Advertisement

A Unified View on Patch Aggregation

  • Alexandre Saint-Dizier
  • Julie DelonEmail author
  • Charles Bouveyron
Article
  • 19 Downloads

Abstract

Patch-based methods are widely used in various topics of image processing, such as image restoration or image editing and synthesis. Patches capture local image geometry and structure and are much easier to model than whole images: in practice, patches are small enough to be represented by simple multivariate priors. An important question arising in all patch-based methods is the one of patch aggregation. For instance, in image restoration, restored patches are usually not compatible, in the sense that two overlapping restored patches do not necessarily yield the same values to their common pixels. A standard way to overcome this difficulty is to see the values provided by different patches at a given pixel as independent estimators of a true unknown value and to aggregate these estimators. This aggregation step usually boils down to a simple average, with uniform weights or with weights depending on the trust we have on these different estimators. In this paper, we propose a probabilistic framework aiming at a better understanding of this crucial and often neglected step. The key idea is to see the aggregation of two patches as a fusion between their models rather than a fusion of estimators. The proposed fusion operation is pretty intuitive and generalizes previous aggregation methods. It also yields a novel interpretation of the Expected Patch Log Likelihood (EPLL) proposed in Zoran and Weiss (in: 2011 IEEE international conference on computer vision (ICCV), 2011).

Keywords

Patch-based image restoration and editing Patch aggregation Patch fusion Stochastic image modeling 

Notes

Acknowledgements

We would like to thank Arthur Leclaire and Pablo Arias for their insightful comments. We also would like to thank the three reviewers of this paper, whose comments and remarks have greatly helped to improve this work. This work has been partially funded by the French Research Agency (ANR) under Grant No ANR-14-CE27-001 (MIRIAM).

