International Journal of Computer Vision

, Volume 118, Issue 3, pp 275–299 | Cite as

Cyclic Schemes for PDE-Based Image Analysis

  • Joachim Weickert
  • Sven Grewenig
  • Christopher Schroers
  • Andrés Bruhn
Article

Abstract

We investigate a class of efficient numerical algorithms for many partial differential equations (PDEs) in image analysis. They are applicable to parabolic or elliptic PDEs that have bounded coefficients and lead to space discretisations with symmetric matrices. Our schemes are easy to implement and well-suited for parallel implementations on GPUs, since they are based on the explicit diffusion scheme in the parabolic case, and the Jacobi method in the elliptic case. By supplementing these methods with cyclically varying time step sizes or relaxation parameters, we achieve efficiency gains of several orders of magnitude. We call the resulting algorithms Fast Explicit Diffusion (FED) and Fast Jacobi (FJ) methods. To achieve a good compromise between efficiency and accuracy, we show that one should use parameter cycles that result from factorisations of box filters. For these cycles we establish stability results in the Euclidean norm. Our schemes perform favourably in a number of applications, including isotropic nonlinear diffusion filters with widely varying diffusivities as well as anisotropic diffusion methods for image filtering, inpainting, and regularisation in computer vision. Moreover, they are equally suited for higher dimensional problems as well as higher order PDEs, and they can also be interpreted as efficient first order methods for smooth optimisation problems.

Keywords

Explicit scheme Jacobi method Partial differential equations (PDEs) PDE-based image analysis Fast explicit diffusion Super time stepping First order optimisation methods 

