Low Rank Priors for Color Image Regularization

  • Thomas Möllenhoff
  • Evgeny Strekalovskiy
  • Michael Moeller
  • Daniel Cremers
Part of the Lecture Notes in Computer Science book series (LNCS, volume 8932)


In this work we consider the regularization of vectorial data such as color images. Based on the observation that edge alignment across image channels is a desirable prior for multichannel image restoration, we propose a novel scheme of minimizing the rank of the image Jacobian and extend this idea to second derivatives in the framework of total generalized variation. We compare the proposed convex and nonconvex relaxations of the rank function based on the Schatten-q norm to previous color image regularizers and show in our numerical experiments that they have several desirable properties. In particular, the nonconvex relaxations lead to better preservation of discontinuities. The efficient minimization of energies involving nonconvex and nonsmooth regularizers is still an important open question. We extend a recently proposed primal-dual splitting approach for nonconvex optimization and show that it can be effectively used to minimize such energies. Furthermore, we propose a novel algorithm for efficiently evaluating the proximal mapping of the ℓ q norm appearing during optimization. We experimentally verify convergence of the proposed optimization method and show that it performs comparably to sequential convex programming.


Color Channel Proximal Mapping Rank Minimization Nuclear Norm Proximal Operator 
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.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Rudin, L.I., Osher, S., Fatemi, E.: Nonlinear total variation based noise removal algorithms. Physica D 60, 259–268 (1992)CrossRefzbMATHGoogle Scholar
  2. 2.
    Chambolle, A., Caselles, V., Cremers, D., Novaga, M., Pock, T.: An introduction to total variation for image analysis. In: Theoretical Foundations and Numerical Methods for Sparse Recovery. De Gruyter (2010)Google Scholar
  3. 3.
    Attouch, H., Buttazzo, G., Michaille, G.: Variational Analysis in Sobolev and BV Spaces: Applications to PDEs and Optimization. Mps-Siam Series on Optimization, vol. 6. SIAM (2005)Google Scholar
  4. 4.
    Blomgren, P., Chan, T.F.: Color TV: Total variation methods for restoration of vector valued images. IEEE Trans. Image Processing 7, 304–309 (1998)CrossRefGoogle Scholar
  5. 5.
    Sapiro, G., Ringach, D.: Anisotropic diffusion of multivalued images with applications to color filtering. IEEE Trans. Img. Proc. 5(11), 1582–1586 (1996)CrossRefGoogle Scholar
  6. 6.
    Bresson, X., Chan, T.F.: Fast dual minimization of the vectorial total variation norm and applications to color image processing. Inverse Problems and Imaging 2(4), 255–284 (2008)CrossRefMathSciNetGoogle Scholar
  7. 7.
    Condat, C.: Joint demosaicking and denoising by total variation minimization. In: IEEE Conference on Image Processing, pp. 2781–2784 (2012)Google Scholar
  8. 8.
    Miyata, T., Sakai, Y.: Vectorized total variation defined by weighted l infinity norm for utilizing inter channel dependency. In: 2012 19th IEEE International Conference on Image Processing (ICIP), pp. 3057–3060 (September 2012)Google Scholar
  9. 9.
    Goldluecke, B., Cremers, D.: An approach to vectorial total variation based on geometric measure theory. In: IEEE Conference on Computer Vision and Pattern Recognition (2010)Google Scholar
  10. 10.
