Moving Objects Using Temporal Integration,” Proc. European Conf. Computer Vision, pp. 282-287, May 1992.
[15] D. Le Gall, “MPEG: A Video Compression Standard for Multimedia Applications,” Comm. ACM, Apr. 1991.
[16] C.H. Lee and T. Huang,“Finding point correspondences and determining motion of a rigid objectfrom two weak perspective views,” Proc. IEEE Conf. Computer Vision and Pattern Recognition, (CVPR’88), pp. 398-403, 1988.
[17] V. Markandey and B.E. Flinchbaugh,“Multispectal constraints for optical flow computation, Proc. Third Int’l Conf. Computer Vision (ICCV’90), pp. 38-41, Dec. 1990.
[18] H.H. Nagel,“On change detection and displacement vector estimation in imagesequences,” Pattern Recognition Letters, vol. 1, pp. 55-59, 1982.
[19] H.H. Nagel,“Displacement vectors derived from second order intensity variations inimage sequences,” Computer Vision, Graphics, and Image Processing, pp. 85-117, 1983.
[20] H. H. Nagel,W. Enkelmann,“An Investigation of smoothness constraints for the estimation of displacement vector fields from image sequences,” Trans. Pattern Analysis and Machine Intelligence, vol. 8, no. 5, pp. 565-593, 1986
[21] J. Pearl, Heuristics: Intelligent Search Strategies for Computer Problem Solving, Addison-Wesley, Reading, Mass., 1984.
[22] A. Pentland,“Photometric motion,” Proc. Third Int’l Conf. Computer Vision (ICCV’90), pp. 178-187, Dec. 1990.
[23] J.R. Quinlan,"Induction of decision trees," Machine Learning, vol. 1, pp. 81-106, 1986.
[24] J. Rissanen, Universal Coding, Information, Prediction and Estimation IEEE Trans. Information Theory, vol. 30, no. 4, pp. 629-636, 1984.
[25] H. Samet, The Design and Analysis of Spatial Data Structures. Addison-Wesley, 1990.
[26] R. Srinivasan and R. Rao,“Predictive coding based on efficient motion estimation,” IEEE Trans. Comm., vol. 33, no. 8, pp. 888-896, Aug. 1985.
[27] L. Stockmeyer,“Optimal orientations of cells in slicing floorplan designs,” Information and Control, vol. 57, pp. 91-101, 1983.
[28] J. Weng,N. Ahuja,, and T. S. Huang,“Matching two perspective views,” Trans. Pattern Analysis and Machine Intelligence Intell., vol. 14, no. 8, pp. 806-825, 1992.

Index Terms:
Image motion, optic flow, video compression, machine learning, Occam algorithms.
Haim (Shvaytser) Schweitzer, "Occam Algorithms for Computing Visual Motion," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 17, no. 11, pp. 1033-1042, Nov. 1995, doi:10.1109/34.473229
Usage of this product signifies your acceptance of the Terms of Use.