References

  1. Abramowitz, M., & Stegun, I. A. (Eds.). (1972). Orthogonal polynomials (Chapter 22). Handbook of mathematical functions with formulas, graphs, and mathematical tables (9th printing) (pp. 771–802). New York: Dover.Google Scholar
  2. Alcantarilla, P. F., & Bartoli, A. (2013). Fast explicit diffusion for accelerated features in nonlinear scale spaces. In T. Burghardt, D. Damen, W. Mayol-Cuevas, & M. Mirmehdi (Eds.), Proceedings of 2013 British Machine Vision Conference (pp. 13.1–13.11). Bristol: BMVA Press.CrossRefGoogle Scholar
  3. Alexiades, V. (1995). Overcoming the stability restriction of explicit schemes via super-time-stepping. Proceedings of Dynamic Systems and Applications, Atlanta, Georgia, 2, 39–44.MathSciNetMATHGoogle Scholar
  4. Alexiades, V., Amiez, G., & Gremaud, P. A. (1996). Super-time-stepping acceleration of explicit schemes for parabolic problems. Communications in Numerical Methods in Engineering, 12, 31–42.CrossRefMATHGoogle Scholar
  5. Anderssen, R.S., & Golub, G. H. (1972). Richardson’s non-stationary matrix iterative procedure. Technical Report, STAN-CS-72-304, Computer Science Department, Stanford University.Google Scholar
  6. Bänsch, E., & Mikula, K. (1997). A coarsening finite element strategy in image selective smoothing. Computation and Visualization in Science, 1, 53–61.CrossRefMATHGoogle Scholar
  7. Ben-Ari, R., Raveh, G. (2011). Variational depth from defocus in real-time. In Computer Vision Workshops, 2011 IEEE International Conference on Computer Vision (pp. 522–529).Google Scholar
  8. Bornemann, F., & Deuflhard, P. (1996). The cascadic multigrid method for elliptic problems. Numerische Mathematik, 75, 135–152.MathSciNetCrossRefMATHGoogle Scholar
  9. Boyd, S., & Vandenberghe, L. (2004). Convex optimization. Cambridge, UK: Cambridge University Press.CrossRefMATHGoogle Scholar
  10. Brakhage, H. (1960). Über die numerische Behandlung von Integralgleichungen nach der Quadraturformelmethode. Numerische Mathematik, 2, 183–196.MathSciNetCrossRefMATHGoogle Scholar
  11. Brandt, A. (1977). Multi-level adaptive solutions to boundary-value problems. Mathematics of Computation, 31(138), 333–390.MathSciNetCrossRefMATHGoogle Scholar
  12. Bruhn, A., Weickert, J., Kohlberger, T., & Schnörr, C. (2006). A multigrid platform for real-time motion computation with discontinuity-preserving variational methods. International Journal of Computer Vision, 70(3), 257–277.CrossRefGoogle Scholar
  13. Calvetti, D., & Reichel, L. (1996). Adaptive Richardson iteration based on Leja points. Journal of Computational and Applied Mathematics, 71, 267–286.MathSciNetCrossRefMATHGoogle Scholar
  14. Calvetti, D., & Reichel, L. (2003). On the evaluation of polynomial coefficients. Numerical Algorithms, 33(1–4), 153–161.MathSciNetCrossRefMATHGoogle Scholar
  15. Catté, F., Lions, P. L., Morel, J. M., & Coll, T. (1992). Image selective smoothing and edge detection by nonlinear diffusion. SIAM Journal on Numerical Analysis, 32, 1895–1909.CrossRefMATHGoogle Scholar
  16. Chambolle, A. (2004). An algorithm for total variation minimization and applications. Journal of Mathematical Imaging and Vision, 20(1–2), 89–97.MathSciNetGoogle Scholar
  17. Chambolle, A., & Pock, T. (2011). A first-order primal-dual algorithm for convex problems with applications to imaging. Journal of Mathematical Imaging and Vision, 40(1), 120–145.MathSciNetCrossRefMATHGoogle Scholar
  18. Charbonnier, P., Blanc-Féraud, L., Aubert, G., & Barlaud, M. (1994). Two deterministic half-quadratic regularization algorithms for computed imaging. Proceedings of 1994 IEEE International Conference on Image Processing (Vol. 2, pp. 168–172). Austin: IEEE Computer Society Press.Google Scholar
  19. Crank, J., & Nicolson, P. (1947). A practical method for numerical evaluation of solutions of partial differential equations of the heat-conduction type. Proceedings of the Cambridge Philosophical Society, 43, 50–67.MathSciNetCrossRefMATHGoogle Scholar