References

  1. 1.
    Aguerrebere, C., Almansa, A., Delon, J., Gousseau, Y., Musé, P.: A bayesian hyperprior approach for joint image denoising and interpolation, with an application to hdr imaging. IEEE Trans. Comput. Imaging 3(4), 633–646 (2017)MathSciNetCrossRefGoogle Scholar
  2. 2.
    Aharon, M., Elad, M., Bruckstein, A.: \(K\)-SVD: an algorithm for designing overcomplete dictionaries for sparse representation. IEEE Trans. Signal Process. 54(11), 4311–4322 (2006)zbMATHGoogle Scholar
  3. 3.
    Buades, A., Coll, B., Morel, J.-M.: A review of image denoising algorithms, with a new one. Multiscale Model. Simul. 4(2), 490–530 (2005)MathSciNetzbMATHCrossRefGoogle Scholar
  4. 4.
    Carrera, D., Boracchi, G., Foi, A., Wohlberg, B.: Sparse overcomplete denoising: aggregation versus global optimization. IEEE Signal Process. Lett. 24(10), 1468–1472 (2017)CrossRefGoogle Scholar
  5. 5.
    Cho, T.S., Butman, M., Avidan, S., Freeman, W.T.: The patch transform and its applications to image editing. In: IEEE Conference on Computer Vision and Pattern Recognition, 2008. CVPR 2008. IEEE, pp. 1–8 (2008)Google Scholar
  6. 6.
    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)MathSciNetCrossRefGoogle Scholar
  7. 7.
    Danielyan, A., Katkovnik, V., Egiazarian, K.: BM3D frames and variational image deblurring. IEEE Trans. Image Process. 21(4), 1715–1728 (2012)MathSciNetzbMATHCrossRefGoogle Scholar
  8. 8.
    Deledalle, C.-A., Duval, V., Salmon, J.: Non-local methods with shape-adaptive patches (NLM-SAP). J. Math. Imaging Vis. 43(2), 103–120 (2012)MathSciNetzbMATHCrossRefGoogle Scholar
  9. 9.
    Dengwen, Z., Xiaoliu, S.: Image denoising using weighted averaging. In: WRI International Conference on Communications and Mobile Computing, 2009. CMC’09. IEEE, vol. 1, pp. 400–403 (2009)Google Scholar
  10. 10.
    Efros, A.A., Freeman, W.T.: Image quilting for texture synthesis and transfer. In: Proceedings of the 28th Annual Conference on Computer Graphics and Interactive Techniques, ACM, pp. 341–346 (2001)Google Scholar
  11. 11.
    Efros, A.A., Leung, T.K.: Texture synthesis by non-parametric sampling. In: Computer Vision, 1999. IEEE, vol. 2, pp. 1033–1038 (1999)Google Scholar
  12. 12.
    Elad, M., Aharon, M.: Image denoising via sparse and redundant representations over learned dictionaries. IEEE Trans. Image Process. 15(12), 3736–3745 (2006)MathSciNetCrossRefGoogle Scholar
  13. 13.
    Elad, M., Milanfar, P.: Style transfer via texture synthesis. IEEE Trans. Image Process. 26(5), 2338–2351 (2017)MathSciNetzbMATHCrossRefGoogle Scholar
  14. 14.
    Feng, J., Song, L., Huo, X., Yang, X., Zhang, W.: An optimized pixel-wise weighting approach for patch-based image denoising. IEEE Signal Process. Lett. 22(1), 115–119 (2015)CrossRefGoogle Scholar
  15. 15.
    Frigo, O., Sabater, N., Delon, J., Hellier, P.: Split and match: example-based adaptive patch sampling for unsupervised style transfer. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 553–561 (2016)Google Scholar
  16. 16.
    Guleryuz, O.G.: Weighted averaging for denoising with overcomplete dictionaries. IEEE Trans. Image Process. 16(12), 3020–3034 (2007)MathSciNetCrossRefGoogle Scholar
  17. 17.
    HaCohen, Y., Shechtman, E., Goldman, D.B., Lischinski, D.: Non-rigid dense correspondence with applications for image enhancement. In: ACM Transactions on Graphics (TOG), ACM, vol. 30, pp. 70 (2011)CrossRefGoogle Scholar
  18. 18.
    Hoeting, J.A., Madigan, D., Raftery, A.E., Volinsky, C.T.: Bayesian model averaging: a tutorial. Stat. Sci. 382–401 (1999)Google Scholar
  19. 19.
    Houdard, A., Bouveyron, C., Delon, J.: High-dimensional mixture models for unsupervised image denoising (HDMI). SIAM J. Imaging Sci. 11(4), 2815–2846 (2018)MathSciNetzbMATHCrossRefGoogle Scholar
  20. 20.
    Kervrann, C.: PEWA: Patch-based exponentially weighted aggregation for image denoising. In: Advances in Neural Information Processing Systems, pp. 2150–2158 (2014)Google Scholar
  21. 21.
