Skip to main content
SpringerLink
Log in

Improved BM3D method with modified block-matching and multi-scaled images

  • Published:
Multimedia Tools and Applications Aims and scope Submit manuscript

Abstract

BM3D-based denoising has been showing high performance in restoring images damaged by additive white noise and there has been intense research on this method and its variants. In this paper, we make three improvements on BM3D (Block-matching and 3-dimensional filtering). Block matching performs poor and affect denoising performance, especially if noise intensity is high. In the paper, we first proposed a new block similarity metric that accounts for characteristic of noise contained in the observed images in order to guarantee accuracy of block matching even in presence of high intensity noise. Second, block size is a crucial hyperparameter for BM3D. The optimal block size varies with the characteristic of images. However, it is difficult to determine such an optimal block size. We proposed a method to mitigate this difficulty in determining optimal block sizes by combining BM3D and multi-scaled images. Finally, in Aggregation of BM3D, the same weight is assigned to every block in three-dimensional structures. In fact, however, the degree with which noise is removed in each block is different. We presented a method of assigning different weights to blocks according to their respective denoising degrees. Experimental results show that the proposed method is competitive with BM3D and even many of state-of-the-art methods. Actually, it brings about 0.1 ~ 0.6 dB pickup in the PSNR (Peak Signal to Noise Ratio) value. Also, we recommend that it may get better results by applying ideas proposed in this paper individually to state-of-the-art methods.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5

Similar content being viewed by others

References

  1. Abramov SK et al (2008) Segmentation-based method for blind evaluation of noise variance in images. J Appl Remote Sens 2(1)

  2. Buades A, Coll B, Morel JM (2005) A review of image denoising algorithms, with a new one. Multiscale Model Simulation (SIAM interdisciplinary journal) 4(2):490–530

    Article  MathSciNet  Google Scholar 

  3. Chamblle A (2004) An algorithm for Total variation minimization and applications. J Math Imaging Vision 20:89–97

    Article  MathSciNet  Google Scholar 

  4. Chavan P, Cholke A, Prabhudev S (2015) Video scene text Deblurring using text-specific multiscale dictionaries: a survey. Int J Sci, Eng Technol Res (IJSETR) 4(12)

  5. Claude K, Zwicker M (2015) Dual-domain filtering. SIAM J Imaging Sci 8(3):1396–1420

    Article  MathSciNet  Google Scholar 

  6. Dabov K, … Egiazarian K (2007) Image denoising by sparse 3D transform-domain collaborative filtering. in IEEE Trans Image Process 16(8):2080–2095

    Article  MathSciNet  Google Scholar 

  7. Dabov K et al (2006) Image denoising with block-matching and 3D filtering. Proc SPIE Electro Imaging 2006: Image Process: Algorithms Syst V (6064A-30)

  8. K. Dabov, A. Foi, V. Katkovnik, K. Egiazarian; “A nonlocal and shape-adaptive transform-domain collaborative filtering,” in Proc. 2008 Int. Workshop on Local and Non-Local Approximation in Image Processing, LNLA 2008, 2008.

  9. Dabov K et al (2009) BM3D image Denoising with shape-adaptive principal component analysis. Proc Workshop Signal Process Adaptive Sparse Structured Representations (SPARS09):1–6

  10. Dai T et al (2015) Pmpa: a patch-based multiscale products algorithm for image denoising. Image Process (ICIP), 2015 IEEE Int Conf IEEE:4406–4410

  11. Dai T et al (2017) Foveated nonlocal dual denoising. Image Process (ICIP), 2017 IEEE Int Conf IEEE

  12. Danielyan A, Foi A (2009) Noise variance estimation in nonlocal transform domain. 2009 Int Workshop Local Non-Local Approximation Image Process:41–45. https://doi.org/10.1109/LNLA.2009.5278404

  13. Dong W et al (2013) Nonlocally centralized sparse representation for image restoration. IEEE Trans Image Process 22(4):1620–1630. https://doi.org/10.1109/TIP.2012.2235847

    Article  MathSciNet  MATH  Google Scholar 

  14. Gu S et al (2014) Weighted nuclear norm minimization with application to image denoising. Proceed IEEE Int Conf Comp Vision Patt Recog:2862–2869. https://doi.org/10.1109/CVPR.2014.366

  15. Hasan M, El-Sakka MR (2018) Improved BM3D image denoising using SSIM-optimized wiener filter. EURASIP J Image Video Process 2018(25):1–12

