Pith. sign in

REVIEW 3 major objections 6 minor 59 references

A Progressive Image Restoration Network for High-order Degradation Imaging in Remote Sensing

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read This paper claims that remote sensing images ruined by repeated rounds of blur, noise, and downsampling can be restored by a network whose stages mirror the reverse of those rounds.

desk verdict Solid empirical gains and a legitimately new architecture, but the mathematical-interpretability claim rests on an asserted split that the derivation does not support. read the letter →

arxiv 2412.07195 v2 pith:FIPUXPOH submitted 2024-12-10 cs.CV cs.LGeess.IV

classification cs.CVcs.LGeess.IV
keywords high-orderdegradationremotesensingimagerestorationdeepunfoldingnetworkMarkovchainMAPestimationproximalmappingNeumannseriesdual-domainlearning
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 claims that remote sensing images degraded by multiple successive processes—blur, downsampling, noise, repeated—can be restored by a network whose architecture mirrors the reverse of that degradation chain. The authors build HDI-PRNet from a Markov-chain view of high-order degradation and MAP estimation, with each stage containing a denoising module learned as a proximal mapping, a super-resolution module, and a deblurring module based on Neumann-series expansion with dual-domain degradation learning. If correct, this means a mathematically interpretable, cascaded network can outperform both general and remote-sensing-specific restoration baselines on synthetic high-order degradations and on real satellite images. The reader should care because most existing restoration models assume a single first-order degradation and are black boxes; this paper claims to make the architecture transparent by construction while improving quality.

What carries the argument

The central object is the $k$-th order degradation chain $g = D_k(\cdots D_2(D_1(u)))$ with each $D_\ell = N_\ell \circ S_\ell \circ B_\ell$, treated as a Markov chain so that the joint degradation probability factors into stage transitions. The reverse of this chain is solved as MAP estimation, and the paper's key move is splitting the resulting energy minimization into three subproblems: a proximal-mapping denoising problem (13), a super-resolution constraint (14), and a deblurring problem (15) whose inverse operator is approximated by a truncated Neumann series $B_\ell^{-1} \approx \sum_{i=0}^{m}(I-B_\ell)^i$. The network HDI-PRNet unfolds these steps: a multi-scale residual channel-attention encoder-decoder learns the proximal operator, a convolution-plus-interpolation module changes resolution, and the dual-domain degradation learning block splits blur into spatial and frequency branches that are merged by channel and spatial interaction aggregation. Intermediate losses tie each module to its subproblem, which is the mechanism the paper uses to argue the network is interpretable by construction.

What would settle it

A decisive check would be to train a single-stage network with the same total parameters on the same high-order degraded data, and also to permute the degradation order (for example, noise before blur before downsampling) while keeping the same degradation types; if the single-stage matches HDI-PRNet's PSNR, or if HDI-PRNet fails on the permuted order, the Markov-chain split is not the source of the gains.

Watch

Extended reading notes

Core claim

The paper's central discovery claim is that a deep unfolding network constructed from the reverse of a Markov-chain degradation process can restore remote sensing images better than fourteen existing restoration models. It formalizes the high-order forward model as $g_\ell = N_\ell(S_\ell(B_\ell(g_{\ell-1})))$ for stages $\ell=1,\dots,k$ and derives the reverse process from the MAP estimate $\arg\max_{g_{0:k-1}} \prod_{\ell=1}^k p_\theta(g_{\ell-1}\mid g_\ell)$, which it splits into a proximal-mapping denoising step, a super-resolution constraint, and a Neumann-series deblurring step. Each network stage implements one such split, and intermediate losses supervise each submodule, which the paper presents as architectural transparency rather than black-box fitting. On synthetic high-order degradations and real satellite images, HDI-PRNet achieves the best or near-best PSNR, SSIM, and BRISQUE scores across five public test datasets and five satellite sources. Ablation results show low-order networks lose several decibels on higher-order inputs, which the paper takes as evidence that the high-order Markov structure, not just model size, drives the gains.

Load-bearing premise

The paper's load-bearing premise is that doing denoising, then super-resolution, then deblurring, once per stage, in that fixed order, actually solves the overall restore-the-best-image problem; that step is asserted in the algorithm section without a proof of convergence or equivalence.

