Pith. sign in

REVIEW 4 major objections 3 minor 31 references

All-in-One Medical Image Restoration with Latent Diffusion-Enhanced Vector-Quantized Codebook Prior

T0 review · 4 major / 3 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read DiffCode proposes a task-adaptive codebook bank plus latent diffusion that restores MRI, CT, and PET images with one model and reports the best average scores among twelve methods.

desk verdict A plausible codebook-prior architecture for all-in-one MedIR, undermined by weak evidence for its signature latent-diffusion component. read the letter →

arxiv 2507.19874 v1 pith:LOWQTOFN submitted 2025-07-26 cs.CV

classification cs.CV
keywords medicalimagerestorationall-in-onevectorquantizationcodebookpriorlatentdiffusionMRIsuper-resolutionCTdenoisingPETsynthesis
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

DiffCode is a unified framework that restores degraded medical images from three different modalities—MRI, CT, and PET—with one set of weights. The paper's central claim is that a task-adaptive bank of vector-quantized codebooks, one codebook per task, supplies high-quality prior features that compensate for the different kinds of information each degradation destroys, and that a short latent diffusion process makes retrieval from that bank accurate enough to matter. On the IXI MRI, LDCT, and clinical PET datasets, DiffCode reports the best average scores among twelve methods, with PSNR 34.6186, SSIM 0.9336, and RMSE 12.3728. The authors argue that this is the first effective use of a vector-quantized codebook prior for all-in-one medical image restoration.

What carries the argument

The central object is the task-adaptive codebook bank: a set of vector-quantized codebooks, one per task, each containing 8,192 code items of dimension 256 and updated by residual vector quantization across eight iterations. The latent diffusion stage, built from a five-layer denoising network run for eight time steps, maps degraded-image-conditioned Gaussian noise back toward high-quality latent features before those features are quantized against the bank. A task-aware global routing module with four experts, one activated per image, routes each task's features through specialized subnetworks to reduce task interference.

What would settle it

On a held-out split, extract the true high-quality features for each test image, quantize them against the codebook bank, and compare the retrieved codes or feature distances from the diffusion path versus a one-step direct prediction from the degraded features. If the diffusion path does not improve retrieval accuracy or feature similarity, the paper's explanation for its gains is unsupported.

Watch

Extended reading notes

Core claim

The paper proposes that the obstacle to all-in-one medical image restoration is not the restoration backbone but the retrieval of high-quality prior information: each task destroys different features, so a single shared prior cannot compensate for all losses. DiffCode therefore builds a codebook bank in which each task has its own residual-quantized codebook of high-quality latent features, and then trains a latent diffusion model that, conditioned on the degraded image's features, iteratively refines a noisy latent toward the correct high-quality feature before quantization. The retrieved codes are decoded into a reference image that is concatenated with the degraded input and fed to a restoration network with task-aware routing. The paper reports that this pipeline outperforms eleven task-specific, general, and all-in-one baselines on MRI super-resolution, CT denoising, and PET synthesis, and that each component—codebook bank, latent diffusion, and routing—contributes positively in the ablations.

Load-bearing premise

The load-bearing premise is that a five-layer diffusion network can recover the correct high-quality latent feature from Gaussian noise in only eight denoising steps, conditioned on the degraded image; the paper validates this only through final restoration metrics, not by checking the retrieved features themselves.

Editorial extensions

If this is right

  • One trained model can serve MRI super-resolution, CT denoising, and PET synthesis without task-specific fine-tuning, simplifying clinical deployment.
  • Adding a new task may only require appending a new codebook and retraining the retrieval stage, rather than redesigning the whole restoration network.
  • The codebook prior supplies high-quality reference images that could be used for downstream quality control or as inputs to other analysis models.
  • Because the codebook bank and diffusion retriever are built once and then frozen, they can guide any restoration backbone, not only the NAF-block architecture used here.
  • The reported gains over AMIR, the closest all-in-one baseline, suggest that the bottleneck in all-in-one restoration is prior retrieval rather than routing alone.

