Abstract
The Beltrami framework is a successful technique for color image denosing by regarding color images as manifolds embedded in a five dimensional spatial-chromatic space. It can ideally model the coupling between the color channels rather than treating them as if they were independent. However, the resulting model with high nonlinearity makes the related optimization problems difficult to solve numerically. In this paper, we propose an operator-splitting method for a variant of the Beltrami regularization model. From the optimality conditions associated with the minimization of the Beltrami regularized functional, we derive an initial value problem (gradient flow). We solve the gradient flow problem by an operator-splitting scheme involving three fractional steps. All three subproblem solutions can be obtained in closed form or computed by one-step Newton’s method. We demonstrate the efficiency and robustness of the proposed algorithm by conducting a series of experiments on real image denoising problems, where more than half of the computational time is saved compared to the existing augmented Lagrangian method (ALM) based algorithm for solving the Beltrami minimization model.
Similar content being viewed by others
Data Availability
Enquiries about data availability should be directed to the authors.
References
Dabov, K., Foi, A., Katkovnik, V., Egiazarian, K.: Image denoising by sparse 3-D transform-domain collaborative filtering. IEEE Trans. Image Process. 16(8), 2080–2095 (2007)
Mairal, J., Elad, M., Sapiro, G.: Sparse representation for color image restoration. IEEE Trans. Image Process. 17(1), 53–69 (2007)
Xu, J., Zhang, L., Zhang, D., Feng, X.: Multi-channel weighted nuclear norm minimization for real color image denoising. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 1096–1104 (2017)
Kim, S.: PDE-based image restoration: A hybrid model and color image denoising. IEEE Trans. Image Process. 15(5), 1163–1170 (2006)
Lefkimmiatis, S.: Non-local color image denoising with convolutional neural networks. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 3587–3596 (2017)
Deng, G., Cahill, L.: An adaptive gaussian filter for noise reduction and edge detection. In: IEEE Conference Record Nuclear Science Symposium and Medical Imaging Conference, pp. 1615–1619 (1993)
Chen, T., Ma, K.-K., Chen, L.-H.: Tri-state median filter for image denoising. IEEE Trans. Image Process. 8(12), 1834–1838 (1999)
Tomasi, C., Manduchi, R.: Bilateral filtering for gray and color images. In: Sixth IEEE International Conference on Computer Vision, pp. 839–846 (1998)
Kazubek, M.: Wavelet domain image denoising by thresholding and wiener filtering. IEEE Signal Process. Lett. 10(11), 324–326 (2003)
Buades, A., Coll, B., Morel, J.-M.: A non-local algorithm for image denoising. In: IEEE Computer Society Conference on Computer Vision and Pattern Recognition, vol. 2, pp. 60–65 (2005)
Gong, Y., Sbalzarini, I.F.: Curvature filters efficiently reduce certain variational energies. IEEE Trans. Image Process. 26(4), 1786–1798 (2017)
Tschumperle, D., Deriche, R.: Diffusion PDEs on vector-valued images. IEEE Signal Process. Mag. 19(5), 16–25 (2002)
Tschumperlé, D., Deriche, R.: Vector-valued image regularization with PDEs: A common framework for different applications. IEEE Trans. Pattern Anal. Mach. Intell. 27(4), 506–517 (2005)
Barbu, T., Barbu, V., Biga, V., Coca, D.: A PDE variational approach to image denoising and restoration. Nonlinear Anal. Real World Appl. 10(3), 1351–1361 (2009)
Bettahar, S., Stambouli, A.B., Lambert, P., Benoit, A.: PDE-based enhancement of color images in RGB space. IEEE Trans. Image Process. 21(5), 2500–2512 (2011)
Tschumperlé, D.: Fast anisotropic smoothing of multi-valued images using curvature-preserving PDE’s. Int. J. Comput. Vision 68(1), 65–82 (2006)
Shan, X., Sun, J., Guo, Z.: Multiplicative noise removal based on the smooth diffusion equation. J. Math. Imaging Vision 61(6), 763–779 (2019)
Rudin, L.I., Osher, S., Fatemi, E.: Nonlinear total variation based noise removal algorithms. Physica D 60(1–4), 259–268 (1992)
Blomgren, P., Chan, T.F.: Color TV: Total variation methods for restoration of vector-valued images. IEEE Trans. Image Process. 7(3), 304–309 (1998)
Goldluecke, B., Strekalovskiy, E., Cremers, D.: The natural vectorial total variation which arises from geometric measure theory. SIAM J. Imag. Sci. 5(2), 537–563 (2012)
Ono, S., Yamada, I.: Decorrelated vectorial total variation. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4090–4097 (2014)