  20. Drori, Y., & Teboulle, M. (2014). Performance of first-order methods for smooth convex minimization: A novel approach. Mathematical Programming, 145(1–2), 454–482.MathSciNetMATHGoogle Scholar
  21. Fedorenko, R. P. (1962). A relaxation method for solving elliptic difference equations. USSR Computational Mathematics and Mathematical Physics, 1(4), 1092–1096.CrossRefMATHGoogle Scholar
  22. Frankel, S. (1950). Convergence rates of iterative treatments of partial differential equations. Mathematical Tables and Other Aids to Computation, 4, 65–75.MathSciNetCrossRefGoogle Scholar
  23. Galić, I., Weickert, J., Welk, M., Bruhn, A., Belyaev, A., & Seidel, H. P. (2008). Image compression with anisotropic diffusion. Journal of Mathematical Imaging and Vision, 31(2–3), 255–269.MathSciNetMATHGoogle Scholar
  24. Gentzsch, W. (1979). Numerical solution of linear and non-linear parabolic differential equations by a time discretisation of third order accuracy. In E. H. Hirschel (Ed.), Proceedings of the Third GAMM-Conference on Numerical Methods in Fluid Mechanics. Brunswick: Friedr. Vieweg & Sohn.Google Scholar
  25. Gentzsch, W., & Schlüter, A. (1978). Über ein Einschrittverfahren mit zyklischer Schrittweitenänderung zur Lösung parabolischer Differentialgleichungen. Zeitschrift für Angewandte Mathematik und Mechanik, 58, T415–T416. (in German).MATHGoogle Scholar
  26. Gordeziani, D. G., & Meladze, G. V. (1974). Simulation of the third boundary value problem for multidimensional parabolic equations in an arbitrary domain by one-dimensional equations. USSR Computational Mathematics and Mathematical Physics, 14(1), 249–253.MathSciNetCrossRefMATHGoogle Scholar
  27. Grewenig, S., Weickert, J., & Bruhn, A. (2010). From box filtering to fast explicit diffusion. In M. Goesele, S. Roth, A. Kuijper, B. Schiele, & K. Schindler (Eds.), Pattern recognition (Vol. 6376, pp. 543–552)., Lecture Notes in Computer Science Berlin: Springer.CrossRefGoogle Scholar
  28. Gurski, K. F., O’Sullivan, S. (2010). An explicit super-time-stepping scheme for non-symmetric parabolic problems. In AIP Conference Proceedings: International Conference of Numerical Analysis and Applied Mathematics, Rhodes (Greece) (Vol. 1281, pp. 761–764).Google Scholar
  29. Gwosdek, P., Zimmer, H., Grewenig, S., Bruhn, A., & Weickert, J. (2012). A highly efficient GPU implementation for variational optic flow based on the Euler–Lagrange framework. In E. E. Kutulakos (Ed.), Trends and topics in computer vision (Vol. 6554, pp. 372–383)., Lecture Notes in Computer Science Berlin, Heidelberg: Springer.CrossRefGoogle Scholar
  30. Hackbusch, W. (1985). Multigrid methods and applications. New York: Springer.CrossRefMATHGoogle Scholar
  31. Hellwig, G. (1977). Partial differential equations. Stuttgart: Teubner.MATHGoogle Scholar
  32. Hoffmann, S., Mainberger, M., Weickert, J., & Puhl, M. (2013). Compression of depth maps with segment-based homogeneous diffusion. In A. Kuijper, K. Bredies, T. Pock, & H. Bischof (Eds.), Scale space and variational methods in computer vision (Vol. 7893, pp. 319–330). Berlin: Springer.CrossRefGoogle Scholar
  33. Jawerth, B., Lin, P., & Sinzinger, E. (1999). Lattice Boltzmann models for anisotropic diffusion of images. Journal of Mathematical Imaging and Vision, 11, 231–237.MathSciNetCrossRefMATHGoogle Scholar
  34. Krivá, Z., & Mikula, K. (2002). An adaptive finite volume scheme for solving nonlinear diffusion equations in image processing. Journal of Visual Communication and Image Representation, 13(1/2), 22–35.CrossRefGoogle Scholar
  35. Laasonen, P. (1949). Über eine Methode zur Lösung der Wärmeleitungsgleichung. Acta Mathematica, 81, 309–317.MathSciNetCrossRefMATHGoogle Scholar