Reading between the lines

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

  • A testable implication the paper leaves implicit is whether the diffusion retriever's benefit grows with degradation severity; if retrieval errors increase as blur or noise worsens, the fixed eight-step schedule may need task-dependent length.
  • The codebook bank could in principle be extended to unseen tasks by adding codebooks trained on small amounts of high-quality data, turning restoration into a plug-and-play prior library; the paper does not demonstrate this.
  • Comparing the retrieved code indices against ground-truth indices would isolate where the diffusion stage helps, since final PSNR conflates retrieval accuracy with the backbone's ability to ignore bad codes.
  • The same latent-diffusion-plus-codebook pattern may transfer to non-medical multi-degradation restoration, where task heterogeneity is equally pronounced; this is an extrapolation, not a claim in the paper.
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

4 major / 3 minor

Summary. The paper proposes DiffCode, a three-stage framework for all-in-one medical image restoration. Stage I builds a task-adaptive vector-quantized codebook bank with residual quantization on high-quality images for each task. Stage II trains a small latent diffusion model, conditioned on low-quality features, to estimate features that are then quantized against the codebook bank; the stated goal is to improve prior retrieval by refining the latent feature distribution. Stage III decodes the retrieved codebook features into reference images and concatenates them with the low-quality input to guide a task-aware routing restoration backbone. The authors report state-of-the-art results on MRI super-resolution, CT denoising, and PET synthesis, with average PSNR 34.6186, SSIM 0.9336, and RMSE 12.3728, and they provide component ablations in Table 2.

Significance. If the results are reproducible and the diffusion mechanism is genuinely improving prior retrieval, the task-adaptive codebook bank is a sensible and potentially useful contribution to all-in-one medical image restoration, where task heterogeneity is a recognized challenge. The paper is clearly written, the architecture is specified in detail, and the experimental comparison against eleven methods across three tasks is comprehensive. The component ablation shows that each module adds some quantitative benefit. However, the central novelty of the paper is the latent diffusion stage, and that stage is currently supported only by a small end-task metric gain with no direct evidence about retrieval behavior or distribution alignment; this weakens the significance of the claimed mechanism.

major comments (4)
  1. [Sec. 2.2 and Table 2(a)] The central claim that the latent diffusion model 'enhances prior retrieval' is supported only by a 0.0913 dB average PSNR improvement (V3: 34.5273 vs. DiffCode: 34.6186) and a 0.0013 SSIM improvement, reported without error bars or significance tests. No direct evidence is provided that the diffusion process changes which codebook entries are retrieved or that the estimated features align with ground-truth HQ features. This is load-bearing because the diffusion stage is the paper's main novelty; I request direct retrieval analysis (e.g., retrieved code index agreement with the HQ code indices, feature-space distance or MMD before vs. after diffusion) and repeated-seed statistics to rule out that the gain comes from added stochasticity or parameter count rather than from improved retrieval.
  2. [Sec. 2.2, Eq. (5)] At inference the reverse diffusion is initialized from randomly sampled Gaussian noise, but the experimental tables report a single run with no seed or variance information. Given that the largest reported advantage over the strongest baseline AMIR is 0.2601 dB average PSNR (Table 1), run-to-run variability could plausibly change the ranking. The authors should report mean and standard deviation over at least three to five seeds for all main comparisons and ablations, or fix a seed and state it explicitly.
  3. [Sec. 2.2, Eq. (6)] The training objective L_stage2 = ||z^n - \hat z^n||_1 trains the denoiser to map Gaussian noise to HQ features conditioned on LQ features, and at inference the reverse process starts from pure noise. This is not a 'refinement' of the LQ-degraded distribution in the sense of gradually correcting an input feature; it is a conditional generation from noise. The paper does not explain why this procedure should land near the codebook manifold, and no codebook reconstruction loss or nearest-neighbor constraint is used in Stage II. I ask the authors to show that \hat z^n lies close to the codebook manifold (e.g., nearest-neighbor distances before and after diffusion) and to ablate the initialization choice (Gaussian noise vs. LQ feature plus noise).
  4. [Sec. 3.1, PET Synthesis dataset] The description 'subsampling full scans with a 12x dose reduction, and reconstructed using the standard OSEM method' omits essential details: how the sinogram subsampling is performed, the OSEM iteration and subset counts, whether any post-filtering is applied, and how the high-quality reference is exactly defined. Without these details the PET synthesis result cannot be reproduced or compared fairly against the cited prior work [9], and the quantitative gain on that task is not fully verifiable.
minor comments (3)
  1. [Sec. 1, Contribution 1] The claim of being 'the first to exploit the VQ codebook prior for all-in-one MedIR' should be supported by a brief discussion of prior VQ-based restoration works and why none of them addresses the all-in-one setting; otherwise the novelty claim is too strong.
  2. [Fig. 1] Several labels in the overview figure appear as unreadable '???' placeholders, particularly around the equations and block annotations; please provide a clean vector version so that the architecture can be verified.
  3. [Sec. 3.1 and References] The implementation description gives NAF block counts and channels but omits the exact structure of the condition encoder E_lq, the number of learnable parameters, and the total compute or storage overhead of the codebook bank and diffusion stage; adding a parameter/FLOPs table would strengthen the efficiency claims in the Introduction. Also, reference [18] is missing its publication venue (ICLR 2017).

