Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
149 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
45 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

On the Posterior Distribution in Denoising: Application to Uncertainty Quantification (2309.13598v2)

Published 24 Sep 2023 in cs.CV, cs.LG, and stat.ML

Abstract: Denoisers play a central role in many applications, from noise suppression in low-grade imaging sensors, to empowering score-based generative models. The latter category of methods makes use of Tweedie's formula, which links the posterior mean in Gaussian denoising (\ie the minimum MSE denoiser) with the score of the data distribution. Here, we derive a fundamental relation between the higher-order central moments of the posterior distribution, and the higher-order derivatives of the posterior mean. We harness this result for uncertainty quantification of pre-trained denoisers. Particularly, we show how to efficiently compute the principal components of the posterior distribution for any desired region of an image, as well as to approximate the full marginal distribution along those (or any other) one-dimensional directions. Our method is fast and memory-efficient, as it does not explicitly compute or store the high-order moment tensors and it requires no training or fine tuning of the denoiser. Code and examples are available on the project webpage in https://hilamanor.github.io/GaussianDenoisingPosterior/ .

Definition Search Book Streamline Icon: https://streamlinehq.com
References (54)
  1. NTIRE 2017 challenge on single image super-resolution: Dataset and study. In Proceedings of the IEEE conference on computer vision and pattern recognition workshops, pp.  126–135, 2017. URL https://data.vision.ee.ethz.ch/cvl/DIV2K/.
  2. Image-to-image regression with distribution-free uncertainty quantification and applications in imaging. In International Conference on Machine Learning, pp.  717–730. PMLR, 2022.
  3. Contour detection and hierarchical image segmentation. IEEE transactions on pattern analysis and machine intelligence, 33(5):898–916, 2010. URL https://www2.eecs.berkeley.edu/Research/Projects/CS/vision/grouping/resources.html.
  4. Peter Arbenz. Lecture notes on solving large scale eigenvalue problems. 2016.
  5. Label-efficient semantic segmentation with diffusion models. In International Conference on Learning Representations, 2022.
  6. The generalized cross entropy method, with applications to probability density estimation. Methodology and Computing in Applied Probability, 13(1):1–27, 2011.
  7. Turning a denoiser into a super-resolver using plug and play priors. In 2016 IEEE International Conference on Image Processing (ICIP), pp.  1404–1408. IEEE, 2016.
  8. An efficient statistical method for image noise level estimation. In Proceedings of the 2015 IEEE International Conference on Computer Vision (ICCV), pp.  477–485, 2015.
  9. Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, pp.  248–255. Ieee, 2009.
  10. GENIE: Higher-order denoising diffusion solvers. In Alice H. Oh, Alekh Agarwal, Danielle Belgrave, and Kyunghyun Cho (eds.), Advances in Neural Information Processing Systems, 2022.
  11. Bradley Efron. Tweedie’s formula and selection bias. Journal of the American Statistical Association, 106(496):1602–1614, 2011.
  12. Rich Franzen. Kodak lossless true color image suite. source: http://r0k. us/graphics/kodak, 1999.
  13. Dropout as a bayesian approximation: Representing model uncertainty in deep learning. In international conference on machine learning, pp.  1050–1059. PMLR, 2016.
  14. Rémi Gribonval. Should penalized least squares regression be interpreted as maximum a posteriori estimation? IEEE Transactions on Signal Processing, 59(5):2405–2410, 2011.
  15. Gaussian error linear units (gelus). arXiv preprint arXiv:1606.08415, 2016.
  16. Denoising diffusion probabilistic models. Advances in Neural Information Processing Systems, 33:6840–6851, 2020.
  17. Conffusion: Confidence intervals for diffusion models. arXiv preprint arXiv:2211.09795, 2022.
  18. A style-based generator architecture for generative adversarial networks. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.  4401–4410, 2019. URL https://github.com/NVlabs/ffhq-dataset.
  19. Training generative adversarial networks with limited data. Advances in neural information processing systems, 33:12104–12114, 2020.
  20. Natural image noise level estimation based on local statistics for blind noise reduction. The Visual Computer: International Journal of Computer Graphics, 34(4):575–587, 2018.
  21. Adam: A method for stochastic optimization. International Conference on Learning Representations, 3, 2015.
  22. Additive white gaussian noise level estimation for natural images using linear scale-space features. Circuits, Systems, and Signal Processing, 40(1):353–374, 2021.