  36. Lebedev, V. I., & Finogenov, V. N. (1971). Ordering the iteration parameters in the cyclic Chebychev iterative method. USSR Computational Mathematics and Mathematical Physics, 11(2), 155–170.MathSciNetCrossRefMATHGoogle Scholar
  37. Lu, T., Neittaanmäki, P., & Tai, X. C. (1991). A parallel splitting up method and its application to Navier–Stokes equations. Applied Mathematics Letters, 4(2), 25–29.Google Scholar
  38. Luxenburger, A., Zimmer, H., Gwosdek, P., & Weickert, J. (2011). Fast PDE-based image analysis in your pocket. In A. M. Bruckstein, B. ter Haar Romeny, A. M. Bronstein, & M. M. Bronstein (Eds.), Scale Space and Variational Methods in Computer Vision (Vol. 6667, pp. 544–555)., Lecture Notes in Computer Science Berlin: Springer.CrossRefGoogle Scholar
  39. Mang, A., Schuetz, T. A., Becker, S., Toma, A., & Buzug, T. M. (2012). Cyclic numerical time integration in variational non-rigid image registration based on quadratic regularisation. In Proceedings of Vision, Modeling, and Visualization (2012). Eurographics Digital Library (pp. 143–150). Germany: Magdeburg.Google Scholar
  40. Meijerink, J. A., & van der Vorst, H. A. (1977). An iterative solution method for linear systems of which the coefficient matrix is a symmetric \(M\)-matrix. Mathematics of Computation, 31(137), 148–162.MathSciNetMATHGoogle Scholar
  41. Nesterov, Y. (2004). Introductory lectures on convex optimization: A basic course, applied optimization (Vol. 87). Boston: Kluwer.MATHGoogle Scholar
  42. Nocedal, J., & Wright, S. J. (2006). Numerical optimization. New York: Springer.MATHGoogle Scholar
  43. Ochs, P., Brox, T., & Pock, T. (2015). iPiasco: Inertial proximal algorithm for strongly convex optimization. Journal of Mathematical Imaging and Vision, 53(2), 171–181.MathSciNetCrossRefMATHGoogle Scholar
  44. Opfer, G., & Schober, G. (1984). Richardson’s iteration for nonsymmetric matrices. Linear Algebra and its Applications, 58, 343–361.MathSciNetCrossRefMATHGoogle Scholar
  45. Peaceman, D. W., & Rachford, H. H, Jr. (1955). The numerical solution of parabolic and elliptic differential equations. Journal of the Society for Industrial and Applied Mathematics, 3(1), 28–41.MathSciNetCrossRefMATHGoogle Scholar
  46. Perona, P., & Malik, J. (1990). Scale space and edge detection using anisotropic diffusion. IEEE Transactions on Pattern Analysis and Machine Intelligence, 12, 629–639.CrossRefGoogle Scholar
  47. Peter, P. (2013). Three-dimensional data compression with anisotropic diffusion. In J. Weickert, M. Hein, & B. Schiele (Eds.), Pattern recognition (Vol. 8142, pp. 231–236)., Lecture Notes in Computer Science Berlin: Springer.CrossRefGoogle Scholar
  48. Pock, T., Chambolle, A. (2011). Diagonal preconditioning for first order primal-dual algorithms in convex optimization. In Proceedings 13th International Conference on Computer Vision, Barcelona (pp. 1762–1769).Google Scholar
  49. Rakêt, L. L., & Markussen, B. (2014). Approximate inference for spatial functional data on massively parallel processors. Computational Statistics and Data Analysis, 72, 1723–1730.MathSciNetCrossRefGoogle Scholar
  50. Reichel, L. (1990). Newton interpolation at Leja points. BIT Numerical Mathematics, 30(2), 332–346.MathSciNetCrossRefMATHGoogle Scholar
  51. Richardson, L. F. (1910). The approximate arithmetical solution by finite differences of physical problems involving differential equation, with an application to the stresses in a masonry dam. Transactions of the Royal Society of London Series, A(210), 307–357.Google Scholar
  52. Rosman, G., Dascal, L., Sidi, A., & Kimmel, R. (2009). Efficient Beltrami image filtering via vector extrapolation methods. SIAM Journal on Imaging Sciences, 2(3), 858–878.MathSciNetCrossRefMATHGoogle Scholar
  53. Rudin, L. I., Osher, S., & Fatemi, E. (1992). Nonlinear total variation based noise removal algorithms. Physica D, 60, 259–268.MathSciNetCrossRefMATHGoogle Scholar