Editorial extensions

If this is right

  • First-order restoration networks trained on simple bicubic downsampling lose substantial PSNR when tested on second- and third-order degraded images, so real satellite restoration should model multiple degradation stages instead of one.
  • A staged cascade with intermediate supervision makes each submodule's role identifiable, so the same design can be used to audit whether denoising, deblurring, and super-resolution modules are doing their assigned jobs.
  • Learning deblurring as a truncated Neumann series with spatial-frequency dual branches is presented as more robust and more parameter-efficient than direct inverse deblurring, with ablation showing more expansion terms improve quality.
  • The second-order network is the practical sweet spot, nearly matching third-order reconstruction quality with far fewer parameters and FLOPs.
  • On real satellite images without ground truth, HDI-PRNet obtains lower BRISQUE scores than most baselines, suggesting the high-order model transfers beyond simulated degradations.

Reading between the lines

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

  • Beyond the paper: the same Markov-chain unfolding should transfer to the degradation types the authors list as limitations—clouds, haze, and artifacts—because any invertible degradation step could be inserted as another module in the chain; this is a testable extension they flag but do not implement.
  • The ablations imply that the ordering of degradations matters: a network trained on one order may not invert a permuted chain, so real imaging pipelines should fix the physical order of blur, undersampling, and noise when generating training data.
  • The Neumann-series deblurring module could be reused as a plug-in for any restoration network where the blur operator is close to identity, independent of the high-order framing.
  • If the Markov-chain interpretation is doing the work, then removing the progressive split (replacing the cascade with a single-stage network of matched capacity) should reduce both quality and the interpretability of intermediate outputs; this comparison is not presented in the paper and would be a decisive experiment.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper proposes HDI-PRNet, a multi-stage deep unfolding network for remote sensing image restoration under a high-order degradation model. The degradation process is modeled as a Markov chain, and restoration is formulated as MAP estimation. The MAP energy in Eq. (12) is split into three subproblems: a denoising problem solved by a learned proximal operator, a super-resolution feasibility problem solved by interpolation, and a deblurring problem solved by a truncated Neumann series with dual-domain degradation learning. The network is trained with a reconstruction loss plus an intermediate loss intended to supervise each module. Experiments on AID and WHU Building training data and five remote sensing test datasets show consistent PSNR and SSIM improvements over fourteen comparison methods at scales x2, x3, and x4, supplemented by an evaluation on real satellite images using BRISQUE.

Significance. The empirical contribution is substantial: the comparisons are broad, the gains over strong baselines such as DAT and TTST are consistent across datasets and scales, and the paper includes ablations of the degradation order, loss weight, denoising module scale, Neumann order, and dual-domain block. If the performance claim stands, HDI-PRNet is a competitive remote sensing restoration method. However, the paper's advertised theoretical contribution, namely that the architecture is mathematically interpretable as a MAP solver for a high-order degradation model, is not established by the derivation in Section III-C. The empirical results are independent of that derivation and could survive a re-framing, but the current theoretical claims overreach the evidence.

major comments (3)
  1. [III-C, Eqs. (12)-(15)] The paper asserts that the MAP energy (12) "can be divided" into the three subproblems (13)-(15), and Section III-D then presents the architecture as an unfolding of this split. This equivalence is not proved and, as written, is not correct. Eq. (13) minimizes ||g - g_ell||^2 + lambda_ell f_ell(g), which is a denoising problem on the observed image g_ell and contains no reference to the composition S_ell(B_ell(·)). Eqs. (14) and (15) are hard feasibility constraints rather than least-squares fidelity terms; if exact solutions exist, the composite output satisfies S_ell(B_ell(g_db)) = g_dn, so the data-fidelity term of (12) equals ||g_ell - g_dn||^2 while the regularizer is evaluated at g_db, not at g_dn. There is no argument that a single pass through these three steps minimizes (12) for any lambda_ell, nor a convergence or equivalence proof. Because the abstract and Section III-D present mathematical interpretability as a MAP solver for the high-order model, this is a load-bearing gap. The empirical results can stand without the MAP-interpretability claim, but the claim itself needs either a proof, a clearly stated approximation assumption, or a substantial revision of the interpretation.
  2. [III-E, Eq. (23)] The intermediate loss in Eq. (23) sums ||g_i - u^gt_i||^2 over subscripts i in A, but the ground-truth images u^gt_i are never defined. Different intermediate predictions g_i live in different spaces: the denoising module outputs at the input resolution, the SR module at an intermediate resolution, and the deblurring module at the restored resolution. It is not stated how u^gt_i are produced from the HR ground truth, nor which subscripts A contains. Without this definition, the claim that the intermediate loss makes each submodule "execute its designated function" (Section III-E) and the associated interpretability claim are not verifiable.
  3. [III-D.4, Eqs. (18)-(19)] The Neumann series expansion in Eq. (18) is valid only under a spectral-radius or norm condition such as rho(I - B_ell) < 1. The sentence following Eq. (19) claims that even when this condition is not met, the truncated series can still approximate B_ell^{-1} and acts as implicit regularization. This is not generally true: truncating a non-convergent Neumann series does not yield an approximation to the inverse, and no error bound, learned-correction estimate, or experiment isolating the failure regime is provided. Since the deblurring module is described as a truncated Neumann expansion of the inverse blur operator, this weakens the mathematical grounding of that module; the module may still work as a learned polynomial in B, but it is not justified as an inverse approximation.
minor comments (6)
  1. [IV-B.2, Table II] The Table II caption says "AIRS DATASET", but the experiment in Section IV-B.2 is carried out on WHU-RS19; the caption appears to be a typo.
  2. [IV-C] The text contains the typo "geesubmit-neralizability"; this should be "generalizability".
  3. [IV-D, Table XI] Since lower BRISQUE is better, the phrase "MHAN records the highest score on Jilin-1" should read "lowest BRISQUE". In that row HDI-PRNet is not the best, so the text should either acknowledge this or support the visual over-smoothing argument with a quantitative criterion.
  4. [IV-C] The text introducing Tables VII-X says "WHU-RS9"; the correct dataset name is WHU-RS19.
  5. [III-B, Eq. (11)] The first line of Eq. (11) writes the negative log-likelihood as proportional to (1/(2 sigma_ell^2)) ||g_ell - S_ell(B_ell(g))|| without a square, whereas Eq. (12) uses the squared norm; the notation should be made consistent.
  6. [IV-A.5] No code or trained models are released; given the breadth of the comparisons, a code availability statement would aid reproducibility.