Lefkimmiatis, S., Roussos, A., Maragos, P., Unser, M.: Structure tensor total variation. SIAM J. Imag. Sci. 8(2), 1090–1122 (2015)
Jia, Z., Ng, M.K., Wang, W.: Color image restoration by saturation-value total variation. SIAM J. Imag. Sci. 12(2), 972–1000 (2019)
Batard, T., Bertalmío, M.: On covariant derivatives and their applications to image regularization. SIAM J. Imag. Sci. 7(4), 2393–2422 (2014)
Batard, T., Sochen, N.: A class of generalized laplacians on vector bundles devoted to multi-channel image processing. J. Math. Imaging Vision 48(3), 517–543 (2014)
Sochen, N., Kimmel, R., Malladi, R.: A general framework for low level vision. IEEE Trans. Image Process. 7(3), 310–318 (1998)
Sochen, N.A., Gilboa, G., Zeevi, Y.Y.: Color Image Enhancement by a Forward-and-backward Adaptive Beltrami Flow. In: International Workshop on Algebraic Frames for the Perception-Action Cycle, pp. 319–328. Springer, Berlin (2000)
Kaftory, R., Sochen, N., Zeevi, Y.: Color image denoising and blind deconvolution using the Beltrami operator. In: International Symposium on Image and Signal Processing and Analysis, vol. 1, pp. 1–4 (2003)
Bresson, X., Vandergheynst, P., Thiran, J.-P.: Multiscale active contours. Int. J. Comput. Vision 70(3), 197–211 (2006)
Dou, Z., Han, Y., Sheng, W., Ma, X.: Image dehaze using alternating Laplacian and Beltrami regularizations. J. Electron. Imaging 24(2), 023004 (2015)
Kimmel, R., Sochen, N., Malladi, R.: From high energy physics to low level vision. In: International Conference on Scale-Space Theories in Computer Vision, pp. 236–247 (1997)
Spira, A., Kimmel, R., Sochen, N.: A short-time Beltrami kernel for smoothing images and manifolds. IEEE Trans. Image Process. 16(6), 1628–1636 (2007)
Rosman, G., Dascal, L., Sidi, A., Kimmel, R.: Efficient Beltrami image filtering via vector extrapolation methods. SIAM J. Imag. Sci. 2(3), 858–878 (2009)
Rosman, G., Dascal, L., Tai, X.-C., Kimmel, R.: On semi-implicit splitting schemes for the Beltrami color image filtering. J. Math. Imaging Vision 40(2), 199–213 (2011)
Rosman, G., Tai, X.-C., Dascal, L., Kimmel, R.: Polyakov action minimization for efficient color image processing. In: Proceedings of the Springer European Conference on Computer Vision, pp. 50–61 (2010)
Wang, W., Ng, M.K.: Color image restoration by saturation-value total variation regularization on vector bundles. SIAM J. Imag. Sci. 14(1), 178–197 (2021)
Glowinski, R., Pan, T.-W., Tai, X.-C.: Some Facts About Operator-splitting and Alternating Direction Methods. In: Splitting Methods in Communication, Imaging, Science, and Engineering, pp. 19–94. Springer, Cham (2016)
Glowinski, R., Osher, S.J., Yin, W.: Splitting Methods in Communication, Imaging, Science, and Engineering. Springer, Berlin (2017)
Deng, L.-J., Glowinski, R., Tai, X.-C.: A new operator splitting method for the Euler elastica model for image smoothing. SIAM J. Imag. Sci. 12(2), 1190–1230 (2019)
Liu, H., Tai, X.-C., Kimmel, R., Glowinski, R.: A color elastica model for vector-valued image regularization. SIAM J. Imag. Sci. 14(2), 717–748 (2021)
Chambolle, A., Lions, P.-L.: Image recovery via total variation minimization and related problems. Numer. Math. 76(2), 167–188 (1997)
Rudin, L., Lions, P.-L., Osher, S.: Multiplicative Denoising and Deblurring: Theory and Algorithms. In: Geometric Level Set Methods in Imaging, Vision, and Graphics, pp. 103–119. Springer, Berlin (2003)
Glowinski, R., Luo, S., Tai, X.-C.: Fast Operator-splitting Algorithms for Variational Imaging Models: Some Recent Developments. In: Handbook of Numerical Analysis, vol. 20, pp. 191–232. Elsevier, Netherlands (2019)
MacNamara, S., Strang, G.: Operator Splitting. In: Splitting Methods in Communication, Imaging, Science, and Engineering, pp. 95–114. Springer, Cham (2016)
Liu, H., Tai, X.-C., Glowinski, R.: An operator-splitting method for the gaussian curvature regularization model with applications in surface smoothing and imaging. arXiv preprint, arXiv:2108.01914 (2021)
Koczka, G., Auberhofer, S., Biro, O., Preis, K.: Optimal convergence of the fixed-point method for nonlinear eddy current problems. IEEE Trans. Magn. 45(3), 948–951 (2009)
Amiri, A., Cordero, A., Darvishi, M.T., Torregrosa, J.R.: Preserving the order of convergence: Low-complexity jacobian-free iterative schemes for solving nonlinear systems. J. Comput. Appl. Math. 337, 87–97 (2018)
Biazar, J., Ghanbary, B.: A new technique for solving systems of nonlinear equations. Appl. Math. Sci. 2(55), 2699–2703 (2008)