    Google Scholar 

  16. He N, … Lu K (2016) Non-local sparse regularization model with application to image denoising. Multimed Tools Appl 75(5):2579–2594. https://doi.org/10.1007/s11042-015-2471-2

    Article  Google Scholar 

  17. Jain P, Tyagi V (2017) An adaptive edge-preserving image denoising technique using patch-based weighted-SVD filtering in wavelet domain. Multimed Tools Appl 76(2):1659–1679. https://doi.org/10.1007/s11042-015-3154-8

    Article  Google Scholar 

  18. Jainand V, Seung H (2008) Natural image denoising with convolutional networks. Adv Neural Inf Process Syst (NIPS) 21:769–776

    Google Scholar 

  19. Katkovnik V et al (2017) Complex domain nonlocal block-matching denoising based on high-order singular value decomposition (HOSVD). EURASIP, 2017 25th Eur Signal Process Conf((EUSIPCO):748–752

  20. Knaus C, Zwicker M (2013, 2013) Dual-domain image Denoising. IEEE Int Conf Image Process:440–444

  21. Lebrun M (2012) An analysis and implementation of the BM3D image denoising method. Image Process On Line 2(25):175–213. https://doi.org/10.5201/ipol.2012.l-bm3d

    Article  Google Scholar 

  22. Li YJ (2017) Improved BM3D denoising method. IET Image Process 11(12):1197–1204

    Article  Google Scholar 

  23. Mairal J, … Elad M (2008) Learning multiscale sparse representations for image and video restoration. Soc Industrial Appl Math 7(1):214–241

    MathSciNet  MATH  Google Scholar 

  24. Monagi H, Alkinani, Mahmoud R, El-Sakka (2016) A modified block matching 3D algorithm for additive noise reduction. Math Appl 5:93–103

    MathSciNet  MATH  Google Scholar 

  25. Perona P, Malik J (1990) Scale-space and edge detection using anisotropic diffusion. IEEE Trans Patt Anal Mach Intell 12(7):629–639

    Article  Google Scholar 

  26. Poderico M et al (2010) Sigmoid shrinkage for BM3D denoising algorithm. 2010 IEEE Int Workshop Multimedia Signal Process:423–426

  27. Singh R et al (2017) Image Denoising using BM3D and principal component analysis. IEEE Trans Image Process 6(5):7–13

    Google Scholar 

  28. Singh V (2017) Comparative study of algorithms/techniques for Denoising of Gaussian noise. IJARC 8(8):78–82

    Article  Google Scholar 

  29. Theodoridis S (2009) Konstantinos Koutroumbas; "pattern recognition", 4th edn

    Google Scholar 

  30. Xu J, Zhang L, Zuo W, Zhang D, Feng X (2015) Patch group based nonlocal self- similarity prior learning for image denoising. Proceed IEEE Int Conf Comp Vision:244–252, 10.1109/ICCV.2015.36

  31. X. Zhang; “Center pixel weight based on wiener filter for non-local means image denoising”, in: Proceed IEEE Int Conf Comp Vision Patt Recog, pp. 2862–2869, 2021. https://doi.org/10.1109/CVPR.2014.366.

  32. Zhang YS et al (2012) BM3D Denoising algorithm with adaptive block-match thresholds. Appl Mech Mat . 229–231:1715–1720. https://doi.org/10.4028/www.scientific.net/AMM.229-231.1715

  33. Zoran D, Weiss Y (2011) From learning models of natural image patches to whole image restoration. Proceed IEEE Int Conf Comp Vision:479–486. https://doi.org/10.1109/ICCV.2011.6126278

  34. Zou B-J, … Chen Z-L (2018) 3D filtering by block matching and convolutional neural network for image Denoising. J Comp Sci Technol 33(4):838–848

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations

  1. Faculty of Mathematics, Kim Il Sung University, Pyongyang, Democratic People’s Republic of Korea

    Gwang-Il Ri, Song-Jun Kim & Man-Su Kim

Authors
  1. Gwang-Il Ri
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Song-Jun Kim
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Man-Su Kim
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Gwang-Il Ri.

Ethics declarations

Conflict of interest

The authors declare no conflict of interest.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Ri, GI., Kim, SJ. & Kim, MS. Improved BM3D method with modified block-matching and multi-scaled images. Multimed Tools Appl 81, 12661–12679 (2022). https://doi.org/10.1007/s11042-022-12270-y

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11042-022-12270-y

Keywords

Search

Navigation