Pith. sign in

REVIEW 1 major objections 1 minor 59 references

Rotation-Equivariant Self-Supervised Method in Image Denoising

T0 review · 1 major / 1 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper claims that replacing every convolution in a U-Net with rotation-equivariant Fconv layers yields an approximately rotation-equivariant denoiser whose error vanishes as resolution grows, and that this built-in rotation prior…

desk verdict A useful empirical finding about rotation-equivariant self-supervised denoising, but the stated Theorem 3 guarantee is vacuous for ReLU networks. read the letter →

arxiv 2505.19618 v1 pith:2UGOZY3B submitted 2025-05-26 cs.CV

Hanze Liu , Jiahong Fu , Qi Xie , Deyu Meng This is my paper · ORCID
classification cs.CV MSC 68T0768U10
keywords rotationequivarianceself-supervisedimagedenoisingU-NetFconverrorboundsadaptivefusionNoise2NoiseNoise2Void
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper sets out to add a rotation-equivariance prior to self-supervised image denoising by changing the network architecture itself. It claims that replacing every convolution in the U-Net with a rotation-equivariant Fconv layer turns the whole U-Net into an approximately rotation-equivariant denoiser, with rigorous bounds on the equivariance error coming from the downsampling, upsampling, and convolutional layers. If true, self-supervised methods like Noise2Noise, Noise2Void, and Recorrupted-to-Recorrupted inherit the same kind of built-in symmetry that translation-equivariant convolutions already provide for translations, without requiring clean data or rotation augmentation. The paper reports consistent PSNR improvements from the equivariant networks and further gains from an adaptive fusion of equivariant and vanilla branches, selected per image region.

What carries the argument

The load-bearing object is the Fconv rotation-equivariant convolution, a Fourier series expansion based filter parametrization that makes convolutions equivariant to rotations in the continuous domain and only approximately equivariant after discretization. The proof machinery decomposes the U-Net into blocks, one equivariant convolution plus a downsampling operator per downsampling block, and one upsampling operator plus two equivariant convolutions per upsampling block, and bounds each block's equivariance error using first-order smoothness bounds on feature maps. These block errors are then accumulated into the whole-network bound. The adaptive part is a MaskNetwork that produces a per-pixel mask fusing the vanilla branch output and the equivariant branch output element-wise, followed by a residual self-correcting block.

What would settle it

Train the N2N-EQ network and measure its relative rotation equivariance error $\|UNet(\tilde{\pi}_\theta(X))-\tilde{\pi}_\theta(UNet(X))\|/\|UNet(\tilde{\pi}_\theta(X))\|$ at increasing input resolutions. The bound in Theorem 3 predicts the error should decline at the rate $O(h)$ plus $O(h^2)$; if the measured error does not shrink as the resolution increases, the theoretical guarantee does not describe the trained network.

Watch

Extended reading notes

Core claim

The core claim is that the U-Net, despite its upsampling and downsampling layers, can be made a reliable rotation-equivariant network simply by replacing all convolutions with Fconv layers. Theorem 1 bounds the equivariance error of max-pooling and stride downsampling by $2\sqrt{2}Gh$, Theorem 2 bounds nearest-neighbor and bilinear upsampling by $2(\sqrt{2}+1)Gh$, and Theorem 3 accumulates blockwise errors into a whole-network bound $R_1h + R_2h^2$ under smoothness conditions on feature maps and filters; Corollary 1 extends the bound to arbitrary rotation angles. The empirical discovery is that the resulting N2N-EQ, N2V-EQ, and R2R-EQ networks consistently improve PSNR over the vanilla baselines, and that the AdaReNet mask, which mixes the equivariant and vanilla outputs, improves them further. The paper also claims that among tested equivariant convolution designs, Fconv's Fourier-series parametrization is the one that preserves enough representation accuracy for low-level vision.