  54. Saad, Y. (2003). Iterative Methods for Sparse Linear Systems (2nd ed.). Philadelphia: SIAM.Google Scholar
  55. Saul’yev, V. K. (1964). Integration of equations of parabolic type by the method of nets. Oxford: Pergamon.MATHGoogle Scholar
  56. Schmidt-Richberg, A., Ehrhardt, J., Werner, R., & Handels, H. (2012). Fast explicit diffusion for registration with direction-dependent regularization. In B. M. Dawant, G. E. Christensen, J. M. Fitzpatrick, & D. Rueckert (Eds.), Biomedical image registration (Vol. 7359, pp. 220–228)., Lecture Notes in Computer Science Berlin, Heidelberg: Springer.CrossRefGoogle Scholar
  57. Schroers, C., Zimmer, H., Valgaerts, L., Bruhn, A., Demetz, O., & Weickert, J. (2012). Anisotropic range image integration. In A. Prinz, T. Pock, H. Bischof, & F. Leberl (Eds.), Pattern recognition (Vol. 7476, pp. 73–82)., Lecture Notes in Computer Science Berlin: Springer.CrossRefGoogle Scholar
  58. Setzer, S., Steidl, G., & Morgenthaler, J. (2013). On cyclic gradient descent reprojection. Computational Optimization and Applications, 54(2), 417–440.MathSciNetCrossRefMATHGoogle Scholar
  59. Spira, A., Kimmel, R., & Sochen, N. (2007). A short-time Beltrami kernel for smoothing images and manifolds. IEEE Transactions on Image Processing, 16(3), 1628–1636.MathSciNetCrossRefGoogle Scholar
  60. Varga, R. S. (1962). Matrix Iterative Analysis. Englewood Cliffs: Prentice Hall.Google Scholar
  61. Weickert, J. (1998). Anisotropic diffusion in image processing. Stuttgart: Teubner.MATHGoogle Scholar
  62. Weickert, J. (1999). Coherence-enhancing diffusion filtering. International Journal of Computer Vision, 31(2/3), 111–127.CrossRefGoogle Scholar
  63. Weickert, J. (2001). Applications of nonlinear diffusion in image processing and computer vision. Acta Mathematica Universitatis Comenianae, 70(1), 33–50.MathSciNetMATHGoogle Scholar
  64. Weickert, J., ter Haar Romeny, B. M., & Viergever, M. A. (1998). Efficient and reliable schemes for nonlinear diffusion filtering. IEEE Transactions on Image Processing, 7(3), 398–410.CrossRefGoogle Scholar
  65. Weickert, J., Hagenburg, K., Breuß, M., & Vogel, O. (2013). Lecture Notes in Computer Science. Energy minimisation methods in computer vision and pattern recognition (Vol. 8081, pp. 29–39). Berlin: Springer.Google Scholar
  66. Welk, M., Steidl, G., & Weickert, J. (2008). Locally analytic schemes: A link between diffusion filtering and wavelet shrinkage. Applied and Computational Harmonic Analysis, 24, 195–224.MathSciNetCrossRefMATHGoogle Scholar
  67. Wells, W. M. (1986). Efficient synthesis of Gaussian filters by cascaded uniform filters. IEEE Transactions on Pattern Analysis and Machine Intelligence, 8, 234–239.CrossRefGoogle Scholar
  68. Young, D. (1954). On Richardson’s method for solving linear systems with positive definite matrices. Journal of Mathematics and Physics, 32, 243–255.CrossRefMATHGoogle Scholar
  69. Young, D. M. (1950) Iterative methods for solving partial difference equations of elliptic type. PhD Thesis, Department of Mathematics, Harvard University, Cambridge, MA.Google Scholar
  70. Yuan’Chzhao-Din (1958). Some difference schemes for the solution of the first boundary value problem for linear differential equations with partial derivatives. PhD Thesis, Moscow State University (in Russian).Google Scholar

Copyright information

© Springer Science+Business Media New York 2015

Authors and Affiliations

  • Joachim Weickert
    • 1
  • Sven Grewenig
    • 1
  • Christopher Schroers
    • 1
  • Andrés Bruhn
    • 2
  1. 1.Mathematical Image Analysis Group, Department of Mathematics and Computer ScienceSaarland UniversitySaarbrückenGermany
  2. 2.Computer Vision and Intelligent Systems Group, Institute for Visualization and Interactive SystemsUniversity of StuttgartStuttgartGermany

Personalised recommendations