    Lefkimmiatis, S., Roussos, A., Unser, M., Maragos, P.: Convex generalizations of total variation based on the structure tensor with applications to inverse problems. In: Pack, T. (ed.) SSVM 2013. LNCS, vol. 7893, pp. 48–60. Springer, Heidelberg (2013)CrossRefGoogle Scholar
  11. 11.
    Ehrhardt, M.J., Arridge, S.: Vector-valued image processing by parallel level sets. IEEE Trans. on Image Processing 23, 9–18 (2014)CrossRefMathSciNetGoogle Scholar
  12. 12.
    Moeller, M., Brinkmann, E., Burger, M., Seybold, T.: Color bregman tv Preprint. On ArXiv,
  13. 13.
    Huang, J., Mumford, D.: Statistics of natural images and models. In: Int. Conf. on Computer Vision and Pattern Recognition (CVPR) (1999)Google Scholar
  14. 14.
    Krishnan, D., Fergus, R.: Fast Image Deconvolution using Hyper-Laplacian Priors. In: Proc. Neural Information Processing Systems, pp. 1033–1041 (2009)Google Scholar
  15. 15.
    Bredies, K., Kunisch, K., Pock, T.: Total generalized variation. SIAM J. Img. Sci. 3(3), 492–526 (2010)CrossRefzbMATHMathSciNetGoogle Scholar
  16. 16.
    Ochs, P., Dosovitskiy, A., Pock, T., Brox, T.: An iterated L1 Algorithm for Non-smooth Non-convex Optimization in Computer Vision. In: IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2013)Google Scholar
  17. 17.
    Bredies, K.: Recovering piecewise smooth multichannel images by minimization of convex functionals with total generalized variation penalty. In: Bruhn, A., Pock, T., Tai, X.-C. (eds.) Global Optimization Methods. LNCS, vol. 8293, pp. 44–77. Springer, Heidelberg (2014)Google Scholar
  18. 18.
    Pock, T., Cremers, D., Bischof, H., Chambolle, A.: An algorithm for minimizing the piecewise smooth Mumford-Shah functional. In: IEEE Int. Conf. on Comp. Vis. (ICCV) (2009)Google Scholar
  19. 19.
    Esser, E., Zhang, X., Chan, T.: A general framework for a class of first order primal-dual algorithms for convex optimization in imaging science. SIAM J. Img. Sci. 3(4), 1015–1046 (2010)CrossRefzbMATHMathSciNetGoogle Scholar
  20. 20.
    Chambolle, A., Pock, T.: A first-order primal-dual algorithm for convex problems with applications to imaging. J. Math. Imaging Vis. 40, 120–145 (2011)CrossRefzbMATHMathSciNetGoogle Scholar
  21. 21.
    Strekalovskiy, E., Cremers, D.: Real-Time Minimization of the Piecewise Smooth Mumford-Shah Functional. In: Proceedings of the European Conference on Computer Vision (2014)Google Scholar
  22. 22.
    Storath, M., Weinmann, A., Demaret, L.: Jump-sparse and sparse recovery using potts functionals. CoRR, pp. 1–1 (2013)Google Scholar
  23. 23.
    Möllenhoff, T., Strekalovskiy, E., Möller, M., Cremers, D.: The Primal-Dual Hybrid Gradient Method for Semiconvex Splittings (preprint, 2014),
  24. 24.
    Ochs, P., Chen, Y., Brox, T., Pock, T.: iPiano: Inertial Proximal Algorithm for Non-convex Optimization. SIAM Journal on Imaging Sciences (SIIMS) (Preprint, 2014)Google Scholar
  25. 25.
    Bouaziz, S., Tagliasacchi, A., Pauly, M.: Sparse Iterative Closest Point. Computer Graphics Forum (Symposium on Geometry Processing) 32(5), 1–11 (2013)CrossRefGoogle Scholar
  26. 26.
    McKelvey, J.P.: Simple transcendental expressions for the roots of cubic equations. Amer. J. Phys. 52(3), 269–270 (1984)CrossRefMathSciNetGoogle Scholar
  27. 27.
    Mirsky, L.: Symmetric gauge functions and unitarily invariant norms. Quart. J. Math. Oxford Ser. (2), 50–59 (1960)Google Scholar

Copyright information

© Springer International Publishing Switzerland 2015

Authors and Affiliations

  • Thomas Möllenhoff
    • 1
  • Evgeny Strekalovskiy
    • 1
  • Michael Moeller
    • 1
  • Daniel Cremers
    • 1
  1. 1.Department of Computer ScienceTechnical University of MunichGermany

Personalised recommendations