Load-bearing premise

The error bounds assume the network's internal feature maps and filters vary smoothly, with bounded slopes and curvatures; trained ReLU networks are piecewise linear, so the constants in the bound may not be small or even finite for the network actually trained.

Editorial extensions

If this is right

  • A U-Net with all convolutions replaced by Fconv layers is approximately rotation equivariant, with whole-network error bounded by $R_1h + R_2h^2$, which vanishes as the input resolution grows.
  • Downsampling and upsampling layers contribute equivariance error only at order $O(h)$; because the equivariant convolutions contribute $O(h^2)$, the sampling operators become the dominant error source at finite resolution.
  • The equivariant versions of Noise2Noise, Noise2Void, and Recorrupted-to-Recorrupted beat their vanilla baselines on Gaussian denoising (for instance, N2V-EQ raises BSD500 PSNR from 28.17 to 29.05 dB at $\sigma=25$).
  • A learned mask that fuses the equivariant and vanilla branch outputs yields further consistent gains, using the vanilla branch more near high-frequency edges and the equivariant branch in smooth regions.
  • Without rotation augmentation, the adaptive equivariant network produces the largest improvements, indicating that the architecture-level prior partially substitutes for data augmentation.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the tested methods, the same convolution-replacement recipe should transfer to other U-Net-based self-supervised restoration tasks, such as inpainting or medical-image reconstruction, because the theoretical argument concerns the U-Net skeleton rather than the denoising loss.
  • The mask's observed preference for vanilla output near high-frequency edges implies a testable correlation: mask values should track local gradient magnitude, and forcing the mask to be uniform should reproduce the behavior of the equivariant branch alone.
  • Because the proof assumes smooth activations while ReLU is piecewise linear, one could test whether smoother activations such as SiLU give smaller measured equivariance errors, which would indicate whether the bound governs real trained networks.
  • The specific constants $R_1$, $R_2$, and $R_3$ are stated in the supplementary material, so the size of the bound, not just its $O(h)$ scaling, can be checked there.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

1 major / 1 minor

Summary. This manuscript proposes to incorporate rotation-equivariant convolutions (Fconv) into self-supervised image denoising networks. It states theorems bounding the equivariance error introduced by downsampling and upsampling operations (Theorems 1 and 2) and by a full U-Net with Fconv layers and ReLU activations (Theorem 3 and Corollary 1). It then introduces AdaReNet, which fuses a vanilla CNN branch and an equivariant branch through a learned mask and a self-correcting module, with a loss combining the main denoising objective and per-branch regularizers. Experiments on Noise2Noise, Noise2Void, and R2R report PSNR/SSIM improvements over the vanilla baselines on several Gaussian-noise benchmarks, along with ablations on the choice of equivariant convolution and on rotation augmentation. The code is publicly released.

Significance. If the theoretical claims can be repaired, the paper makes a useful contribution by introducing rotation-equivariant priors into self-supervised denoising at the network architecture level. The empirical results are broadly positive, the ablations across equivariant convolution families are informative, and the public code is a practical asset. The adaptive fusion idea is a reasonable response to the fact that natural images are only locally rotation-equivariant. However, the central theoretical guarantee, which is presented as the paper's headline contribution, is not established as stated because the smoothness assumptions of Theorem 3 are incompatible with the ReLU activation used in the experiments; this issue must be resolved before the 'rigorous theoretical analysis' claim can be accepted.

major comments (1)
  1. [Section 3.1.1, Eq. (4) and (5)] The notation in Eq. (4) treats the feature map as a function on E(2) = R2 ⋊ O(2), while Eq. (3) defines it on R2 × S. The relation between these two domains, and the precise definition of the action πE_\tilde{A} on e(x,A), should be stated more carefully, because the subsequent error bounds depend on which function space and transformation action is used. In particular, after discretization, the sampled feature map is only defined on a finite set of rotations S, and the continuous rotation of such a sample is not unique; the paper should specify how ilde{π}^E_\tilde{A}(F) is computed in the analysis.
minor comments (1)
  1. [Abstract and Introduction] The claim of being the 'first time' with 'comprehensive theoretical analysis' should be tempered, since the theoretical analysis as stated is not valid for ReLU networks; the novelty claim can stand, but it should not be tied to an unproven guarantee.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation: the U-Net equivariance bounds are derived from stated smoothness assumptions and prior published Fconv equivariance results; the main validity concern is a smoothness gap with ReLU, not circularity.

