Abstract
The main goal of many computer vision tasks can in summary be described by finding a function which is optimal according to some criteria. Examples of such functions are the two-dimensional intensity/color function of the image itself in image restoration or deblurring, two-dimensional vector fields like optical flow, or the course of a curve separating the image into multiple areas (which are all presented as examples in this chapter). This is the domain of variational optimization, which aims at estimating those functional relationships. The functional quantifying the quality of a particular solution typically consists of two terms: one for measuring the fidelity of the solution to the observed data and a second term for incorporating prior assumptions about the expected solution, e.g., smoothness constraints. There exist several ways of finding the solution, such as closed-form solutions via the so-called Euler-Lagrange equation, or iterative gradient-based schemes. Despite of the iterative and therefore inherently slow nature of the last-mentioned approach, quite simple iterative update rules can be found for some applications, which allow for a very fast implementation on massively parallel hardware like GPUs. Therefore, variational methods currently are an active area of research in computer vision.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsNotes
- 1.
Actually, (4.32) describes two ROF models: one for \( {{\mathbf{u}}_1} \) (where \( d=1 \)) and one for \( {{\mathbf{u}}_2} \) (where \( d=2 \)). Consequently, we have to perform two (separate) optimizations.
References
Bakushinsky A, Goncharsky A (1994) Ill-posed problems: theory and applications. Kluwer Academic Publishers, Dordrecht. ISBN 978-0792330738
Briggs WL, Henson VE, McCormick SF (2000) A multigrid tutorial, 2nd edn. Society for Industrial and Applied Mathematics, Philadelphia
Chambolle A (2004) An algorithm for total variation minimization and applications. J Math Imaging Vis 20:89–97
Chambolle A (2005) Total variation minimization and a class of binary MRF models. Int Conf Energy Minimization Methods Comput Vision Pattern Recogn 5:136–152
Fergus R, Singh H, Hertzmann A, Rotweis S.T, Freeman W.T (2006) Removing camera shake from a single photograph. In: SIGGRAPH, Boston
Fischen, J (1999) Introduction to the calculus of variations. http://online.redwoods.cc.ca.us/instruct/darnold/staffdev/assignments/calcvarb.pdf. Published online
Golub GH, Van Loan CF (1996) Matrix computations, 3rd edn. Johns Hopkins University Press, Baltimore. ISBN 0-8018-5414-8
Horn K, Schunck B (1981) Determining optical flow. Artif Intell 17:185–203
Jordan M, Ghahramani Z, Jaakkola T, Saul L (1999) An introduction to variational methods for graphical models. Mach Learn 38:183–233
Kass M, Witkin A, Terzopoulos D (1988) Snakes: active contour models. Int J Comput Vis 1:321–331
Kirk D.B, Hwu W.W (2012) Programming massively parallel processors: a hands-on approach, 2nd edn. Morgan Kaufman, San Francisco. ISBN 978-0124159921
Kolev K, Klodt M, Brox T, Cremers D (2009) Continuous global optimization in multiview 3D reconstruction. Int J Comput Vis 84(1):80–96
Kundur D, Hatzinakos D (1996) Blind image deconvolution. IEEE Signal Proc Mag 13(3):43–64
Lellman J, Becker F, Schnörr C (2009) Convex optimization for multi-class image labeling with a novel family of total variation based regularizers. IEEE Int Conf Comput Vis 12:646–653
Levin A, Weiss Y, Durand F, Freeman WT (2009) Understanding blind deconvolution algorithms. Extended Technical Report
Levin A, Weiss Y, Durand F, Freeman WT (2011) Understanding blind deconvolution algorithms. IEEE Signal Proc Mag 33(12):2354–2367
Miskin J, MacKay DC (2000) Ensemble learning for blind image separation and deconvolution. In: Girolani M (ed) Advances in independent component analysis. Springer, London
Nixon M, Aguado A (2007) Feature extraction and image processing. Academic Press, New York. ISBN 978-0-12-372538-7
Oliveira JP, Bioucas-Dias JM, Figueiredo M (2009) Adaptive total variation image deblurring: a majorization-minimization approach. Signal Process 89(9):1683–1693
Pock T (2008) Fast total variation for computer vision. PhD thesis, TU, Graz
Richardson WH (1972) Bayesina-based iterative method of image restoration. J Opt Soc Am 62(1):55–59
Rudin LI, Osher S, Fatemi E (1992) Nonlinear total variation based noise removal algorithms. Physica D 60:259–268
Shan Q, Jia J, Agarwala A (2008) High-quality motion deblurring from a single image. ACM Trans Graphic 27(3)
Tang J (2009) A multi-direction GVF snake for the segmentation of skin cancer images. Pattern Recogn 42:1172–1179
Terzopoulos D (1983) Multilevel computational processes for visual surface reconstruction. Comput Vis Graph Image Process 24:52–96
Tikhonov AN, Arsenin VY (1997) Solutions of ill-posed problems. V.H. Winston, Washington, DC. ISBN 047-099-1240
Treiber M (2010) An introduction to object recognition – selected algorithms for a wide variety of applications. Springer, London. ISBN 978-1849962346
Wanner S, Goldluecke B (2012) Globally consistent depth labeling of 4D light fields. IEEE Conf Comput Vis Pattern Recogn 25:41–48
Wesselink JW (1996) Variational modeling of curves and surfaces. PhD thesis, Technische Universiteit Eindhoven
Whyte O, Sivic J, Zisserman A, Ponce J (2010) Non-uniform deblurring for shaken images. IEEE Conf Comput Vision Pattern Recogn 23:491–498
Xu C, Prince JL (1998) Snakes, shapes and gradient vector flow. IEEE Trans Image Process 7(3):359–369
Zach C, Pock T, Bischof H (2007) A duality based approach for realtime TV-L1 optical flow. In: Proceedings of the 29th DAGM conference on pattern recognition. Springer, Berlin, pp 214–223
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2013 Springer-Verlag London
About this chapter
Cite this chapter
Treiber, M.A. (2013). Variational Methods. In: Optimization for Computer Vision. Advances in Computer Vision and Pattern Recognition. Springer, London. https://doi.org/10.1007/978-1-4471-5283-5_4
Download citation
DOI: https://doi.org/10.1007/978-1-4471-5283-5_4
Published:
Publisher Name: Springer, London
Print ISBN: 978-1-4471-5282-8
Online ISBN: 978-1-4471-5283-5
eBook Packages: Computer ScienceComputer Science (R0)