Pith. sign in

REVIEW 4 major objections 6 minor 60 references

Diffusion Prior Interpolation for Flexibility Real-World Face Super-Resolution

T0 review · 4 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A masked diffusion sampler claims state-of-the-art real-world face super-resolution with a user-adjustable consistency–diversity trade-off.

desk verdict Solid incremental face SR method with a real weakness: the headline consistency gain is measured on synthetic data generated from the same degradation model used to train its corrector. read the letter →

arxiv 2412.16552 v1 pith:6Q3UYQCJ submitted 2024-12-21 cs.CV cs.AI

Jiarui Yang , Tao Dai , Yufei Zhu , Naiqi Li , Jinmin Li , Shutao Xia This is my paper · ORCID
classification cs.CVcs.AI
keywords diffusionpriorinterpolationfacesuper-resolutionblindrestorationconditionmaskingposteriorsamplingcorrectormodelsrecognitionconsistency
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

The paper introduces Diffusion Prior Interpolation (DPI), a way to steer a pre-trained diffusion model for blind face super-resolution without retraining the generator. It claims that by masking the sampling process with a fixed grid condition in early steps and a random edge-adaptive condition in later steps, faces can be reconstructed with both stable identity and rich texture. A small condition corrector network refines the conditioning signal during sampling, and DPI reports the best perceptual metrics among compared face super-resolution methods on synthetic and real-world benchmarks, including face recognition consistency at extreme 16x upscaling.

What carries the argument

The load-bearing mechanism is the two-mask interpolation of posterior samples plus the condition corrector (CRT). A fixed condition mask (FCM) places low-resolution pixel values on a grid; a randomly adaptive condition mask (RACM) stochastically selects grid sites weighted by the normalized Laplacian edge map of the current condition, so high-frequency face regions receive more conditioning. CRT is a small U-Net that takes the masked noisy condition, the initial condition, and the time step, predicts the clean masked ground-truth condition, and is trained with a loss blending prior and gap terms; it is applied at every sampling step to pull conditions back toward the prior manifold. The scalars $\tau$, $s$, and $\omega$ are the user dials that trade consistency against diversity.

What would settle it

Take a paired real-world benchmark with ground-truth high-resolution faces degraded by camera optics and in-the-wild compression that are not drawn from the paper's degradation model, run DPI without retraining the corrector, and measure face recognition accuracy at 16x; if accuracy falls to the level of unguided diffusion baselines rather than the reported 72.8% on CelebA1000, the transfer claim collapses.

Watch

Extended reading notes

Core claim

The central claim is that strong and weak condition masks, interpolated through a pre-trained diffusion denoiser, can jointly satisfy the two competing requirements of face super-resolution: pixel-level consistency with the low-resolution input and high-fidelity generation of facial details. DPI divides sampling into two stages separated by a scalar tau: for t >= tau a fixed grid mask enforces structural consistency, and for t < tau a randomly adaptive mask derived from the face edge map opens the prior space for texture and diversity. The condition corrector updates mask values with a small U-Net on the fly, establishing a reciprocal loop in which samples refine conditions and conditions refine samples. The paper reports that this design beats prior face restoration methods, and at 16x upscaling it raises face recognition accuracy on a 1000-face test set from 59.8% to 72.8%.

Load-bearing premise

The load-bearing premise is that the condition corrector, trained on faces degraded by the paper's synthetic blur–noise–JPEG model, generalizes to real-world degradations; if real-world blur, noise, or compression falls outside that model's span, the corrected conditions become inaccurate and the claimed identity-consistency gains may shrink.

Editorial extensions

If this is right

  • DPI can be layered onto any already-trained DDPM or DDIM sampler without retraining the generative model, so it inherits the prior's quality while gaining conditioning at small extra computational cost.
  • Users can tune consistency against diversity by adjusting $\tau$, $s$, and $\omega$, with the same pipeline covering mild, medium, and heavy real-world degradations.
  • The method reports best or second-best perceptual metrics (LPIPS, FID, IS, PSNR) on CelebA1000 and FFHQ1000 across x4, x8, and x16 upscaling factors.
  • On face-recognition consistency checks, DPI reports the highest accuracy and lowest feature distance among compared methods, with the largest gain at 16x upscaling.
  • The light-weight corrector adds only a modest per-step overhead and works with accelerated DDIM sampling, not just full DDPM sampling.