full rationale

The paper's derivation chain is not circular. Theorem 3 bounds the full U-Net equivariance error by R1h + R2h^2 using (i) the Fconv layer equivariance error from the published work of Xie et al. [49], (ii) new bounds for downsampling and upsampling in Theorems 1 and 2, and (iii) block-wise composition over the U-Net. The target result — the equivariance error of the complete U-Net — is not assumed by any of these inputs. The Fconv result is cited from prior published work by overlapping authors, but it is a parameter-free, externally published derivation about the convolution layer itself and does not include the present U-Net conclusion; it is therefore independent support rather than a circular input. The decomposition of the network into downsampling blocks and upsampling blocks is a mathematical construction, not a fitting step, and the constants R1, R2, R3 are stated to depend on the assumed smoothness bounds and network sizes, not on the test data. The hyperparameters alpha1 = alpha2 = 0.1 in the loss are empirical settings, not quantities fitted to the benchmark predictions. There is a genuine correctness concern: condition (15) requires the latent input functions and filters to have uniformly bounded Hessians, while the theorem explicitly uses ReLU activations, and ReLU networks produce piecewise-linear feature maps whose Hessian is not uniformly bounded; this means the advertised finite equivariance guarantee may not apply to the trained network. That concern, however, is about the validity of an assumption, not about an equation or prediction reducing to its own inputs. No fitted parameter is renamed as a prediction, no uniqueness theorem from the authors is imported to forbid alternatives, and no known result is merely relabeled. Accordingly, the paper receives a low circularity score reflecting only the same-group citations, not a circular argument.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central claim does not introduce new physical entities. It depends on smoothness assumptions about features and filters, on the published Fconv construction, and on a specific U-Net decomposition. The only fitted-style parameters are the two loss weights and the rotation group size; learned network weights are standard model parameters, not free scientific parameters.

free parameters (2)
  • Number of rotation group elements t (subgroup size)
    The size of the rotation subgroup S is a design choice that controls the arbitrary-angle equivariance error bound R3*t^{-1}*h in Corollary 1; the value used in experiments is not reported in the main text.
  • Loss weights alpha_1 and alpha_2 = 0.1 each
    Set empirically in Eq. (22) to balance the auxiliary outputs; no tuning procedure or validation search is described.
assumptions (5)
  • domain assumption Images and intermediate feature maps are cell-center discretizations of smooth functions with bounded gradient and Hessian (Eqs. (2)-(3) and conditions (8), (12), (15)).
    Theorems 1-3 use these smoothness bounds to control the discretization error of rotation; real images and ReLU-generated features are not C^2, so the bound may fail at discontinuities.
  • standard math Fconv from Xie et al. [49] is exactly rotation equivariant in the continuous domain and has small discretization error.
    Imported from published work by the same group; the U-Net equivariance claim in this paper inherits this result. Because [49] is a published derivation with external benchmarks, it is treated as independent support rather than circular input.
  • standard math ReLU pointwise nonlinearity commutes with the rotation action and therefore contributes no equivariance error.
    ReLU is applied per channel and the group action permutes channels and rotates coordinates; this is standard for regular-representation equivariant networks.
  • domain assumption The continuous-domain definitions of maxpooling, stride downsampling, nearest-neighbor and bilinear upsampling in Eqs. (6)-(11) faithfully model the discrete operators used in the implementation.
    The implemented U-Nets use discrete TensorFlow/PyTorch operators with specific padding and coordinate conventions; the idealization in the proofs may not match exactly.
  • domain assumption The U-Net can be decomposed into downsampling and upsampling blocks with exactly the stated numbers of E-Conv layers, as in Eq. (14).
    The total error bound is derived for this particular block decomposition; if the real network has different block structures or extra layers, the bound needs extension.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Rotation-Equivariant Self-Supervised Method in Image Denoising." pith.science (2026). https://pith.science/paper/2UGOZY3B