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the codebook and latent diffusion stages are supervised training with independent held-out evaluation.

full rationale

DiffCode's derivation chain is self-contained. Stage I learns each task codebook via self-reconstruction on HQ images (Eq. 3), so the codebook prior is a trained representation, not an output defined by the final restoration. Stage II trains the latent diffusion denoiser to predict the Stage I HQ features z^n from LQ-conditioned noisy latents with Lstage2 = ||z^n - \hat z^n||_1 (Eq. 6); this is a standard supervised regression target, and the retrieved features are then decoded and used as an input to the Stage III restoration network. The final PSNR/SSIM/RMSE results in Tables 1 and 2 are computed on held-out test data and compared against eleven external baselines, so the central claim is falsifiable and not forced by construction. The only self-citation is [22] for the task-aware routing module; that module is an externally published architectural component and the paper's contribution does not reduce to it. The absence of direct retrieval-accuracy or distribution-alignment measurements weakens the mechanistic evidence for the diffusion component, but that is an evidence gap, not circularity.

Assumptions & free parameters 7 free parameters · 4 assumptions · 2 invented entities

All free parameters are standard hyperparameters taken from cited works or chosen by hand; none are fitted to the test set in a circular way. The main assumptions are that a finite codebook can encode the HQ prior, that a few-step diffusion can map corrupted features to clean ones, and that L1 losses are sufficient for restoration quality. These are domain assumptions that are common in the field but not rigorously justified in the paper.

free parameters (7)
  • codebook size per task (8192 entries, dimension 256) = 8192 x 256
    Chosen by hand following prior work [10]; not fitted to data but an architectural choice that affects performance. It is a free parameter in the sense that the paper does not search over it.
  • number of RQ iterations (8) = 8
    Set following the paper [10]; number of quantization depths, a hand-chosen hyperparameter.
  • diffusion total time steps T (8) = 8
    Following the paper [21]; not tuned, but a key design choice.
  • beta schedule (linear from 0.1 to 0.99) = beta_1=0.1, beta_T=0.99
    Standard DDPM choice, not explored in the paper.
  • commitment loss weight delta (0.25) = 0.25
    Taken directly from [10]; influences codebook training.
  • number of experts in TARM (4) and top-k (1) = 4 experts, k=1
    Following AMIR [22]; the paper does ablate top-k (Table 2b) but only for 1, 2, 3, not systematically for the number of experts.
  • NAF block configuration [2,2,4,4] and channels [64,128,256,256] = NAF blocks, channels per level
    Network width/depth chosen by hand, not ablated.
assumptions (4)
  • domain assumption The VQ codebook learned on clean images captures the full distribution of HQ medical image features necessary for restoration.
    The paper assumes that a codebook of finite size (8192 vectors of dimension 256) is sufficient to encode all relevant HQ details for each task, and that the decoder can reconstruct them. This is a standard assumption in VQ-based methods and is not verified quantitatively in terms of reconstruction quality of the codebook.
  • domain assumption Latent diffusion with T=8 steps can reliably map a random Gaussian latent to the correct HQ prior feature when conditioned on LQ features.
    This is the core premise of Stage II. The paper relies on the result of DiffIR [21] that few-step latent diffusion works for image restoration, but in DiffIR the diffusion is applied in the latent space of a pre-trained autoencoder; here it is applied in the feature space of the codebook encoder, which may behave differently.
  • domain assumption The LQ condition encoder Elq provides sufficient information for the diffusion model to recover task-specific HQ features.
    The condition encoder is only described as extracting features from LQ images; no architecture details are given for Elq. If the LQ features are too corrupted, the diffusion process may not converge to useful priors.
  • domain assumption Using L1 losses only (reconstruction, diffusion, restoration) is sufficient to train a stable, high-quality restoration model.
    The paper does not use adversarial or perceptual losses; it assumes L1 alone gives state-of-the-art results, which is unusual for image synthesis and may limit sharpness or perceptual quality.
invented entities (2)
  • Task-adaptive codebook bank
    purpose: Stores task-specific HQ prior features in a concatenated set of codebooks, one per task, to provide customized compensation for diverse information losses.
    This is a novel architectural component introduced by the paper; it is evaluated only through the final downstream restoration metrics, not through any direct measure of prior retrieval accuracy or codebook utility.
  • Latent diffusion-enhanced prior retrieval
    purpose: Refines the latent feature distribution before codebook lookup, so that LQ features better align with HQ codebook entries.
    A new mechanism in this paper; its effectiveness is only shown via ablation (V3 vs. DiffCode). No independent evidence exists for this component outside the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of All-in-One Medical Image Restoration with Latent Diffusion-Enhanced Vector-Quantized Codebook Prior." pith.science (2026). https://pith.science/paper/LOWQTOFN

