Abstract
We propose a novel idea to introduce regularization based on second order total generalized variation (\(\text {TGV}\)) into optimization frameworks based on functional lifting. The proposed formulation extends a recent sublabel-accurate relaxation for multi-label problems and thus allows for accurate solutions using only a small number of labels, significantly improving over previous approaches towards lifting the total generalized variation. Moreover, even recent sublabel accurate methods exhibit staircasing artifacts when used in conjunction with common first order regularizers such as the total variation (\(\text {TV}\)). This becomes very obvious for example when computing derivatives of disparity maps computed with these methods to obtain normals, which immediately reveals their local flatness and yields inaccurate normal maps. We show that our approach is effective in reducing these artifacts, obtaining disparity maps with a smooth normal field in a single optimization pass.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Alberti, G., Bouchitté, G., Dal Maso, G.: The calibration method for the Mumford-Shah functional and free-discontinuity problems. Calc. Var. Partial Differ. Equ. 16(3), 299–333 (2003)
Blake, A., Zisserman, A.: Visual Reconstruction. MIT Press, Cambridge (1987)
Bouchitté, G.: Recent convexity arguments in the calculus of variations. In: Lecture Notes from the 3rd International Summer School on the Calculus of Variations, Pisa (1998)
Boykov, Y., Veksler, O., Zabih, R.: Markov random fields with efficient approximations. In: Proceedings of International Conference on Computer Vision and Pattern Recognition, Santa Barbara, California, pp. 648–655 (1998)
Bredies, K., Kunisch, K., Pock, T.: Total generalized variation. SIAM J. Imaging Sci. 3(3), 492–526 (2010)
Chambolle, A.: Convex representation for lower semicontinuous envelopes of functionals in \(L^1\). J. Convex Anal. 8(1), 149–170 (2001)
Cremers, D., Strekalovskiy, E.: Total cyclic variation and generalizations. J. Math. Imaging Vis. 47(3), 258–277 (2012)
Ferstl, D., Reinbacher, C., Ranftl, R., Rüther, M., Bischof, H.: Image guided depth upsampling using anisotropic total generalized variation. In: Proceedings of International Conference on Computer Vision, pp. 993–1000. IEEE (2013)
Geman, S., Geman, D.: Stochastic relaxation, Gibbs distributions, and the Bayesian restoration of images. IEEE Trans. Pattern Anal. Mach. Intell. 6, 721–741 (1984)
Goldluecke, B., Strekalovskiy, E., Cremers, D.: Tight convex relaxations for vector-valued labeling. SIAM J. Imaging Sci. 6(3), 1626–1664 (2013)
Honauer, K., Johannsen, O., Kondermann, D., Goldluecke, B.: A dataset and evaluation methodology for depth estimation on 4D light fields. In: Lai, S.-H., Lepetit, V., Nishino, K., Sato, Y. (eds.) ACCV 2016. LNCS, vol. 10113, pp. 19–34. Springer, Cham (2017). https://doi.org/10.1007/978-3-319-54187-7_2
Ishikawa, H.: Exact optimization for Markov random fields with convex priors. IEEE Trans. Pattern Anal. Mach. Intell. 25(10), 1333–1336 (2003)
Kolmogorov, V., Zabih, R.: What energy functions can be minimized via graph cuts. IEEE Trans. Pattern Anal. Mach. Intell. 26(2), 147–159 (2004)
Lellmann, J., Becker, F., Schnörr, C.: Convex optimization for multi-class image labeling with a novel family of total variation based regularizers. In: IEEE International Conference on Computer Vision (ICCV) (2009)
Moellenhoff, T., Laude, E., Moeller, M., Lellmann, J., Cremers, D.: Sublabel-accurate relaxation of nonconvex energies. In: Proceedings of International Conference on Computer Vision and Pattern Recognition (2016)
Möllenhoff, T., Cremers, D.: Sublabel-accurate discretization of nonconvex free-discontinuity problems. In: International Conference on Computer Vision (ICCV), Venice, Italy, October 2017
Mumford, D., Shah, J.: Optimal approximations by piecewise smooth functions and associated variational problems. Commun. Pure Appl. Math. 42, 577–685 (1989)
Pock, T., Cremers, D., Bischof, H., Chambolle, A.: Global solutions of variational models with convex regularization. SIAM J. Imaging Sci. 3, 1122–1145 (2010)
Ranftl, R., Gehrig, S., Pock, T., Bischof, H.: Pushing the limits of stereo using variational stereo estimation. In: IEEE Intelligent Vehicles Symposium (2012)
Ranftl, R., Pock, T., Bischof, H.: Minimizing TGV-based variational models with non-convex data terms. In: Kuijper, A., Bredies, K., Pock, T., Bischof, H. (eds.) SSVM 2013. LNCS, vol. 7893, pp. 282–293. Springer, Heidelberg (2013). https://doi.org/10.1007/978-3-642-38267-3_24
Rudin, L.I., Osher, S., Fatemi, E.: Nonlinear total variation based noise removal algorithms. Phys. D 60, 259–268 (1992)
Strecke, M., Alperovich, A., Goldluecke, B.: Accurate depth and normal maps from occlusion-aware focal stack symmetry. In: Proceedings of International Conference on Computer Vision and Pattern Recognition (2017)
Zach, C., Häne, C., Pollefeys, M.: What is optimized in convex relaxations for multilabel problems: connecting discrete and continuously inspired map inference. IEEE Trans. Pattern Anal. Mach. Intell. 36(1), 157–170 (2014)
Acknowledgements
This work was supported by the ERC Starting Grant “Light Field Imaging and Analysis” (LIA 336978, FP7-2014) and the SFB Transregio 161 “Quantitative Methods for Visual Computing”.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2019 Springer Nature Switzerland AG
About this paper
Cite this paper
Strecke, M., Goldluecke, B. (2019). Sublabel-Accurate Convex Relaxation with Total Generalized Variation Regularization. In: Brox, T., Bruhn, A., Fritz, M. (eds) Pattern Recognition. GCPR 2018. Lecture Notes in Computer Science(), vol 11269. Springer, Cham. https://doi.org/10.1007/978-3-030-12939-2_19
Download citation
DOI: https://doi.org/10.1007/978-3-030-12939-2_19
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-12938-5
Online ISBN: 978-3-030-12939-2
eBook Packages: Computer ScienceComputer Science (R0)