@misc{pith2026250519618,
  author       = {Pith},
  title        = {Pith review of: Rotation-Equivariant Self-Supervised Method in Image Denoising},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2UGOZY3B}},
  note         = {Machine review of arXiv:2505.19618}
}
read the original abstract

Self-supervised image denoising methods have garnered significant research attention in recent years, for this kind of method reduces the requirement of large training datasets. Compared to supervised methods, self-supervised methods rely more on the prior embedded in deep networks themselves. As a result, most of the self-supervised methods are designed with Convolution Neural Networks (CNNs) architectures, which well capture one of the most important image prior, translation equivariant prior. Inspired by the great success achieved by the introduction of translational equivariance, in this paper, we explore the way to further incorporate another important image prior. Specifically, we first apply high-accuracy rotation equivariant convolution to self-supervised image denoising. Through rigorous theoretical analysis, we have proved that simply replacing all the convolution layers with rotation equivariant convolution layers would modify the network into its rotation equivariant version. To the best of our knowledge, this is the first time that rotation equivariant image prior is introduced to self-supervised image denoising at the network architecture level with a comprehensive theoretical analysis of equivariance errors, which offers a new perspective to the field of self-supervised image denoising. Moreover, to further improve the performance, we design a new mask mechanism to fusion the output of rotation equivariant network and vanilla CNN-based network, and construct an adaptive rotation equivariant framework. Through extensive experiments on three typical methods, we have demonstrated the effectiveness of the proposed method.

Figures

Figures reproduced from arXiv: 2505.19618 by the authors.

Figure 1
Figure 1. Illustration of the output feature map of a typical image [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The network architecture of the equivariant N2N [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Illustrations of our proposed adaptive network AdaReNet. Specifically, [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: (a) An image from the Kodak dataset, (b) the heatmap [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: N2N: image denoising results of one image from kodak with [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: N2V: image denoising results of one image from BSD500 with [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 50 canonical work pages

  1. [1]

    Renoir–a dataset for real low-light image noise reduction

    Josue Anaya and Adrian Barbu. Renoir–a dataset for real low-light image noise reduction. Journal of Visual Commu- nication and Image Representation, 51:144–154, 2018. 1

    work page 2018

  2. [2]

    Real image denoising with feature attention

    Saeed Anwar and Nick Barnes. Real image denoising with feature attention. In Proceedings of the IEEE/CVF inter- national conference on computer vision , pages 3155–3164,

    work page

  3. [3]

    Noise2self: Blind denoising by self-supervision

    Joshua Batson and Loic Royer. Noise2self: Blind denoising by self-supervision. In International Conference on Machine Learning, pages 524–533. PMLR, 2019. 1, 3

    work page 2019

  4. [4]

    A non-local algorithm for image denoising

    Antoni Buades, Bartomeu Coll, and J-M Morel. A non-local algorithm for image denoising. In 2005 IEEE computer so- ciety conference on computer vision and pattern recognition (CVPR’05), pages 60–65. Ieee, 2005. 2

    work page 2005

  5. [5]

    Non-local means denoising

    Antoni Buades, Bartomeu Coll, and Jean-Michel Morel. Non-local means denoising. Image Processing On Line , 1: 208–212, 2011. 2

    work page 2011

  6. [6]