Circularity Check

2 steps flagged · score 5.0 of 10

Performance comparisons are independent, but the MAP/unfolding interpretability claim is partly manufactured: the intermediate loss forces each module to match its named subproblem, and the 'Neumann inversion' blur operator is defined as the learned network, making those correspondences true by construction.

  1. self definitional [Section III-D1 (Overall Architecture); Section III-E2, Eq. (23); ablation in Section IV-B2]
    "Additionally, to reduce information loss between network stages, we add intermediate losses to ensure that the submodule learns the corresponding subproblem, to improve the spatio-temporal feature representation of the framework, and also to ensure the interpretability of the algorithm."

    The paper presents module-to-subproblem alignment as evidence of interpretability, but that alignment is imposed by the intermediate loss (Eq. 23), which minimizes each module output gi against a precomputed target ugt_i for the named denoising/SR/deblurring state. A module trained with this loss matches its designated subproblem by construction, so the claimed correspondence cannot fail and is not an independent consequence of the Markov/MAP derivation. The α=0 versus α=1 ablation (Section IV-B2) only shows that the loss, not the architecture or the MAP split, creates the visible module roles. The interpretability claim is therefore a self-confirming training objective.

  2. self definitional [Section III-D4, Eqs. (18)-(20)]
    "Even if the condition ρ(I − Bℓ) < 1 is not met, the truncated Neumann series (19) can still approximate the inverse operator B−1ℓ, providing a form of implicit regularization [44]. To obtain accurately blur degradation, we use the Dual Domain Degradation Learning Block (DDLB) designed based on the ResFFT-ReLU Block [45] as follows I − Bℓ = I + DEA(BSpatialℓ, BFrequencyℓ) (20) to learn (19)."

    Eq. (19) presents the truncated Neumann series as an approximation to B−1 for a genuine forward blur operator, but Eq. (20) defines the operator by I − B = I + DEA(...), i.e., B = −DEA(...), where DEA is the learned DDLB network whose repeated application implements the series. Substituting (20) into (19) reduces the 'Neumann expansion' to the DDLB cascade itself: g_db ≈ Σ(−DEA)^i(g_sr). The claim that the module solves the deblurring subproblem by Neumann inversion is therefore an identity: the blur operator is defined in terms of the network, and the network is then said to invert that operator. No independent blur operator or verified inverse is involved.

full rationale

The central empirical claim—superior performance on synthetic and real remote sensing images—is grounded in independent benchmark comparisons against 14 methods retrained on the same protocol, and the real-image results use standard metrics and external datasets. That part of the paper is not circular. The interpretability/unfolding claim, however, contains two construction-imposed equivalences. First, the intermediate loss (Eq. 23) forces each submodule to output the named denoised/SR/deblurred intermediate states, and the paper then cites the resulting alignment (and the α=0/α=1 ablation) as evidence that the modules 'execute their designated functions'; the correspondence is manufactured by the loss. Second, the Neumann-series deblurring argument defines the blur operator B in terms of the learned DDLB (Eq. 20), so 'B−1' is an alias for the same network and the inversion claim is true by definition rather than by theorem. The asserted division of the MAP energy (12) into subproblems (13)-(15) is mathematically unsupported—those equalities do not follow from (12) and the composite output does not minimize it—but this is a correctness gap rather than a circular identity, so per the hard rules it does not itself raise the circularity score. The paper's self-citations ([6], [42], [43]) are minor and not load-bearing. Overall, the empirical contribution remains independent, but the headline 'mathematical interpretability' is partially circular, giving a score of 5.