    Kervrann, C., Boulanger, J.: Optimal spatial adaptation for patch-based image denoising. IEEE Trans. Image Process. 15(10), 2866–2878 (2006)CrossRefGoogle Scholar
  22. 22.
    Kwatra, V., Essa, I., Bobick, A., Kwatra, N.: Texture optimization for example-based synthesis. ACM Trans. Gr. (ToG) 24(3), 795–802 (2005)CrossRefGoogle Scholar
  23. 23.
    Lebrun, M., Buades, A., Morel, J.-M.: A nonlocal bayesian image denoising algorithm. SIAM J. Imaging Sci. 6(3), 1665–1688 (2013)MathSciNetzbMATHCrossRefGoogle Scholar
  24. 24.
    Lebrun, M., Colom, M., Buades, A., Morel, J.-M.: Secrets of image denoising cuisine. Acta Numer. 21, 475–576 (2012)MathSciNetzbMATHCrossRefGoogle Scholar
  25. 25.
    Newson, A., Almansa, A., Fradet, M., Gousseau, Y., Pérez, P.: Video inpainting of complex scenes. SIAM J. Imaging Sci. 7(4), 1993–2019 (2014)MathSciNetzbMATHCrossRefGoogle Scholar
  26. 26.
    Paulino, I.F.R.: PACO: Signal restoration via patch consensus. arXiv preprint arXiv:1808.06942 (2018)
  27. 27.
    Pierazzo, N., Morel, J.-M., Facciolo, G.: Multi-scale DCT denoising. Image Process. On Line 7, 288–308 (2017)MathSciNetzbMATHCrossRefGoogle Scholar
  28. 28.
    Raad, L., Desolneux, A., Morel, J.-M.: A conditional multiscale locally gaussian texture synthesis algorithm. J. Math. Imaging Vis. 56(2), 260–279 (2016)MathSciNetzbMATHCrossRefGoogle Scholar
  29. 29.
    Romano, Y., Elad, M.: Boosting of image denoising algorithms. SIAM J. Imaging Sci. 8(2), 1187–1219 (2015)MathSciNetzbMATHCrossRefGoogle Scholar
  30. 30.
    Roth, S., Lempitsky, V., Rother, C.: Discrete-continuous optimization for optical flow estimation. In: Statistical and Geometrical Approaches to Visual Motion Analysis, Springer, pp. 1–22 (2009)Google Scholar
  31. 31.
    Salmon, J., Strozecki, Y.: From patches to pixels in non-local methods: Weighted-average reprojection. In: 2010 17th IEEE International Conference on Image Processing (ICIP). IEEE, pp. 1929–1932 (2010)Google Scholar
  32. 32.
    Sezer, O.G., Altunbasak, Y.: Weighted average denoising with sparse orthonormal transforms. In: 2009 16th IEEE International Conference on Image Processing (ICIP), IEEE, pp. 3849–3852 (2009)Google Scholar
  33. 33.
    Tabti, S., Deledalle, C-A., Denis, L., Tupin, F.: Modeling the distribution of patches with shift-invariance: application to SAR image restoration. In: 2014 IEEE International Conference on Image Processing (ICIP), IEEE, pp. 96–100 (2014)Google Scholar
  34. 34.
    Talebi, H., Zhu, X., Milanfar, P.: How to SAIF-ly boost denoising performance. IEEE Trans. Image Process. 22(4), 1470–1485 (2013)MathSciNetzbMATHCrossRefGoogle Scholar
  35. 35.
    Teodoro, A.M., Almeida, M.S.C., Figueiredo, M.A.T.: Single-frame image denoising and inpainting using gaussian mixtures. In: ICPRAM (2), pp. 283–288 (2015)Google Scholar
  36. 36.
    Van De Ville, D., Kocher, M.: SURE-based non-local means. IEEE Signal Process. Lett. 16(11), 973–976 (2009)CrossRefGoogle Scholar
  37. 37.
    Wang, Y.-Q., Morel, J.-M.: SURE guided Gaussian mixture image denoising. SIAM J. Imaging Sci. 6(2), 999–1034 (2013)MathSciNetzbMATHCrossRefGoogle Scholar
  38. 38.
    Guoshen, Y., Sapiro, G., Mallat, S.: Solving inverse problems with piecewise linear estimators: from Gaussian mixture models to structured sparsity. IEEE Trans. Image Process. 21(5), 2481–2499 (2012)MathSciNetzbMATHCrossRefGoogle Scholar
  39. 39.
    Zontak, M., Mosseri, I., Irani, M.: Separating signal from noise using patch recurrence across scales. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1195–1202 (2013)Google Scholar
  40. 40.
    Zoran, D., Weiss, Y.: From learning models of natural image patches to whole image restoration. In: 2011 IEEE International Conference on Computer Vision (ICCV), IEEE, pp. 479–486 (2011)Google Scholar

Copyright information

© Springer Science+Business Media, LLC, part of Springer Nature 2019

Authors and Affiliations

  1. 1.MAP5, Université Paris DescartesParisFrance
  2. 2.Laboratoire J.A. Dieudonné, Université Côte d’AzurNiceFrance

Personalised recommendations