    GAN2GAN: Generative Noise Learning for Blind Denoising with Single Noisy Images

    Sungmin Cha, Taeeon Park, and Taesup Moon. Gan2gan: Generative noise learning for blind image denoising with sin- gle noisy images. arXiv preprint arXiv:1905.10488, 3, 2019. 1, 2, 3

    work page Pith review arXiv 1905

  7. [7]

    Spatial- adaptive network for single image denoising

    Meng Chang, Qi Li, Huajun Feng, and Zhihai Xu. Spatial- adaptive network for single image denoising. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXX 16 , pages 171–187. Springer, 2020. 1, 3

    work page 2020

  8. [8]

    Group equivariant convo- lutional networks

    Taco Cohen and Max Welling. Group equivariant convo- lutional networks. In International conference on machine learning, pages 2990–2999. PMLR, 2016. 3, 7

    work page 2016

Show all 59 references
  1. [9]

    Image denoising by sparse 3-d transform- domain collaborative filtering

    Kostadin Dabov, Alessandro Foi, Vladimir Katkovnik, and Karen Egiazarian. Image denoising by sparse 3-d transform- domain collaborative filtering. IEEE Transactions on image processing, 16(8):2080–2095, 2007. 2

    2007

  2. [10]

    Kodak lossless true color image suite

    Rich Franzen. Kodak lossless true color image suite. source: http://r0k. us/graphics/kodak, 4(2):9, 1999. 7, 8

    1999

  3. [11]

    Rotation equivariant proximal operator for deep unfolding methods in image restoration

    Jiahong Fu, Qi Xie, Deyu Meng, and Zongben Xu. Rotation equivariant proximal operator for deep unfolding methods in image restoration. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024. 2, 4

    2024

  4. [12]

    Weighted nuclear norm minimization with applica- tion to image denoising

    Shuhang Gu, Lei Zhang, Wangmeng Zuo, and Xiangchu Feng. Weighted nuclear norm minimization with applica- tion to image denoising. In Proceedings of the IEEE con- ference on computer vision and pattern recognition , pages 2862–2869, 2014. 2

    2014

  5. [13]

    Weighted nuclear norm min- imization and its applications to low level vision

    Shuhang Gu, Qi Xie, Deyu Meng, Wangmeng Zuo, Xi- angchu Feng, and Lei Zhang. Weighted nuclear norm min- imization and its applications to low level vision. Interna- tional journal of computer vision, 121:183–208, 2017. 2

    2017

  6. [14]

    Toward convolutional blind denoising of real pho- tographs

    Shi Guo, Zifei Yan, Kai Zhang, Wangmeng Zuo, and Lei Zhang. Toward convolutional blind denoising of real pho- tographs. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 1712–1722,

  7. [15]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016. 2

    2016

  8. [16]

    Measurement of signal intensities in the presence of noise in mr images

    R Mark Henkelman. Measurement of signal intensities in the presence of noise in mr images. Medical physics, 12(2): 232–233, 1985. 1

    1985

  9. [17]

    Hexaconv

    Emiel Hoogeboom, Jorn WT Peters, Taco S Cohen, and Max Welling. Hexaconv. In International Conference on Learn- ing Representations, 2018. 3

    2018

  10. [18]

    Neighbor2neighbor: Self-supervised de- noising from single noisy images

    Tao Huang, Songjiang Li, Xu Jia, Huchuan Lu, and Jianzhuang Liu. Neighbor2neighbor: Self-supervised de- noising from single noisy images. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 14781–14790, 2021. 1, 2, 3

    2021

  11. [19]

    Foc- net: A fractional optimal control network for image de- noising

    Xixi Jia, Sanyang Liu, Xiangchu Feng, and Lei Zhang. Foc- net: A fractional optimal control network for image de- noising. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 6054–6063,

  12. [20]

    Transfer learning from synthetic to real-noise denois- ing with adaptive instance normalization

    Yoonsik Kim, Jae Woong Soh, Gu Yong Park, and Nam Ik Cho. Transfer learning from synthetic to real-noise denois- ing with adaptive instance normalization. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 3482–3492, 2020. 1