  23. Noise2void-learning denoising from single noisy images. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.  2129–2137, 2019.
  24. Probabilistic noise2void: Unsupervised content-aware denoising. Frontiers in Computer Science, 2:5, 2020.
  25. Conformal prediction masks: Visualizing uncertainty in medical imaging. In ICLR 2023 Workshop on Trustworthy Machine Learning for Healthcare, 2023.
  26. Yan LeCun. The MNIST database of handwritten digits. 1998. URL http://yann.lecun.com/exdb/mnist/.
  27. Maskgan: Towards diverse and interactive facial image manipulation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.  5549–5558, 2020.
  28. Swinir: Image restoration using swin transformer. In Proceedings of the IEEE/CVF international conference on computer vision, pp.  1833–1844, 2021.
  29. Enhanced deep residual networks for single image super-resolution. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, July 2017. URL https://github.com/limbee/NTIRE2017.
  30. Gwo Dong Lin. Recent developments on the moment problem. Journal of Statistical Distributions and Applications, 4(1):5, 2017.
  31. Additive white gaussian noise level estimation in svd domain for images. IEEE Transactions on Image processing, 22(3):872–883, 2012.
  32. Single-image noise level estimation for blind denoising. IEEE transactions on image processing, 22(12):5226–5237, 2013.
  33. Maximum likelihood training for score-based diffusion odes by high order denoising score matching. In International Conference on Machine Learning, pp.  14429–14460. PMLR, 2022.
  34. Waterloo exploration database: New challenges for image quality assessment models. IEEE Transactions on Image Processing, 26(2):1004–1016, 2016. URL https://ece.uwaterloo.ca/~k29ma/exploration/.
  35. A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics. In Proceedings Eighth IEEE International Conference on Computer Vision. ICCV 2001, volume 2, pp.  416–423. IEEE, 2001. URL https://github.com/clausmichele/CBSD68-dataset.
  36. Estimating high order gradients of the data distribution by denoising. In M. Ranzato, A. Beygelzimer, Y. Dauphin, P.S. Liang, and J. Wortman Vaughan (eds.), Advances in Neural Information Processing Systems, volume 34, pp.  25359–25369. Curran Associates, Inc., 2021.
  37. Koichi Miyasawa et al. An empirical bayes estimator of the mean of a normal population. Bull. Inst. Internat. Statist, 38(181-188):1–2, 1961.
  38. High-order langevin diffusion yields an accelerated MCMC algorithm. Journal of Machine Learning Research, 22(42):1–41, 2021.
  39. Interval neural networks: Uncertainty scores. arXiv preprint arXiv:2003.11566, 2020.
  40. The little engine that could: Regularization by denoising (RED). SIAM Journal on Imaging Sciences, 10(4):1804–1844, 2017.
  41. Yousef Saad. Numerical methods for large eigenvalue problems: revised edition. SIAM, 2011.
  42. Higher order langevin monte carlo algorithm. Electronic Journal of Statistics, 13(2):3805–3850, 2019.
  43. Semantic uncertainty intervals for disentangled latent spaces. In Alice H. Oh, Alekh Agarwal, Danielle Belgrave, and Kyunghyun Cho (eds.), Advances in Neural Information Processing Systems, 2022.
  44. Deep unsupervised learning using nonequilibrium thermodynamics. In International Conference on Machine Learning, pp.  2256–2265. PMLR, 2015.
  45. Generative modeling by estimating gradients of the data distribution. Advances in neural information processing systems, 32, 2019.
  46. Charles M Stein. Estimation of the mean of a multivariate normal distribution. The Annals of Statistics, 9(6):1135–1151, 1981.
  47. Image restoration by iterative denoising and backward projections. IEEE Transactions on Image Processing, 28(3):1220–1234, 2018.
  48. Plug-and-play priors for model based reconstruction. In 2013 IEEE Global Conference on Signal and Information Processing, pp.  945–948. IEEE, 2013.
  49. Pascal Vincent. A connection between score matching and denoising autoencoders. Neural computation, 23(7):1661–1674, 2011.
  50. Beyond a gaussian denoiser: Residual learning of deep cnn for image denoising. IEEE transactions on image processing, 26(7):3142–3155, 2017a.
  51. Learning deep CNN denoiser prior for image restoration. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.  3929–3938, 2017b.
  52. Plug-and-play image restoration with deep denoiser prior. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(10):6360–6376, 2021.
  53. Color demosaicking by local directional interpolation and nonlocal adaptive thresholding. Journal of Electronic Imaging, 20(2):023016, 2011.
  54. A poisson-gaussian denoising dataset with real fluorescence microscopy images. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.  11710–11718, 2019. URL https://github.com/yinhaoz/denoising-fluorescence.
Citations (7)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now