Assumptions & free parameters 5 free parameters · 7 assumptions · 0 invented entities

The framework rests on a long chain of modeling choices: Markov factorization of the degradation chain, Gaussian likelihood and regularizer forms, a sequential non-iterated split of the MAP energy, a learned Neumann-series deblurring inverse, and an unspecified intermediate supervision target. The final method is a trained network; none of the theoretical components is verified to produce the claimed operators, so the interpretive power of the unfolding is an assumption rather than a result.

free parameters (5)
  • degradation_order_k = 2
    Chosen from ablation Table I as the balance between reconstruction performance and complexity; a third order adds roughly 8M parameters for a smaller gain.
  • intermediate_loss_weight_alpha = 1.0
    Selected in Table II; the sweep is on WHU-RS19, which later appears as a test dataset in Table VI.
  • neumann_truncation_order_m = 5
    Chosen from Table IV; performance improves from order 1 to 5 with extra parameters.
  • denoising_module_scale_and_block_count = 3 scales, 2 blocks
    Selected from Table III; a larger multi-scale encoder-decoder improves PSNR by 0.1 dB over the single-scale baseline.
  • degradation_simulation_distribution = see Section IV-A2
    Hand-specified probabilities for blur kernels, resize operations, and noise types; this synthetic distribution is the training proxy for real remote sensing degradation.
assumptions (7)
  • domain assumption The high-order degradation process can be factorized as a Markov chain q(g_{1:k} | g_0) = product_l q(g_l | g_{l-1}) (Eq. 5).
    Used to justify decomposing the full restoration into k progressive stages; the independence of transitions is assumed, not derived from the imaging physics.
  • domain assumption The reverse restoration process factorizes as p_theta(g_{0:k-1} | g_k) = product_l p_theta(g_{l-1} | g_l) (Eq. 6).
    Assumes the posterior of the reverse chain is Markov with the same ordering as the forward degradation.
  • domain assumption Likelihood and prior take the Gaussian plus regularizer forms in Eq. (11), yielding the energy in Eq. (12).
    The noise likelihood is assumed Gaussian with a known variance and the prior is folded into a regularizer; no justification for the particular forms is given.
  • ad hoc to paper Sequentially solving the subproblems (13), (14), and (15) once in fixed order yields the MAP estimate of (12).
    This is the load-bearing split in Section III-C; no equivalence or convergence proof is provided, and the modules are not iterated.
  • domain assumption The truncated Neumann series (19) approximates B_l^{-1} even when the spectral radius condition rho(I - B_l) < 1 fails.
    The paper states the condition then relies on [44] for implicit regularization; no verified bound is given for the learned blur operators.
  • ad hoc to paper Intermediate ground truths u_gt_i for the intermediate loss (23) exist and are computable.
    The loss function references these targets but the paper never specifies how they are produced for each stage.
  • domain assumption The synthetic high-order degradation distribution of Section IV-A2 adequately mimics real remote sensing degradation.
    Training is performed on simulated degradations; transfer to real satellite images depends on this proxy being representative.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Progressive Image Restoration Network for High-order Degradation Imaging in Remote Sensing." pith.science (2026). https://pith.science/paper/FIPUXPOH

@misc{pith2026241207195,
  author       = {Pith},
  title        = {Pith review of: A Progressive Image Restoration Network for High-order Degradation Imaging in Remote Sensing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FIPUXPOH}},
  note         = {Machine review of arXiv:2412.07195}
}
read the original abstract

Recently, deep learning methods have gained remarkable achievements in the field of image restoration for remote sensing (RS). However, most existing RS image restoration methods focus mainly on conventional first-order degradation models, which may not effectively capture the imaging mechanisms of remote sensing images. Furthermore, many RS image restoration approaches that use deep learning are often criticized for their lacks of architecture transparency and model interpretability. To address these problems, we propose a novel progressive restoration network for high-order degradation imaging (HDI-PRNet), to progressively restore different image degradation. HDI-PRNet is developed based on the theoretical framework of degradation imaging, also Markov properties of the high-order degradation process and Maximum a posteriori (MAP) estimation, offering the benefit of mathematical interpretability within the unfolding network. The framework is composed of three main components: a module for image denoising that relies on proximal mapping prior learning, a module for image deblurring that integrates Neumann series expansion with dual-domain degradation learning, and a module for super-resolution. Extensive experiments demonstrate that our method achieves superior performance on both synthetic and real remote sensing images.