    2020

  13. [21]

    Noise2void-learning denoising from single noisy images

    Alexander Krull, Tim-Oliver Buchholz, and Florian Jug. Noise2void-learning denoising from single noisy images. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 2129–2137, 2019. 1, 3, 7, 8

    2019

  14. [22]

    Probabilistic noise2void: Unsuper- vised content-aware denoising

    Alexander Krull, Tom ´aˇs Vi ˇcar, Mangal Prakash, Manan Lalit, and Florian Jug. Probabilistic noise2void: Unsuper- vised content-aware denoising. Frontiers in Computer Sci- ence, 2:5, 2020. 1

    2020

  15. [23]

    Backpropagation applied to handwrit- ten zip code recognition.Neural computation, 1(4):541–551,

    Yann LeCun, Bernhard Boser, John S Denker, Donnie Henderson, Richard E Howard, Wayne Hubbard, and Lawrence D Jackel. Backpropagation applied to handwrit- ten zip code recognition.Neural computation, 1(4):541–551,

  16. [24]

    Gradient-based learning applied to document recog- nition

    Yann LeCun, L ´eon Bottou, Yoshua Bengio, and Patrick Haffner. Gradient-based learning applied to document recog- nition. Proceedings of the IEEE , 86(11):2278–2324, 1998. 4

    1998

  17. [25]

    Universal denoising networks: a novel cnn architecture for image denoising

    Stamatios Lefkimmiatis. Universal denoising networks: a novel cnn architecture for image denoising. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 3204–3213, 2018. 3

    2018

  18. [26]

    Noise2noise: Learning image restoration without clean data

    Jaakko Lehtinen, Jacob Munkberg, Jon Hasselgren, Samuli Laine, Tero Karras, Miika Aittala, and Timo Aila. Noise2noise: Learning image restoration without clean data. arXiv preprint arXiv:1803.04189, 2018. 1, 3, 5, 7

    2018 arXiv

  19. [27]

    Fm2s: Self- supervised fluorescence microscopy denoising with single noisy image

    Jizhihui Liu, Qixun Teng, and Junjun Jiang. Fm2s: Self- supervised fluorescence microscopy denoising with single noisy image. arXiv preprint arXiv:2412.10031, 2024. 7

    2024

  20. [28]

    Shaking noise exploration and elimination for detecting local flaws of steel wire ropes based on magnetic flux leakages

    Zhiliang Liu and Liyuan Ren. Shaking noise exploration and elimination for detecting local flaws of steel wire ropes based on magnetic flux leakages. IEEE Transactions on Industrial Electronics, 70(4):4206–4216, 2022. 1 9

    2022

  21. [29]

    Ex- act transform-domain noise variance for collaborative filter- ing of stationary correlated noise

    Ymir M ¨akinen, Lucio Azzari, and Alessandro Foi. Ex- act transform-domain noise variance for collaborative filter- ing of stationary correlated noise. In 2019 IEEE interna- tional conference on image processing (ICIP) , pages 185–

    2019

  22. [30]

    Rotation equivariant vector field networks

    Diego Marcos, Michele V olpi, Nikos Komodakis, and Devis Tuia. Rotation equivariant vector field networks. In Pro- ceedings of the IEEE International Conference on Computer Vision, pages 5048–5057, 2017. 3

    2017

  23. [31]

    A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics

    David Martin, Charless Fowlkes, Doron Tal, and Jitendra Malik. 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. IC...

