A Time-Efficient Optimisation Framework for Parameters of Optical Flow Methods

  • Michael Stoll
  • Sebastian Volz
  • Daniel Maurer
  • Andrés Bruhn
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 10269)

Abstract

Due to the increase of optical flow benchmark data, concerning both amount and resolution, learning parameters from training sequences with ground truth has become significantly more challenging in recent years. Moreover, optical flow methods are much more complex than a few years ago resulting in a larger amount of model parameters and a noticeably increased runtime. As a consequence, even optimising a small set of suitable parameters may take hours or even days which makes hand tuning infeasible. Hence, time-efficient strategies for automatic parameter optimisation become more and more important. In this context, our work addresses three important aspects. First, we provide an overview of different optimisation strategies and juxtapose them in the context of different optical flow methods and different evaluation benchmarks. Second, we focus on choosing a suitable subset of the training data to speed up the computation while still obtaining meaningful results. Finally, we also consider different strategies for distributing the evaluation on hardware infrastructures which allows to further reduce the run time. Experiments show that the proposed methodology allows to obtain good results while keeping the overall effort reasonably low.

Keywords

Performance evaluation Parameter optimisation Distributed optimisation Adaptive scheduling Optical flow 

References

  1. 1.
    Baker, S., Roth, S., Scharstein, D., Black, M.J., Lewis, J.P., Szeliski, R.: A database and evaluation methodology for optical flow. In: Proceedings of IEEE International Conference on Computer Vision (ICCV). IEEE Computer Society Press (2007)Google Scholar
  2. 2.
    Barron, J.L., Fleet, D.J., Beauchemin, S.S.: Performance of optical flow techniques. Int. J. Comput. Vis. 12(1), 43–77 (1994)CrossRefGoogle Scholar
  3. 3.
    Boyd, S., Parikh, N., Chu, E., Peleato, B., Eckstein, J.: Distributed optimization and statistical learning via the alternating direction method of multipliers. Found. Trends Mach. Learn. 3(1), 1–122 (2011)CrossRefMATHGoogle Scholar
  4. 4.
    Braun, T.D., Siegel, H.J., Beck, N., Bölöni, L.L., Maheswaran, M., Reuther, A.I., Robertson, J.P., Theys, M.D., Yao, B., Hensgen, D., Freund, R.F.: A comparison of eleven static heuristics for mapping a class of independent tasks onto heterogeneous distributed computing systems. J. Parallel Distrib. Comput. 61(6), 810–837 (2001)CrossRefMATHGoogle Scholar
  5. 5.
    Brox, T., Malik, J.: Large displacement optical flow: descriptor matching in variational motion estimation. IEEE Trans. Pattern Anal. Mach. Intell. 33(3), 500–513 (2011)CrossRefGoogle Scholar
  6. 6.
    Butler, D.J., Wulff, J., Stanley, G.B., Black, M.J.: A naturalistic open source movie for optical flow evaluation. In: Fitzgibbon, A., Lazebnik, S., Perona, P., Sato, Y., Schmid, C. (eds.) ECCV 2012. LNCS, vol. 7577, pp. 611–625. Springer, Heidelberg (2012). doi:10.1007/978-3-642-33783-3_44 CrossRefGoogle Scholar
  7. 7.
    Datta, K., Murphy, M., Volkov, V., Williams, S., Carter, J., Oliker, L., Patterson, D., Shalf, J., Yelick, K.: Stencil computation optimization and auto-tuning on state-of-the-art multicore architectures. In: Proceedings of the 2008 ACM/IEEE Conference on Supercomputing, pp. 4:1–4:12. IEEE Press (2008)Google Scholar
  8. 8.
    Demetz, O.: Feature Invariance versus Change Estimation in Variational Motion Estimation. Ph.D. Thesis, Faculty of Mathematics and Computer Science, Saarland University (2015)Google Scholar
  9. 9.
    Drayer, B., Brox,T.: Combinatorial regularization of descriptor matching for optical flow estimation. In: British Machine Vision Conference (BMVC). BMVA Press (2015)Google Scholar
  10. 10.
    Geiger, A., Lenz, P., Urtasun, R.: Are we ready for autonomous driving? The KITTI vision benchmark suite. In: Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 3354–3361. IEEE Computer Society Press (2012)Google Scholar
