Nonlocal Matrix Rank Minimization Method for Multiplicative Noise Removal

Abstract

Multiplicative noise removal is a challenging problem in image denoising. In this paper, we develop a nonlocal matrix rank minimization method for the multiplicative noise removal problem. By utilizing the logarithm transformation, we convert the problem into an additive noise removal problem and propose a maximum a posteriori (MAP) estimation-based matrix rank minimization model for this kind of additive noise removal. A proximal alternating algorithm is designed to solve the matrix rank minimization model. The convergence of the algorithm is demonstrated by the famous Kurdyka-Łojasiewicz property. Taking advantage of the proposed matrix rank minimization model and its proximal alternating algorithm, a multiplicative noise removal method is finally developed. Numerical experiments illustrate that the proposed method can remove multiplicative noise in images much better than the existing state-of-the-art methods in terms of both image recovered measure quantities and visual qualities.

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Keywords

Multiplicative noise / Matrix rank minimization / Proximal alternating method / Kurdyka-Łojasiewicz property / 68U10 / 94A08

References

[1] AttouchH, BolteJ, RedontP, SoubeyranA. Proximal alternating minimization and projection methods for nonconvex problems: an approach based on the Kurdyka-Łojasiewicz inequality. Math. Oper. Res., 2010, 35(2): 438-457

[2] AttouchH, BolteJ, SvaiterBF. Convergence of descent methods for semi-algebraic and tame problems: proximal algorithms, forward-backward splitting, and regularized Gauss-Seidel methods. Math. Prog., 2013, 137: 91-129

[3] AubertG, AujolJ. A variational approach to remove multiplicative noise. SIAM J. Appl. Math., 2008, 68(4): 925-946

[4] BaiZ-Z, PanJ-YMatrix Analysis and Computations, 2021, Philadelphia. SIAM.

[5] BoF, LuW, WangG, ZhouM, WangQ, FangJ. A blind SAR image despeckling method based on improved weighted nuclear norm minimization. IEEE Geosci. Remote Sens. Lett., 2022, 19: 1-5

[6] Bochnak, J., Coste, M., Roy, M.F.: Real Algebraic Geometry. Ergebrisse der Mathematik und ihser Grenzgebiete. Springer, Berlin (1998)

[7] BolteJ, DaniilidisA, LewisA. The Łojasiewicz inequality for nonsmooth subanalytic functions with applications to subgradient dynamical systems. SIAM J. Optim., 2006, 17(4): 1205-1223

[8] BruneC, SawatzkyA, BurgerM. Primal and dual Bregman methods with application to optical nanoscopy. Int. J. Comput. Vis., 2011, 92: 211-229

[9] Buades, A., Coll, B., Morel, J. M.: A non-local algorithm for image denoising. In: 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, vol.2, 60–65 (2005)

[10] BuadesA, CollB, MorelJM. A review of image denoising algorithms, with a new one. SIAM Multiscale Model. Sim., 2005, 4(2): 490-530

[11] ChambolleA, PockT. A first-order primal-dual algorithm for convex problems with applications to imaging. J. Math. Imag. Vis., 2011, 40(1): 120-145

[12] Chan, P.-S., Balakrishnan, N., Vidakovic, B., Kotz, S., Read, C. B.: Log-Gamma distribution. In: Encyclopedia of Statistical Sciences. John Wiley & Sons, Inc., Hoboken, NJ, USA (2004)

[13] Chen, L., Zhu, F., Wang, X.: Low-rank constraint with sparse representation for image restoration under multiplicative noise. Signal Image Video P. 13, 179–187 (2019)

[14] DabovK, FoiA, KatkovnikV, EgiazarianK. Image denoising by sparse 3-D transform-domain collaborative filtering. IEEE Trans. Image Process., 2007, 16(8): 2080-2095

[15] DongW, ShiG, LiX. Nonlocal image restoration with bilateral variance estimation: a low-rank approach. IEEE Trans. Image Process., 2013, 22(2): 700-711

[16] DongY, ZengT. A convex variational model for restoring blurred images with multiplicative noise. SIAM J. Imag. Sci., 2013, 6(3): 1598-1625

[17] DonohoD, JohnstoneI. Ideal spatial adaptation via wavelet shrinkage. Biometrika, 1994, 81: 425-455

[18] DuttV, GreenleafJ. Adaptive speckle reduction filter for log-compressed B-scan images. IEEE Trans. Med. Imag., 1996, 15(6): 802-813

[19] EladM, AharonM. Image denoising via sparse and redundant representations over learned dictionaries. IEEE Trans. Image Process., 2006, 15(12): 3736-3745

[20] Fazel, M., Hindi, H., Boyd, S. P.: A rank minimization heuristic with application to minimum order system approximation. In: Proceedings of the American Control Conference, 4734–4739 (2001)

[21] GolubGH, Van LoanCFMatrix Computations, 20134Baltimore. Johns Hopkins University Press.