@misc{pith2026250719874,
  author       = {Pith},
  title        = {Pith review of: All-in-One Medical Image Restoration with Latent Diffusion-Enhanced Vector-Quantized Codebook Prior},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LOWQTOFN}},
  note         = {Machine review of arXiv:2507.19874}
}
read the original abstract

All-in-one medical image restoration (MedIR) aims to address multiple MedIR tasks using a unified model, concurrently recovering various high-quality (HQ) medical images (e.g., MRI, CT, and PET) from low-quality (LQ) counterparts. However, all-in-one MedIR presents significant challenges due to the heterogeneity across different tasks. Each task involves distinct degradations, leading to diverse information losses in LQ images. Existing methods struggle to handle these diverse information losses associated with different tasks. To address these challenges, we propose a latent diffusion-enhanced vector-quantized codebook prior and develop \textbf{DiffCode}, a novel framework leveraging this prior for all-in-one MedIR. Specifically, to compensate for diverse information losses associated with different tasks, DiffCode constructs a task-adaptive codebook bank to integrate task-specific HQ prior features across tasks, capturing a comprehensive prior. Furthermore, to enhance prior retrieval from the codebook bank, DiffCode introduces a latent diffusion strategy that utilizes the diffusion model's powerful mapping capabilities to iteratively refine the latent feature distribution, estimating more accurate HQ prior features during restoration. With the help of the task-adaptive codebook bank and latent diffusion strategy, DiffCode achieves superior performance in both quantitative metrics and visual quality across three MedIR tasks: MRI super-resolution, CT denoising, and PET synthesis.

Figures

Figures reproduced from arXiv: 2507.19874 by the authors.

Figure 1
Figure 1. Overview of DiffCode. (a) Stage I constructs a codebook bank to encode HQ priors. (b) Stage II leverages latent diffusion strategy to enhance prior retrieval from the codebook bank. (c) Stage III performs restoration guided by retrieved priors. (d) RQ. (e) DM training. (f) Expert routing in TARM. (g) Network Architecture. DMs gradually reconstruct the complex underlying HQ distribution through controlled noise reduc… view at source ↗
Figure 2
Figure 2. Visual comparison between different methods across three MedIR tasks. each scan, we extract 100 central 256×256 slices. LQ images are generated using a 4× scaling factor following the paper [30]. (2) CT Denoising: The LDCT dataset [15] provides paired standard-dose and quarter-dose CT scans. We select 50 chest scans acquired from a Siemens scanner, divided into 40/5/5 for training/validation/testing, and extract 512… view at source ↗
Figure 3
Figure 3. Visual comparison for component analysis of DiffCode across three tasks. The configurations of V1, V2 and V3 are presented in Tab. 2 (a). the PyTorch framework on NVIDIA A100 GPUs. Quantitative performance is evaluated using PSNR, SSIM, and RMSE [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 17 canonical work pages

  1. [9]

    IEEE transactions on medical imaging37(6), 1478–1487 (2018)