Figures

Figures reproduced from arXiv: 2412.07195 by the authors.

Figure 1
Figure 1. The overall framework of the proposed HDI-PRNet solving the high-order degradation model. [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Visualization comparison for different intermediate loss weight [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: PSNR violin-plot of different ×2 SR methods on WHU-RS19 dataset [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: Visualization comparison of different methods on the real satellite images. [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 32 canonical work pages

  1. [44]

    Mvms-rcn: A dual- domain unified ct reconstruction with multi-sparse-view and multi-scale refinement-correction,

    X. Fan, K. Chen, H. Yi, Y . Yang, and J. Zhang, “Mvms-rcn: A dual- domain unified ct reconstruction with multi-sparse-view and multi-scale refinement-correction,” IEEE Transactions on Computational Imaging , pp. 1–14, 2024

  2. [1]

    Unetformer: A unet-like transformer for efficient semantic segmentation of remote sensing urban scene imagery,

    L. Wang, R. Li, C. Zhang, S. Fang, C. Duan, X. Meng, and P. M. Atkinson, “Unetformer: A unet-like transformer for efficient semantic segmentation of remote sensing urban scene imagery,” ISPRS Journal of Photogrammetry and Remote Sensing , vol. 190, pp. 196–214, 2022

  3. [2]

    Learning enriched features for fast image restoration and enhancement,

    S. W. Zamir, A. Arora, S. Khan, M. Hayat, F. S. Khan, M.-H. Yang, and L. Shao, “Learning enriched features for fast image restoration and enhancement,” IEEE transactions on pattern analysis and machine intelligence, vol. 45, no. 2, pp. 1934–1948, 2022

  4. [3]

    Focal network for image restoration,

    Y . Cui, W. Ren, X. Cao, and A. Knoll, “Focal network for image restoration,” in Proceedings of the IEEE/CVF international conference on computer vision , 2023, pp. 13 001–13 011

  5. [4]

    Rethinking coarse-to-fine approach in single image deblurring,

    S.-J. Cho, S.-W. Ji, J.-P. Hong, S.-W. Jung, and S.-J. Ko, “Rethinking coarse-to-fine approach in single image deblurring,” in Proceedings of the IEEE/CVF international conference on computer vision , 2021, pp. 4641–4650

  6. [5]

    Selective frequency network for image restoration,

    Y . Cui, Y . Tao, Z. Bing, W. Ren, X. Gao, X. Cao, K. Huang, and A. Knoll, “Selective frequency network for image restoration,” in The eleventh international conference on learning representations , 2023

  7. [6]

    Nest-dgil: Nesterov- optimized deep geometric incremental learning for cs image reconstruc- tion,

    X. Fan, Y . Yang, K. Chen, Y . Feng, and J. Zhang, “Nest-dgil: Nesterov- optimized deep geometric incremental learning for cs image reconstruc- tion,” IEEE Transactions on Computational Imaging , vol. 9, pp. 819– 833, 2023

  8. [8]

    Real-world remote sensing image super-resolution via a practical degradation model and a kernel-aware network,

    R. Dong, L. Mou, L. Zhang, H. Fu, and X. X. Zhu, “Real-world remote sensing image super-resolution via a practical degradation model and a kernel-aware network,” ISPRS Journal of Photogrammetry and Remote Sensing, vol. 191, pp. 155–170, 2022

Show all 59 references
  1. [9]

    Real-world image super-resolution as multi-task learning,

    W. Zhang, X. Li, G. Shi, X. Chen, Y . Qiao, X. Zhang, X.-M. Wu, and C. Dong, “Real-world image super-resolution as multi-task learning,” Advances in Neural Information Processing Systems , vol. 36, 2024

  2. [10]

    Revitalizing convolutional network for image restoration,

    Y . Cui, W. Ren, X. Cao, and A. Knoll, “Revitalizing convolutional network for image restoration,” IEEE Transactions on Pattern Analysis and Machine Intelligence , 2024

  3. [11]

    Representing noisy image without denoising,

    S. Qi, Y . Zhang, C. Wang, T. Xiang, X. Cao, and Y . Xiang, “Representing noisy image without denoising,” IEEE Transactions on Pattern Analysis and Machine Intelligence , 2024

  4. [12]