[22] Gu, S., Zhang, L., Zuo, W., Feng, X.: Weighted nuclear norm minimization with application to image denoising. In: 2014 IEEE Conference on Computer Vision and Pattern Recognition, 2862–2867 (2014)

[23] GuanD, XiangD, TangX, KuangG. SAR image despeckling based on nonlocal low-rank regularization. IEEE Trans. Geosci. Remote, 2019, 57(6): 3472-3489

[24] HuangY-M, MoisanL, NgMK, ZengT. Multiplicative noise removal via a learned dictionary. IEEE Trans. Image Process., 2012, 21(11): 4534-4543

[25] HuangY-M, YanH-Y. Weighted nuclear norm minimization based-regularization method for image restoration. Commun. Appl. Math Comput., 2021, 3(3): 371-389

[26] HuangY-M, YanH-Y, WenY-W, YangX. Rank minimization with applications to image noise removal. Inform. Sci., 2018, 429: 147-163

[27] HuangY-M, YanH-Y, ZengT. Multiplicative noise removal based on unbiased Box-Cox transformation. Commun. Comput. Phys., 2017, 22(3): 803-828

[28] Ji, H., Liu, C., Shen, Z., Xu, Y.: Robust video denoising using low rank matrix completion. In: 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 1791–1798 (2010)

[29] Kassam, S.A.: Signal Detection in Non-Gaussian Noise. Springer, Berlin (2012)

[30] KatsaggelosAKDigital Image Restoration, 2012, Berlin. Springer.

[31] KimG, ChoJ, KangM. Cauchy noise removal by weighted nuclear norm minimization. J. Sci. Comput., 2020, 83(1): 1-21

[32] Kryvanos, A., Hesser, J., Steidl, G.: Nonlinear image restoration methods for marker extraction in $3$D fluorescent microscopy. In: Proc. SPIE. vol. 5674, pp. 432–443 (2005)

[33] Liu, X., Lu, J., Shen, L., Xu, C., Xu, Y.: Multiplicative noise removal: nonlocal low-rank model and its proximal alternating reweighted minimization algorithm. SIAM J. Imag. Sci. 13(3), 1595–1629 (2020)

[34] LiuX, TanakaM, OkutomiM. Single-image noise level estimation for blind denoising. IEEE Trans. Image Process., 2013, 22: 5226-5237

[35] Lyu, X.-G., Li, F., Liu, J., Lu, S.-T.: A patch-based low-rank minimization approach for speckle noise reduction in ultrasound images. Adv. Appl. Math. Mech. 14(1), 155–180 (2022)

[36] MittalA, SoundararajanR, BovikAC. Making a “completely blind” image quality analyzer. IEEE Signal Process. Lett., 2013, 20(3): 209-212

[37] Mordukhovich, B.: Variational analysis and generalized differentiation I. Basic theory. In: Grundlehren der Mathematischen Wissenschaften. Springer-Verlag, Berlin, (2006)

[38] Rudin, L., Lions, P., Osher, S.: Multiplicative Denoising and Deblurring: Theory and Algorithms, Geometric Level Sets in Imaging, Vision, and Graphics. Springer, New York (2003)

[39] Rudin, L., Osher, S., Fatemi, E.: Nonlinear total variation based noise removal algorithms. Phys. D 60(1/2/3/4), 259–268 (1992)

[40] ScetbonM, EladM, MilanfarP. Deep K-SVD denoising. IEEE Trans. Image Process., 2021, 30: 5944-5955

[41] SchmittJM, XiangS, YungKM. Speckle in optical coherence tomography. J. Biomed. Opt., 1990, 4(1): 95-105

[42] ShanY, HuD, WangZ, JiaT. Multi-channel nuclear norm minus Frobenius norm minimization for color image denoising. Signal Process., 2023, 207108959

[43] SteidlG, TeuberT. Removing multiplicative noise by Douglas-Rachford splitting methods. J. Math. Imag. Vis., 2010, 36: 168-184

[44] TeuberT, LangA. A new similarity measure for nonlocal filtering in the presence of multiplicative noise. Comput. Statist. Data Anal., 2012, 56(12): 3821-3842

[45] WagnerR, SmithS, SandrikJ, LopezH. Statistics of speckle in ultrasound B-scans. IEEE Trans. Sonics Ultrason., 1983, 30(3): 156-163

[46] WangC, GuoB. A double residual iterative regularization method for SAR image despeckling. IEEE Geosci. Remote Sens. Lett., 2023, 20: 1-5

[47] WangZ, BovikAC, SheikhHR, SimoncelliEP. Image quality assessment: from error visibility to structural similarity. IEEE Trans. Image Process., 2004, 13(4): 600-612

[48] WenY-W, ChanRH, ZengT. Primal-dual algorithms for total variation based image restoration under Poisson noise. Sci. China Math., 2016, 59: 141-160

[49] XieY, GuS, LiuY, ZuoW, ZhangW, ZhangL. Weighted schatten $p$-norm minimization for image denoising and background subtraction. IEEE Trans. Image Process., 2016, 25(10): 4842-4857

[50] YanH-Y, HuangY-M. Cauchy noise removal by a generalized nonlocal low-rank method. SPIE J. Electron. Imag., 2022, 313033022

[51] YanH-Y, HuangY-M, YuY. A matrix rank minimization-based regularization method for image restoration. Digit. Signal Process., 2022, 130103694

[52] YangH, LuJ, LuoY, ZhangG, ZhangH, LiangY, LuJ. Nonlocal ultrasound image despeckling via improved statistics and rank constraint. Pattern Anal. Appl., 2022, 26: 1-21

[53] YuY, PengJ, YueS. A new nonconvex approach to low-rank matrix completion with application to image inpainting. Multidim. Syst. Sign. Process., 2019, 30: 145-174

[54] Zheng, H., Yong, H., Zhang, L.: Deep convolutional dictionary learning for image denoising. In: 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 630–641 (2021)

Funding

National Natural Science Foundation of China(11971215)

Key Scientific Research Project for Colleges and Universities of Henan Province(23A110017)

Scientific and Technological Key Projects of Henan Province Grant(232102310227)

RIGHTS & PERMISSIONS

Shanghai University