Advertisement

Accuracy vs. efficiency trade-offs in optical flow algorithms

  • Hongche Liu
  • Tsai-Hong Hong
  • Martin Herman
  • Rama Chellappa
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1065)

Abstract

There have been two thrusts in the development of optical flow algorithms. One has emphasized higher accuracy; the other faster implementation. These two thrusts, however, have been independently pursued, without addressing the accuracy vs. efficiency trade-offs. Although the accuracy-efficiency characteristic is algorithm dependent, an understanding of a general pattern is crucial in evaluating an algorithm as far as real world tasks are concerned, which often pose various performance requirements. This paper addresses many implementation issues that have often been neglected in previous research, including subsampling, temporal filtering of the output stream, algorithms' flexibility and robustness, etc. Their impacts on accuracy and/or efficiency are emphasized. We present a critical survey of different approaches toward the goal of higher performance and present experimental studies on accuracy vs. efficiency trade-offs. The goal of this paper is to bridge the gap between the accuracy and the efficiency-oriented approaches.

Keywords

Optical Flow Motion Estimation Obstacle Avoidance Noise Sensitivity Output Density 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

References

  1. [1]
    Allen, P., Timcenko, A., Yoshimi, B. and Michelman, P., “Automated Tracking and Grasping of a Moving Object with a Robotic Hand-Eye System”, IEEE Transactions on Robotics and Automations, vol. 9, no. 2, pp. 152–165, 1993.Google Scholar
  2. [2]
    Ancona, N. and Poggio, T., “Optical Flow from 1-D Correlation: Applications to a Simple Time-to-Crash Detector”, Proceedings of IEEE International Conference on Computer Vision, Berlin, Germany, pp. 209–214, 1993.Google Scholar
  3. [3]
    Artieri, A. and Jutand, F., “A Versatile and Powerful Chip for Real-Time Motion Estimation”, Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing, Glasgow, UK, pp. 2453–2456, 1989.Google Scholar
  4. [4]
    Barron, J. L., Fleet, D. J. and Beauchemin, S. S., “Performance of Optical Flow Techniques”, International Journal of Computer Vision, vol. 12, no. 1, pp. 43–77, 1994.Google Scholar
  5. [5]
    Bober, M. and Kittler, J., “Robust Motion Analysis”, Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, Seattle, WA, pp. 947–952, 1994.Google Scholar
  6. [6]
    Bülthoff, H., Little, J. and Poggio, T., “A Parallel Algorithm for Real-Time Computation of Optical Flow”, Nature, vol. 337, pp. 549–553, 1989.Google Scholar
  7. [7]
    Camus, T., “Calculating Time-to-Contact Using Real-Time Quantized Optical Flow”, NISTIR-5609, Gaithersburg, MD, 1995.Google Scholar
  8. [8]
    Camus, T., “Real-Time Quantized Optical Flow”, Proceedings of IEEE Conference on Computer Architectures for Machine Perception, Como, Italy, 1995.Google Scholar
  9. [9]
    Coombs, D., Herman M., Hong T. and Nashman, M., “Real-time Obstacle Avoidance Using Central Flow Divergence and Peripheral Flow”, Proceedings of IEEE International Conference on Computer Vision, Cambridge, MA, 1995.Google Scholar
  10. [10]
    Delbruck, T., “Silicon Retina with Correlation-Based Velocity-Tuned Pixels”, IEEE Transactions on Neural Networks, vol. 4, no. 3, pp. 529–541, 1993.Google Scholar
  11. [11]
    Ens, J. and Li, Z.N., “Real-Time Motion Stereo”, Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, pp.130–135, New York, NY, 1993.Google Scholar
  12. [12]
    Etienne-Cummings, R., Fernando, S.A., Van der Spiegel, J. and Mueller, P., “Real-Time 2D Analog Motion Detector VLSI Circuit”, Proceedings of IEEE International Joint Conference on Neural Networks, vol. 4, pp. 426–431, New York, NY, 1992.Google Scholar
  13. [13]
    Fleet, D.J. and Jepson, A.L., “Computation of Component Image Velocity from Local Phase Information”, International Journal of Computer Vision, vol. 5, no.1, pp. 77–104, 1990.Google Scholar
  14. [14]
    Heeger, D. J., “Optical How Using Spatiotemporal Filters”, International Journal of Computer Vision, vol. 1, no. 4, pp. 279–302, 1988.Google Scholar
  15. [15]
    Inoue, H., Tachikawa, T. and Inaba, M., “Robot Vision System with a Correlation Chip for Real-Time Tracking, Optical How and Depth Map Generation”, Proceedings IEEE Conference on Robotics and Automation, Nice, France, vol. 2, pp. 1621–1626, 1992.Google Scholar
  16. [16]
    Koch, C., Marroquin, J. and Yuille, A., “Analog ‘Neural’ Networks in Early Vision”, Proceedings of the National Academy of Sciences, vol. 83, pp. 4263–4267, 1986.Google Scholar
  17. [17]
    Lee, J.C., Sheu, B. J., Fang, W.C. and Chellappa, R., “VLSI neuroprocessors for Video Motion Detection”, IEEE Transactions on Neural Networks, vol. 4, no. 2, pp. 178–191, 1993.Google Scholar
  18. [18]
    Liu, H., Hong, T., Herman, M. and Chellappa, R., “A General Motion Model and Spatio-temporal Filters for Computing Optical Flow”, University of Maryland TR-3365, November, 1994; NIST-IR 5539, Gaithersburg MD, January, 1995, to appear in International Journal of Computer Vision.Google Scholar