    Simple baselines for image restoration,

    L. Chen, X. Chu, X. Zhang, and J. Sun, “Simple baselines for image restoration,” in European conference on computer vision . Springer, 2022, pp. 17–33

  5. [13]

    Image restoration via frequency selection,

    Y . Cui, W. Ren, X. Cao, and A. Knoll, “Image restoration via frequency selection,” IEEE Transactions on Pattern Analysis and Machine Intelli- gence, 2023

  6. [14]

    Image super- resolution using very deep residual channel attention networks,

    Y . Zhang, K. Li, K. Li, L. Wang, B. Zhong, and Y . Fu, “Image super- resolution using very deep residual channel attention networks,” in Proceedings of the European conference on computer vision (ECCV) , 2018, pp. 286–301

  7. [15]

    Fast ultra high-definition video deblurring via multi-scale separable network,

    W. Ren, S. Deng, K. Zhang, F. Song, X. Cao, and M.-H. Yang, “Fast ultra high-definition video deblurring via multi-scale separable network,” International Journal of Computer Vision , vol. 132, no. 5, pp. 1817– 1834, 2024

  8. [16]

    Learning enriched features for real image restoration and enhancement,

    S. W. Zamir, A. Arora, S. Khan, M. Hayat, F. S. Khan, M.-H. Yang, and L. Shao, “Learning enriched features for real image restoration and enhancement,” in Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXV

  9. [17]

    Swinir: Image restoration using swin transformer,

    J. Liang, J. Cao, G. Sun, K. Zhang, L. Van Gool, and R. Timofte, “Swinir: Image restoration using swin transformer,” in Proceedings of the IEEE/CVF international conference on computer vision , 2021, pp. 1833–1844

  10. [18]

    Springer, 2020, pp. 492–511

  11. [19]

    Gaussian kernel mixture network for single image defocus deblurring,

    Y . Quan, Z. Wu, and H. Ji, “Gaussian kernel mixture network for single image defocus deblurring,” Advances in Neural Information Processing Systems, vol. 34, pp. 20 812–20 824, 2021

  12. [20]

    Context aware edge-enhanced gan for remote sensing image super-resolution,

    Z. Ren, L. He, and J. Lu, “Context aware edge-enhanced gan for remote sensing image super-resolution,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , vol. 17, pp. 1363– 1376, 2024

  13. [21]

    Image super-resolution using deep convolutional networks,

    C. Dong, C. C. Loy, K. He, and X. Tang, “Image super-resolution using deep convolutional networks,” IEEE transactions on pattern analysis and machine intelligence , vol. 38, no. 2, pp. 295–307, 2015

  14. [22]

    Esrgan: Enhanced super-resolution generative adversar- ial networks,

    X. Wang, K. Yu, S. Wu, J. Gu, Y . Liu, C. Dong, Y . Qiao, and C. Change Loy, “Esrgan: Enhanced super-resolution generative adversar- ial networks,” in Proceedings of the European conference on computer vision (ECCV) workshops , 2018, pp. 0–0

  15. [23]

    Residual dense network for image restoration,

    ——, “Residual dense network for image restoration,” IEEE transactions on pattern analysis and machine intelligence , vol. 43, no. 7, pp. 2480– 2495, 2020

  16. [24]

    Residual dense net- work for image super-resolution,

    Y . Zhang, Y . Tian, Y . Kong, B. Zhong, and Y . Fu, “Residual dense net- work for image super-resolution,” inProceedings of the IEEE conference on computer vision and pattern recognition , 2018, pp. 2472–2481

  17. [25]

    Second-order attention network for single image super-resolution,

    T. Dai, J. Cai, Y . Zhang, S.-T. Xia, and L. Zhang, “Second-order attention network for single image super-resolution,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2019, pp. 11 065–11 074

  18. [26]

    Restormer: Efficient transformer for high-resolution image restoration,

    S. W. Zamir, A. Arora, S. Khan, M. Hayat, F. S. Khan, and M.-H. Yang, “Restormer: Efficient transformer for high-resolution image restoration,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 5728–5739

  19. [27]

    Dual aggregation transformer for image super-resolution,

    Z. Chen, Y . Zhang, J. Gu, L. Kong, X. Yang, and F. Yu, “Dual aggregation transformer for image super-resolution,” in Proceedings of the IEEE/CVF international conference on computer vision , 2023, pp. 12 312–12 321

  20. [28]

    Image super-resolution with non-local sparse attention,

    Y . Mei, Y . Fan, and Y . Zhou, “Image super-resolution with non-local sparse attention,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2021, pp. 3517–3526

  21. [29]