  11. 11.
    Hansen, N., Ostermeier, A.: Completely derandomized self-adaptation in evolution strategies. Evol. Comput. 9(2), 159–195 (2001)CrossRefGoogle Scholar
  12. 12.
    Krajsek, K., Mester, R.: A maximum likelihood estimator for choosing the regularization parameters in global optical flow methods. In: Proceedings of IEEE International Conference on Image Processing (ICIP), pp. 1081–1084. IEEE Computer Society (2006)Google Scholar
  13. 13.
    Kulkarni, T., Kohli, P., Tenenbaum, J.B., Mansinghka, V.: Picture: a probabilistic programming language for scene perception. In: Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 4390–4399. IEEE Computer Society Press (2015)Google Scholar
  14. 14.
    Kunisch, K., Pock, T.: A bilevel optimization approach for parameter learning in variational models. SIAM J. Imaging Sci. 6(2), 938–983 (2013)MathSciNetCrossRefMATHGoogle Scholar
  15. 15.
    Li, Y., Huttenlocher, D.P.: Learning for optical flow using stochastic optimization. In: Forsyth, D., Torr, P., Zisserman, A. (eds.) ECCV 2008. LNCS, vol. 5303, pp. 379–391. Springer, Heidelberg (2008). doi:10.1007/978-3-540-88688-4_28 CrossRefGoogle Scholar
  16. 16.
    Nelder, J.A., Mead, R.: A simplex method for function minimization. Comput. J. 7(4), 308–313 (1965)MathSciNetCrossRefMATHGoogle Scholar
  17. 17.
    Memin, E., Heas, P., Herzet, C.: Bayesian inference of models and hyper-parameters for robust optic-flow estimation. IEEE Trans. Image Process. 21(4), 1437–1451 (2012)MathSciNetCrossRefGoogle Scholar
  18. 18.
    Perreira, D.R., Delpiano, J., Papa, J.P.: On the optical flow model selection through metaheuristics. EURASIP J. Image Video Process. 2015, 11 (2015)CrossRefGoogle Scholar
  19. 19.
    Ragan-Kelley, J., Barnes, C., Adams, A., Paris, S., Durand, F., Amarasinghe, S.: Halide: a language and compiler for optimizing parallelism, locality, and recomputation in image processing pipelines. In: Proceedings of the 34th ACM SIGPLAN Conference on Programming Language Design and Implementation, pp. 519–530. ACM (2013)Google Scholar
  20. 20.
    Salmen, J., Caup, L., Igel, C.: Real-time estimation of optical flow based on optimized haar wavelet features. In: Takahashi, R.H.C., Deb, K., Wanner, E.F., Greco, S. (eds.) EMO 2011. LNCS, vol. 6576, pp. 448–461. Springer, Heidelberg (2011). doi:10.1007/978-3-642-19893-9_31 CrossRefGoogle Scholar
  21. 21.
    Samuel, K.G.G., Tappen, M.F.: Learning optimized map estimates in continuously-valued MRF models. In: Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 477–484. IEEE Computer Society Press (2009)Google Scholar
  22. 22.
    Sun, D., Roth, S., Lewis, J.P., Black, M.J.: Learning optical flow. In: Forsyth, D., Torr, P., Zisserman, A. (eds.) ECCV 2008. LNCS, vol. 5304, pp. 83–97. Springer, Heidelberg (2008). doi:10.1007/978-3-540-88690-7_7 CrossRefGoogle Scholar
  23. 23.
    Sun, D., Sudderth, E.B., Black, M.J.: Layered segmentation and optical flow estimation over time. In: Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1768–1775. IEEE Computer Society Press (2012)Google Scholar
  24. 24.
    Sun, D., Roth, S., Black, M.J.: A quantitative analysis of current practices in optical flow estimation and the principles behind them. Int. J. Comput. Vis. 106(2), 115–137 (2013)CrossRefGoogle Scholar
  25. 25.
    Volz, S., Bruhn, A., Valgaerts, L., Zimmer, H.: Modeling temporal coherence for optical flow. In: Proceedings of IEEE International Conference on Computer Vision (ICCV), pp. 1116–1123. IEEE Computer Society Press (2011)Google Scholar
  26. 26.
    Weinzaepfel, P., Revaud, J., Harchaoui, Z., Schmid, C.: DeepFlow: large displacement optical flow with deep matching. In: Proceedings of IEEE International Conference on Computer Vision (ICCV), pp. 1385–1392. IEEE Computer Society Press (2013)Google Scholar

Copyright information

© Springer International Publishing AG 2017

Authors and Affiliations

  • Michael Stoll
    • 1
  • Sebastian Volz
    • 1
  • Daniel Maurer
    • 1
  • Andrés Bruhn
    • 1
  1. 1.Institute for Visualization and Interactive SystemsUniversity of StuttgartStuttgartGermany

Personalised recommendations