    2001

  24. [32]

    Noisier2noise: Learning to denoise from unpaired noisy data

    Nick Moran, Dan Schmidt, Yu Zhong, and Patrick Coady. Noisier2noise: Learning to denoise from unpaired noisy data. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition , pages 12064–12072,

  25. [33]

    Recorrupted-to-recorrupted: unsupervised deep learning for image denoising

    Tongyao Pang, Huan Zheng, Yuhui Quan, and Hui Ji. Recorrupted-to-recorrupted: unsupervised deep learning for image denoising. In Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition , pages 2043–2052, 2021. 1, 2, 3, 7, 8

    2021

  26. [34]

    Densely connected hierarchical network for image denoising

    Bumjun Park, Songhyun Yu, and Jechang Jeong. Densely connected hierarchical network for image denoising. In Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops, pages 0–0, 2019. 1

    2019

  27. [35]

    Benchmarking denoising algorithms with real photographs

    Tobias Plotz and Stefan Roth. Benchmarking denoising algorithms with real photographs. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 1586–1595, 2017. 1

    2017

  28. [36]

    Self2self with dropout: Learning self-supervised denoising from single image

    Yuhui Quan, Mingqin Chen, Tongyao Pang, and Hui Ji. Self2self with dropout: Learning self-supervised denoising from single image. In Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition , pages 1890–1898, 2020. 3

    2020

  29. [37]

    Image denoising using complex-valued deep cnn

    Yuhui Quan, Yixin Chen, Yizhen Shao, Huan Teng, Yong Xu, and Hui Ji. Image denoising using complex-valued deep cnn. Pattern Recognition, 111:107639, 2021. 3

    2021

  30. [38]

    U- net: Convolutional networks for biomedical image segmen- tation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U- net: Convolutional networks for biomedical image segmen- tation. In Medical image computing and computer-assisted intervention–MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, pa...

    2015

  31. [39]

    Pdo-econvs: Partial differential operator based equivari- ant convolutions

    Zhengyang Shen, Lingshen He, Zhouchen Lin, and Jinwen Ma. Pdo-econvs: Partial differential operator based equivari- ant convolutions. In International Conference on Machine Learning, pages 8697–8706. PMLR, 2020. 2, 3, 7

    2020

  32. [40]

    Pdo-es2cnns: Partial differential operator based equivariant spherical cnns

    Zhengyang Shen, Tiancheng Shen, Zhouchen Lin, and Jin- wen Ma. Pdo-es2cnns: Partial differential operator based equivariant spherical cnns. In Proceedings of the AAAI Con- ference on Artificial Intelligence, pages 9585–9593, 2021. 3

    2021

  33. [41]

    Deep image prior

    Dmitry Ulyanov, Andrea Vedaldi, and Victor Lempitsky. Deep image prior. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 9446–9454,

  34. [42]

    Deep gaussian conditional random field network: A model-based deep network for discriminative denoising

    Raviteja Vemulapalli, Oncel Tuzel, and Ming-Yu Liu. Deep gaussian conditional random field network: A model-based deep network for discriminative denoising. InProceedings of the IEEE conference on computer vision and pattern recog- nition, pages 4801–4809, 2016. 1

    2016

  35. [43]

    High-frequency component helps explain the generaliza- tion of convolutional neural networks

    Haohan Wang, Xindi Wu, Zeyi Huang, and Eric P Xing. High-frequency component helps explain the generaliza- tion of convolutional neural networks. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8684–8694, 2020. 7

    2020

  36. [44]

    Blind2unblind: Self-supervised image denoising with visible blind spots

    Zejin Wang, Jiazheng Liu, Guoqing Li, and Hua Han. Blind2unblind: Self-supervised image denoising with visible blind spots. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 2027–2036,

    2027

  37. [45]

    General e (2)-equivariant steerable cnns

    Maurice Weiler and Gabriele Cesa. General e (2)-equivariant steerable cnns. Advances in neural information processing systems, 32, 2019. 3, 7

    2019

  38. [46]

    Learning steerable filters for rotation equivariant cnns

    Maurice Weiler, Fred A Hamprecht, and Martin Storath. Learning steerable filters for rotation equivariant cnns. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 849–858, 2018. 2, 3

    2018

  39. [47]