    Remote sensing image super- resolution via mixed high-order attention network,

    D. Zhang, J. Shao, X. Li, and H. T. Shen, “Remote sensing image super- resolution via mixed high-order attention network,” IEEE Transactions on Geoscience and Remote Sensing, vol. 59, no. 6, pp. 5183–5196, 2020

  22. [30]

    Super-resolution for remote sensing images via local–global combined network,

    S. Lei, Z. Shi, and Z. Zou, “Super-resolution for remote sensing images via local–global combined network,” IEEE Geoscience and Remote Sensing Letters, vol. 14, no. 8, pp. 1243–1247, 2017

  23. [31]

    Ttst: A top-k token selective transformer for remote sensing image super- resolution,

    Y . Xiao, Q. Yuan, K. Jiang, J. He, C.-W. Lin, and L. Zhang, “Ttst: A top-k token selective transformer for remote sensing image super- resolution,” IEEE Transactions on Image Processing , 2024

  24. [32]

    Glf-cr: Sar-enhanced cloud removal with global–local fusion,

    F. Xu, Y . Shi, P. Ebel, L. Yu, G.-S. Xia, W. Yang, and X. X. Zhu, “Glf-cr: Sar-enhanced cloud removal with global–local fusion,” ISPRS Journal of Photogrammetry and Remote Sensing , vol. 192, pp. 268–278, 2022

  25. [33]

    Transformer-based multistage enhancement for remote sensing image super-resolution,

    S. Lei, Z. Shi, and W. Mo, “Transformer-based multistage enhancement for remote sensing image super-resolution,” IEEE Transactions on Geoscience and Remote Sensing , vol. 60, pp. 1–11, 2021

  26. [34]

    Hybrid attention- based u-shaped network for remote sensing image super-resolution,

    J. Wang, B. Wang, X. Wang, Y . Zhao, and T. Long, “Hybrid attention- based u-shaped network for remote sensing image super-resolution,” IEEE Transactions on Geoscience and Remote Sensing , vol. 61, pp. 1–15, 2023

  27. [35]

    Designing a practical degradation model for deep blind image super-resolution,

    K. Zhang, J. Liang, L. Van Gool, and R. Timofte, “Designing a practical degradation model for deep blind image super-resolution,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 4791–4800

  28. [36]

    Real-esrgan: Training real- world blind super-resolution with pure synthetic data,

    X. Wang, L. Xie, C. Dong, and Y . Shan, “Real-esrgan: Training real- world blind super-resolution with pure synthetic data,” in Proceedings of the IEEE/CVF international conference on computer vision , 2021, pp. 1905–1914

  29. [37]

    Towards effective multiple-in-one image restoration: A sequential and prompt learning strategy,

    X. Kong, C. Dong, and L. Zhang, “Towards effective multiple-in-one image restoration: A sequential and prompt learning strategy,” arXiv preprint arXiv:2401.03379, 2024

  30. [38]

    Promptir: Prompting for all-in-one image restoration,

    V . Potlapalli, S. W. Zamir, S. Khan, and F. Khan, “Promptir: Prompting for all-in-one image restoration,” in Thirty-seventh Conference on Neural Information Processing Systems , 2023

  31. [39]

    Unfolding the alternating optimization for blind super resolution,

    Y . Huang, S. Li, L. Wang, T. Tan et al. , “Unfolding the alternating optimization for blind super resolution,” Advances in Neural Information Processing Systems, vol. 33, pp. 5632–5643, 2020

  32. [40]

    Blind super-resolution with iterative kernel correction,

    J. Gu, H. Lu, W. Zuo, and C. Dong, “Blind super-resolution with iterative kernel correction,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2019, pp. 1604–1613

  33. [41]

    Deep generalized unfolding networks for image restoration,

    C. Mou, Q. Wang, and J. Zhang, “Deep generalized unfolding networks for image restoration,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2022, pp. 17 399–17 410

  34. [42]

    Deep unfolding network for image super-resolution,

    K. Zhang, L. V . Gool, and R. Timofte, “Deep unfolding network for image super-resolution,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2020, pp. 3217–3226

  35. [43]

    Deep geometric distillation network for compressive sensing mri,

    X. Fan, Y . Yang, and J. Zhang, “Deep geometric distillation network for compressive sensing mri,” in2021 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI) , 2021, pp. 1–4

  36. [45]

    Intriguing findings of frequency selection for image deblurring,

    X. Mao, Y . Liu, F. Liu, Q. Li, W. Shen, and Y . Wang, “Intriguing findings of frequency selection for image deblurring,” in Proceedings of the AAAI Conference on Artificial Intelligence , vol. 37, no. 2, 2023, pp. 1905– 1913