Kelley, C.T.: Solving Nonlinear Equations with Newton’s Method. SIAM, Philadelphia (2003)
Abbasbandy, S., Ezzati, R.: Newton’s method for solving a system of fuzzy nonlinear equations. Appl. Math. Comput. 175(2), 1189–1199 (2006)
Brigham, E.O.: The Fast Fourier Transform and Its Applications. Prentice-Hall, New Jersey (1988)
Wang, Z., Bovik, A.C., Sheikh, H.R., Simoncelli, E.P.: Image quality assessment: From error visibility to structural similarity. IEEE Trans. Image Process. 13(4), 600–612 (2004)
Yang, J., Yin, W., Zhang, Y., Wang, Y.: A fast algorithm for edge-preserving variational multichannel image restoration. SIAM J. Imag. Sci. 2(2), 569–592 (2009)
Yang, J., Zhang, Y., Yin, W.: An efficient TVL1 algorithm for deblurring multichannel images corrupted by impulsive noise. SIAM J. Sci. Comput. 31(4), 2842–2865 (2009)
Glowinski, R.: Finite element methods for incompressible viscous flow. Handb. Numer. Anal. 9, 3–1176 (2003)
Acknowledgements
This paper is dedicated to the memory of our dear co-worker Prof. Roland Glowinski, who passed away while this paper was being peer-reviewed. The authors would like to thank Dr. Liangjian Deng for sharing the MATLAB code of Lie scheme based operator splitting method [39]. The work was supported by National Natural Science Foundation of China (NSFC 12071345, 11701418), Major Science and Technology Project of Tianjin 18ZXRHSY00160 and Recruitment Program of Global Young Expert. The work was also supported by projects HKBU 12300819, NSF/RGC Grant N-HKBU214-19, ANR/RGC Joint Research Scheme (A-HKBU203-19) and RC-FNRA-IG/19-20/SCI/01.
Funding
The authors have not disclosed any funding.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors have not disclosed any competing interests.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
In memory of Roland Glowinski–a dear friend, mentor, colleague and great leader.
Appendix
Appendix
Now we provide a brief introduction on the Lie and Marchuk-Yanenko schemes for the time-discretization of initial value problems. Consider the following steady-state problem
where the operator A maps the vector space \(\varvec{V}\) into itself. The classical method to solve (40) is to formulated it as the following initial value problem
Suppose that the problem (41) has steady-state solutions. These solutions are necessarily solutions of problem (40). We further assume the operator A has a nontrivial decomposition, namely
where \(J\ge 2\) and all \(A_j, 0\le j\le J\), are individually simpler than A. Let \(\tau >0\) be a time-discretization step, and denote \(n\tau \) by \(t^n\). Assuming that \(\varvec{X}^n\) is the approximation of \(\varvec{X}(t^n)\), the Lie scheme for solving (41) is given by (see Chapter 6 of [55] for its derivation)
then, for \(n\ge 0\), \(\varvec{X}^n\rightarrow \varvec{X}^{n+1}\) are updated as follows
The Lie scheme (43) and (44) is only semiconstructive since it requires to solve a series of initial value problems. Therefore, by discretizing the subproblems (44) using one step of backward Euler scheme, we obtain the following Marchuk-Yanenko scheme
Due to asymptotic properties of the Lie and Marchuk-Yanenko schemes, the following phenomena generically holds [39]
-
If converging for \(j=1,\ldots ,J\), the sequences \((\varvec{X}^{n+j/J})_{n\ge 0}\) converge to different limits than the solution of (40) with the distances between them being \({\mathcal {O}}(\tau )\) at best.
-
None of the above limits is a steady-state solution, but their distance to a steady-state solution converges to 0 as \(\tau \rightarrow 0\) (if a steady-state solution does exist).
The above convergence theory has been proved in Chapter 6 of [55] supposing the space \(\varvec{V}\) being finite dimensional and the operators \(A_j\) being affine. Since the properties for \(A_j\) do not hold in our Beltrami minimization model, the existing convergence theory cannot be applied to our algorithm, which need to be studied separately as our future works. Moreover, when the Lie or Marchuk-Yanenko scheme applies to multivalued operators such as the subdifferential of proper, lower semicontinuous, convex functionals, the first order accuracy is not guaranteed as well (our case in Sect. 3).
Rights and permissions
About this article
Cite this article
Duan, Y., Zhong, Q., Tai, XC. et al. A Fast Operator-splitting Method for Beltrami Color Image Denoising. J Sci Comput 92, 89 (2022). https://doi.org/10.1007/s10915-022-01910-y
Received:
Revised:
Accepted:
Published:
DOI: https://doi.org/10.1007/s10915-022-01910-y