    Harmonic networks: Deep translation and rotation equivariance

    Daniel E Worrall, Stephan J Garbin, Daniyar Turmukham- betov, and Gabriel J Brostow. Harmonic networks: Deep translation and rotation equivariance. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recogni- tion, pages 5028–5037, 2017. 3

    2017

  40. [48]

    Unpaired learning of deep image denoising

    Xiaohe Wu, Ming Liu, Yue Cao, Dongwei Ren, and Wang- meng Zuo. Unpaired learning of deep image denoising. In European conference on computer vision , pages 352–368. Springer, 2020. 1

    2020

  41. [49]

    Fourier series expansion based filter parametrization for equivariant convolutions

    Qi Xie, Qian Zhao, Zongben Xu, and Deyu Meng. Fourier series expansion based filter parametrization for equivariant convolutions. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(4):4537–4551, 2022. 2, 3, 4, 7

    2022

  42. [50]

    Real-world noisy image denoising: A new bench- mark

    Jun Xu, Hui Li, Zhetong Liang, David Zhang, and Lei Zhang. Real-world noisy image denoising: A new bench- mark. arXiv preprint arXiv:1804.02603, 2018. 1

    2018 arXiv

  43. [51]

    Noisy-as-clean: Learning self- supervised denoising from corrupted image

    Jun Xu, Yuan Huang, Ming-Ming Cheng, Li Liu, Fan Zhu, Zhou Xu, and Ling Shao. Noisy-as-clean: Learning self- supervised denoising from corrupted image. IEEE Transac- tions on Image Processing, 29:9316–9329, 2020. 1, 2, 3

    2020

  44. [52]

    Dual adversarial network: Toward real-world noise removal and noise generation

    Zongsheng Yue, Qian Zhao, Lei Zhang, and Deyu Meng. Dual adversarial network: Toward real-world noise removal and noise generation. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part X 16, pages 41–58. Springer, 2020. 3

    2020

  45. [53]

    Learning enriched features for real image restoration and enhancement

    Syed Waqas Zamir, Aditya Arora, Salman Khan, Munawar Hayat, Fahad Shahbaz Khan, Ming-Hsuan Yang, and Ling Shao. Learning enriched features for real image restoration and enhancement. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proce...

    2020

  46. [54]

    On single image scale-up using sparse-representations

    Roman Zeyde, Michael Elad, and Matan Protter. On single image scale-up using sparse-representations. In Curves and 10 Surfaces: 7th International Conference, Avignon, France, June 24-30, 2010, Revised Selected Papers 7 , pages 711–

    2010

  47. [55]

    Unleashing the power of self-supervised image denoising: A comprehensive review

    Dan Zhang, Fangfang Zhou, Felix Albu, Yuanzhou Wei, Xiao Yang, Yuan Gu, and Qiang Li. Unleashing the power of self-supervised image denoising: A comprehensive review. arXiv preprint arXiv:2308.00247, 2023. 1, 3

    2023 arXiv

  48. [56]

    Beyond a gaussian denoiser: Residual learning of deep cnn for image denoising

    Kai Zhang, Wangmeng Zuo, Yunjin Chen, Deyu Meng, and Lei Zhang. Beyond a gaussian denoiser: Residual learning of deep cnn for image denoising. IEEE transactions on image processing, 26(7):3142–3155, 2017. 1, 3

    2017

  49. [57]

    Ffdnet: Toward a fast and flexible solution for cnn-based image denoising

    Kai Zhang, Wangmeng Zuo, and Lei Zhang. Ffdnet: Toward a fast and flexible solution for cnn-based image denoising. IEEE Transactions on Image Processing, 27(9):4608–4622,

  50. [58]

    Oriented response networks

    Yanzhao Zhou, Qixiang Ye, Qiang Qiu, and Jianbin Jiao. Oriented response networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages 519–528, 2017. 3 11

    2017

  51. [730]

    Springer, 2012. 7, 8

    2012

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.