  37. [46]

    Neumann network with recursive kernels for single image defocus deblurring,

    Y . Quan, Z. Wu, and H. Ji, “Neumann network with recursive kernels for single image defocus deblurring,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2023, pp. 5754–5763

  38. [47]

    Satellite image classification via two-layer sparse coding with biased image representation,

    D. Dai and W. Yang, “Satellite image classification via two-layer sparse coding with biased image representation,” IEEE Geoscience and remote sensing letters, vol. 8, no. 1, pp. 173–176, 2010

  39. [48]

    Aid: A benchmark data set for performance evaluation of aerial scene classification,

    G.-S. Xia, J. Hu, F. Hu, B. Shi, X. Bai, Y . Zhong, L. Zhang, and X. Lu, “Aid: A benchmark data set for performance evaluation of aerial scene classification,” IEEE Transactions on Geoscience and Remote Sensing , vol. 55, no. 7, pp. 3965–3981, 2017

  40. [49]

    Dota: A large-scale dataset for object detection in aerial images,

    G.-S. Xia, X. Bai, J. Ding, Z. Zhu, S. Belongie, J. Luo, M. Datcu, M. Pelillo, and L. Zhang, “Dota: A large-scale dataset for object detection in aerial images,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2018, pp. 3974–3983

  41. [50]

    Fully convolutional networks for multisource building extraction from an open aerial and satellite imagery data set,

    S. Ji, S. Wei, and M. Lu, “Fully convolutional networks for multisource building extraction from an open aerial and satellite imagery data set,” IEEE Transactions on geoscience and remote sensing , vol. 57, no. 1, pp. 574–586, 2018

  42. [51]

    Bag-of-visual-words and spatial extensions for land-use classification,

    Y . Yang and S. Newsam, “Bag-of-visual-words and spatial extensions for land-use classification,” in Proceedings of the 18th SIGSPATIAL in- ternational conference on advances in geographic information systems , 2010, pp. 270–279. ACCEPTED TO IEEE TRANSACTIONS ON GEOSCIENCE AND...

  43. [52]

    Deep learning based feature selection for remote sensing scene classification,

    Q. Zou, L. Ni, T. Zhang, and Q. Wang, “Deep learning based feature selection for remote sensing scene classification,” IEEE Geoscience and remote sensing letters , vol. 12, no. 11, pp. 2321–2325, 2015

  44. [53]

    Image quality assessment: from error visibility to structural similarity,

    Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, “Image quality assessment: from error visibility to structural similarity,” IEEE transactions on image processing , vol. 13, no. 4, pp. 600–612, 2004

  45. [54]

    Remote sensing image scene classifi- cation: Benchmark and state of the art,

    G. Cheng, J. Han, and X. Lu, “Remote sensing image scene classifi- cation: Benchmark and state of the art,” Proceedings of the IEEE , vol. 105, no. 10, pp. 1865–1883, 2017

  46. [55]

    Deep back-projection networks for super-resolution,

    M. Haris, G. Shakhnarovich, and N. Ukita, “Deep back-projection networks for super-resolution,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2018, pp. 1664–1673

  47. [56]

    No-reference image quality assessment in the spatial domain,

    A. Mittal, A. K. Moorthy, and A. C. Bovik, “No-reference image quality assessment in the spatial domain,” IEEE Transactions on image processing, vol. 21, no. 12, pp. 4695–4708, 2012

  48. [57]

    Recursive general- ization transformer for image super-resolution,

    Z. Chen, Y . Zhang, J. Gu, L. Kong, and X. Yang, “Recursive general- ization transformer for image super-resolution,” in ICLR, 2024

  49. [59]

    Cross aggregation transformer for image restoration,

    Z. Chen, Y . Zhang, J. Gu, L. Kong, X. Yuan et al., “Cross aggregation transformer for image restoration,” Advances in Neural Information Processing Systems, vol. 35, pp. 25 478–25 490, 2022

  50. [60]

    Contextual transformation network for lightweight remote-sensing image super-resolution,

    S. Wang, T. Zhou, Y . Lu, and H. Di, “Contextual transformation network for lightweight remote-sensing image super-resolution,” IEEE Transactions on Geoscience and Remote Sensing , vol. 60, pp. 1–13, 2021

  51. [61]

    Hybrid-scale self-similarity exploitation for remote sensing image super-resolution,

    S. Lei and Z. Shi, “Hybrid-scale self-similarity exploitation for remote sensing image super-resolution,” IEEE Transactions on Geoscience and Remote Sensing, vol. 60, pp. 1–10, 2021

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.