    Kim, K., Wu, D., Gong, K., Dutta, J., Kim, J.H., Son, Y.D., Kim, H.K., El Fakhri, G., Li, Q.: Penalized pet reconstruction using deep learning prior and local linear fitting. IEEE transactions on medical imaging37(6), 1478–1487 (2018)

  2. [1]

    https://brain-development.org/ixi-dataset/

    Ixi dataset. https://brain-development.org/ixi-dataset/

  3. [2]

    In: European conference on computer vision

    Chen, L., Chu, X., Zhang, X., Sun, J.: Simple baselines for image restoration. In: European conference on computer vision. pp. 17–33. Springer (2022)

  4. [3]

    In: International Conference on Medical Image Computing and Computer-Assisted Intervention

    Chen, Z., Gao, Q., Zhang, Y., Shan, H.: Ascon: Anatomy-aware supervised con- trastive learning framework for low-dose ct denoising. In: International Conference on Medical Image Computing and Computer-Assisted Intervention. pp. 355–365. Springer (2023)

  5. [4]

    In: Proceedings of the IEEE/CVF winter conference on applications of computer vision

    Georgescu, M.I., Ionescu, R.T., Miron, A.I., Savencu, O., Ristea, N.C., Verga, N., Khan, F.S.: Multimodal multi-head convolutional attention with various kernel sizes for medical image super-resolution. In: Proceedings of the IEEE/CVF winter conference on applications of computer vision. pp. 2195–2205 (2023)

  6. [5]

    Advances in neural information processing systems33, 6840–6851 (2020)

    Ho, J., Jain, A., Abbeel, P.: Denoising diffusion probabilistic models. Advances in neural information processing systems33, 6840–6851 (2020)

  7. [6]

    Huang, Z., Zhang, J., Zhang, Y., Shan, H.: Du-gan: Generative adversarial net- workswithdual-domainu-net-baseddiscriminatorsforlow-dosectdenoising.IEEE Transactions on Instrumentation and Measurement71, 1–12 (2021)

  8. [7]

    IEEE transactions on medical imaging 13(4), 601–609 (1994)

    Hudson, H.M., Larkin, R.S.: Accelerated image reconstruction using ordered sub- sets of projection data. IEEE transactions on medical imaging 13(4), 601–609 (1994)

Show all 31 references
  1. [8]

    IEEE transactions on medical imaging (2023) 10 Chen et al

    Jang, S.I., Pan, T., Li, Y., Heidari, P., Chen, J., Li, Q., Gong, K.: Spach trans- former: Spatial and channel-wise transformer based on local and global self- attentions for pet image denoising. IEEE transactions on medical imaging (2023) 10 Chen et al

  2. [10]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Lee, D., Kim, C., Kim, S., Cho, M., Han, W.S.: Autoregressive image genera- tion using residual quantization. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 11523–11532 (2022)

  3. [11]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Li, B., Liu, X., Hu, P., Wu, Z., Lv, J., Peng, X.: All-in-one image restoration for unknown corruption. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 17452–17462 (2022)

  4. [12]

    In: Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition

    Li, G., Rao, C., Mo, J., Zhang, Z., Xing, W., Zhao, L.: Rethinking diffusion model for multi-contrast mri super-resolution. In: Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition. pp. 11365–11374 (2024)

  5. [13]

    Medical Image Analysis77, 102335 (2022)

    Luo, Y., Zhou, L., Zhan, B., Fei, Y., Zhou, J., Wang, Y., Shen, D.: Adaptive rectification based adversarial network with spectrum constraint for high-quality pet image synthesis. Medical Image Analysis77, 102335 (2022)

  6. [14]

    In: Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Luo, Z., Gustafsson, F.K., Zhao, Z., Sjölund, J., Schön, T.B.: Refusion: Enabling large-size realistic image restoration with latent-space diffusion models. In: Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 1680–1691 (2023)

  7. [15]

    https://doi.org/10.7937/9NPB-2637

    McCollough, C., Chen, B., Holmes III, D.R., Duan, X., Yu, Z., Yu, L., Leng, S., Fletcher, J.: Low dose ct image and projection data (ldct-and-projection-data) (version 6) [data set] (2020). https://doi.org/10.7937/9NPB-2637