  19. [19]
    Liu, H., “A General Motion Model and Spatio-temporal Filters for 3-D Motion Interpretations”, PH.D. Dissertation, University of Maryland, September, 1995.Google Scholar
  20. [20]
    Matteucci, P., Regazzani, C.S. and Foresti, G.L., “Real-Time Approach to 3-D Object Tracking in Complex Scenes”, Electronics Letters, vol. 30, no. 6, pp. 475–477, 1994.Google Scholar
  21. [21]
    Mioni, A., “VLSI Vision Chips Homepage”, http://www.eleceng.adelaide.edu.au/Groups/GAAS/Bugeye/visionchips.Google Scholar
  22. [22]
    Mioni, A., Bouzerdoum, A., Yakovleff, A., Abbott, D., Kim, O., Eshraghian, K. and Bogner, R.E., “An Analog Implementation of Early Visual Processing in Insects”, Proceedings International Symposium on VLSI Technology, Systems, and Applications, pp. 283–287, 1993.Google Scholar
  23. [23]
    Moore, A. and Koch C., “A Multiplication Based Analog Motion Detection Chip”, Proceedings of the SPIE, vol. 1473, Visual Information Processing: From Neurons to Chips, pp. 66–75, 1991.Google Scholar
  24. [24]
    Nashman, M., Rippey, W., Hong, T.-H. and Herman, M., “An Integrated Vision Touch-Probe System for Dimensional Inspection Tasks”, NISTIR 5678, National Institute of Standards and Technology, 1995.Google Scholar
  25. [25]
    Nelson, R., “Qualitative Detection of Motion by a Moving Observer”, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Lahaina, HI, pp. 173–178, 1991.Google Scholar
  26. [26]
    Nesi, P., DelBimbo, A.D. and Ben-Tzvi, D. “A Robust Algorithm for Optical Flow Estimation”, Computer Vision and Image Understanding, vol. 62, no. 1, pp. 59–68, 1995.Google Scholar
  27. [27]
    Nishihara, H.K., “Real-Time Stereo-and Motion-Based Figure Ground Discrimination and Tracking Using LOG Sign Correlation”, Conference Record of the Twenty-Seventh Asilomar Conference on Signals, Systems and Computers, vol. 1, pp. 95–100, 1993.Google Scholar
  28. [28]
    Patterson, D. and Hennessy, J., “Computer Organization and Design: the Hardware/Software Interface”, Morgan Kaufman, San Mateo, CA 1994.Google Scholar
  29. [29]
    Rangachar, R., Hong, T.-H., Herman, M., and Luck, R., “Three Dimensional Reconstruction from Optical Flow Using Temporal Integration”, Proceedings of SPIE Advances in Intelligent Robotic Systems: Intelligent Robots and Computer Vision, Boston, MA, 1990.Google Scholar
  30. [30]
    Schunck, B., “Image Flow Segmentation and Estimation by Constraint Line Clustering”, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 11, no.10, pp. 1010–1027, 1989.Google Scholar
  31. [31]
    Singh, A., “Optical Flow Computation: A Unified Perspective”, IEEE Computer Society Press, 1991.Google Scholar
  32. [32]
    Smith, S.M., “ASSET-2 Real-Time Motion Segmentation and Object Tracking”, Proceedings of the Fifth International Conference on Computer Vision, pp. 237–244, Cambridge, MA, 1995.Google Scholar
  33. [33]
    Tanner J. and Mead C., “A Correlating Optical Motion Detector”, MIT Advanced Research in VLSI, pp. 57–64, 1984.Google Scholar
  34. [34]
    Tanner, J. and Mead, C., “An Integrated Analog Optical Motion Sensor”, VLSI Signal Processing II, R.W. Brodersen and H.S. Moscovitz, Eds., pp. 59–87, New York, 1988.Google Scholar
  35. [35]
    Waxman, A.M., Wu J. and Bergholm F. “Convected Activation Profiles and Receptive Fields for Real Time Measurement of Short Range Visual Motion”, Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, Ann Arbor, MI, pp. 717–723, 1988.Google Scholar
  36. [36]
    Weber, J. and Malik, J., “Robust Computation of Optical Flow in a Multi-Scale Differential Framework”, Proceedings of the Fourth International Conference on Computer Vision, Berlin, Germany, 1993.Google Scholar
  37. [37]
    Weber, J. and Malik, J., “Robust Computation of Optical Flow in a Multi-Scale Differential Framework”, International Journal of Computer Vision, vol. 14, pp. 67–81, 1995.Google Scholar
  38. [38]
    Weiss, P. and Christensson, B., “Real Time Implementation of Subpixel Motion Estimation for Broadcast Applications”, IEE Colloquim on Applications of Motion Compensation, London, UK, no. 128, pp. 7/1–3, 1990.Google Scholar
  39. [39]
    Woodfill, J. and Zabin, R., “An Algorithm for Real-Time Tracking of Non-Rigid Objects”, Proceedings of the Ninth National Conference on Artificial Intelligence, pp. 718–723, 1991.Google Scholar
  40. [40]
    Yagi, Y., Kawato, S. and Tsuji, S., “Real-Time Omnidirectional Image Sensor (COPIS) for Vision-Guided Navigation”, IEEE Transactions on Robotics and Automation, vol. 10, no. 1, pp. 11–22, 1994.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1996

Authors and Affiliations

  • Hongche Liu
    • 1
    • 2
  • Tsai-Hong Hong
    • 1
  • Martin Herman
    • 1
  • Rama Chellappa
    • 2
  1. 1.Intelligent Systems DivisionNational Institute of Standards and Technology (NIST)Gaithersburg
  2. 2.Center for Automation Research/Department of Electrical EngineeringUniversity of MarylandCollege Park

Personalised recommendations