  8. [16]

    arXiv preprint arXiv:2409.13094 (2024)

    Öztürk, Ş., Duran, O.C., Çukur, T.: Denomamba: A fused state-space model for low-dose ct denoising. arXiv preprint arXiv:2409.13094 (2024)

  9. [17]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Rombach, R., Blattmann, A., Lorenz, D., Esser, P., Ommer, B.: High-resolution image synthesis with latent diffusion models. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 10684–10695 (2022)

  10. [18]

    Outrageously large neural networks (2017)

    Shazeer, N., Mirhoseini, A., Maziarz, K., Davis, A., Le, Q., Hinton, G., Dean, J.: The sparsely-gated mixture-of-experts layer. Outrageously large neural networks (2017)

  11. [19]

    Advances in neural information processing systems30 (2017)

    Van Den Oord, A., Vinyals, O., et al.: Neural discrete representation learning. Advances in neural information processing systems30 (2017)

  12. [20]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Wang, Z., Cun, X., Bao, J., Zhou, W., Liu, J., Li, H.: Uformer: A general u-shaped transformer for image restoration. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 17683–17693 (2022)

  13. [21]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Xia, B., Zhang, Y., Wang, S., Wang, Y., Wu, X., Tian, Y., Yang, W., Van Gool, L.: Diffir: Efficient diffusion model for image restoration. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 13095–13105 (2023)

  14. [22]

    In: International Con- ference on Medical Image Computing and Computer-Assisted Intervention

    Yang, Z., Chen, H., Qian, Z., Yi, Y., Zhang, H., Zhao, D., Wei, B., Xu, Y.: All- in-one medical image restoration via task-adaptive routing. In: International Con- ference on Medical Image Computing and Computer-Assisted Intervention. pp. 67–77. Springer (2024)

  15. [23]

    In: International Conference on Medical Image Computing and Computer-Assisted Intervention

    Yang, Z., Zhou, Y., Zhang, H., Wei, B., Fan, Y., Xu, Y.: Drmc: A generalist model with dynamic routing for multi-center pet image synthesis. In: International Conference on Medical Image Computing and Computer-Assisted Intervention. pp. 36–46. Springer (2023)

  16. [24]

    IEEE Transactions on Image Processing (2024) Title Suppressed Due to Excessive Length 11

    Yao, M., Xu, R., Guan, Y., Huang, J., Xiong, Z.: Neural degradation representation learning for all-in-one image restoration. IEEE Transactions on Image Processing (2024) Title Suppressed Due to Excessive Length 11

  17. [25]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

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

  18. [26]

    In: Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition

    Zamir, S.W., Arora, A., Khan, S., Hayat, M., Khan, F.S., Yang, M.H., Shao, L.: Multi-stage progressive image restoration. In: Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition. pp. 14821–14831 (2021)

  19. [27]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Zhang, J., Huang, J., Yao, M., Yang, Z., Yu, H., Zhou, M., Zhao, F.: Ingredient- oriented multi-degradation learning for image restoration. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 5825– 5835 (2023)

  20. [28]

    IEEE transactions on medical imaging40(3), 805–817 (2020)

    Zhao, C., Dewey, B.E., Pham, D.L., Calabresi, P.A., Reich, D.S., Prince, J.L.: Smore: a self-supervised anti-aliasing and super-resolution algorithm for mri using deep learning. IEEE transactions on medical imaging40(3), 805–817 (2020)

  21. [29]

    In: Proceedings of the European Conference on Computer Vision (ECCV)

    Zhao, X., Li, H., Shen, X., Liang, X., Wu, Y.: A modulation module for multi- task learning with applications in image retrieval. In: Proceedings of the European Conference on Computer Vision (ECCV). pp. 401–416 (2018)

  22. [30]

    IEEE transactions on image processing28(11), 5649–5662 (2019)

    Zhao, X., Zhang, Y., Zhang, T., Zou, X.: Channel splitting network for single mr image super-resolution. IEEE transactions on image processing28(11), 5649–5662 (2019)

  23. [31]

    Medical image analysis65, 101770 (2020)

    Zhou, L., Schaefferkoetter, J.D., Tham, I.W., Huang, G., Yan, J.: Supervised learn- ing with cyclegan for low-dose fdg pet image denoising. Medical image analysis65, 101770 (2020)

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.