Reading between the lines

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

  • Because the fixed mask is designed around facial structure, the same recipe is unlikely to preserve arbitrary natural-image backgrounds; the paper itself notes that consistency weakens on natural images.
  • CRT is trained under a specific synthetic degradation model, so DPI's real-world gains may depend on that degradation family; a shift in blur, noise, or JPEG compression may require retraining the corrector.
  • The reported recognition-consistency advantage at 16x rests on one face recognition model and threshold, so the identity-preservation claim should be probed with other matchers before treating it as general.
  • A testable extension is to apply DPI to face video or identity-preserving face generation, where the mask could be shared across frames and the corrector updated online.
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

4 major / 6 minor

Summary. The paper proposes Diffusion Prior Interpolation (DPI) for blind face super-resolution using a pre-trained unconditional DDPM. The method splits diffusion sampling into two stages controlled by a scalar τ: a strong Fixed Condition Mask (FCM) stage and a weaker Randomly Adaptive Condition Mask (RACM) stage, with a learned condition Corrector (CRT) that iteratively refines the condition during sampling. Experiments compare DPI with prior-based and trained diffusion FSR methods on synthetic CelebA1000/FFHQ1000 at ×4/×8/×16, on real-world LFW/WebPhoto/WIDER with no-reference metrics, and on face-recognition consistency (ACC/CS) for the synthetic datasets. The paper reports code release and several ablations of the proposed components.

Significance. If the method works as claimed, it is a practically valuable plug-and-play face-super-resolution approach: it avoids retraining the diffusion backbone, uses a small auxiliary corrector, and offers a tunable consistency-diversity trade-off. The paper ships code and evaluates against a wide set of recent baselines, which is a strength. The central limitations are that the real-world consistency claim is not directly tested, and the key approximation in Eq. (16) rests on an imported assumption that is not re-validated in the DPI setting.

major comments (4)
  1. [Real-world Datasets and Face Recognition Results (Tables 2, 3, 6)] The central claim that DPI achieves higher consistency in the real-world setting is not supported by the reported evaluation. Table 3 and Table 6 report ACC and CS only on CelebA1000 and FFHQ1000, whose LR inputs are generated with the same degradation protocols used to train CRT at each scale, so these are in-domain results. The real-world datasets in Table 2 (LFW, WebPhoto, WIDER) are evaluated only with no-reference perceptual metrics; no identity-consistency or paired-HR metric is reported there. Since CRT is trained on the synthetic degradation model of Eq. (19) and real-world degradations include unknown kernels, compression, sensor noise, and misalignment outside that span, the 72.8% ACC at ×16 on CelebA1000 is not evidence that the consistency advantage transfers. Please add a real-world identity-consistency evaluation (for example on identity-labeled face data where a higher-quality original can serve as reference) or explicitly restrict the consistency claim to the in-domain synthetic setting.
  2. [Condition Corrector, Eq. (16)] Equation (16) imports the assumption from Wang et al. 2023d that q(xt|x) is close to q(yt|y) for t>γ and applies it to the corrected condition after each CRT step, but this is not re-derived or validated in the DPI setting. Once CRT updates yt, the condition is no longer the forward-noised initial LR; it is the output of a learned mapping, so the posterior gap can change in a way that the original DR2 assumption does not cover. Since this approximation underpins the reciprocal refinement between condition and sample, please provide a derivation under the CRT update or an empirical check (for example, measure the posterior distance over t on held-out real degradations). As written, Eq. (16) is an unproven load-bearing step.
  3. [Comparison with Previous Work, Tables 1, 2, 7] The abstract and conclusion claim superiority over SOTA FSR methods, but the reported numbers are not uniformly in DPI's favor. In Table 2 on LFW, CodeFormer has MUSIQ 67.30 vs DPI 66.13 and PGDiff has CLIPIQA 0.7137 vs DPI 0.7098; on WIDER, PGDiff has FID 47.63 vs DPI 49.79. No error bars or multiple-seed results are reported for the stochastic sampling procedures, so the magnitude of the reported differences is not quantified. Please qualify the performance claim per metric and add variance estimates for the main comparisons.
  4. [Condition Corrector, Eqs. (17)-(18)] The training procedure for the gap-correcting term is underspecified. Eq. (17) uses x̂crt, the intermediate condition output of CRT, as input, but the text does not state how x̂crt is generated: whether by a frozen teacher, stop-gradient, or unrolled sampling, and how the schedule for Ω(t) is chosen. Without this detail, the CRT training objective in Eq. (18) is not reproducible, and the claim that CRT closes the distribution gap cannot be checked.
minor comments (6)
  1. [Tables 3 and 6] The headers describing the face-recognition comparisons as 'Qualitative' are incorrect; these are quantitative comparisons.
  2. [Table 4] The columns of Table 4 are not labeled; the checkmark-only rows do not identify which variant each column refers to.
  3. [Eq. (10)] The condition 'i, j mod k = 0' should be spelled out as 'i mod k = 0 and j mod k = 0' to avoid ambiguity.
  4. [Related Work] The sentence 'Previous diffusion-based works have not addressed real-world face super-resolution' is contradicted by the paper's own real-world comparisons with PGDiff and DiffFace.
  5. [Eq. (6)] The condition is noised with the model's predicted noise εθ(xt,t); clarify why this is preferred over an independent Gaussian sample for aligning the condition with the posterior.
  6. [Eqs. (15)-(16)] The notation for the ground-truth posterior is inconsistent (IG, I′Gt−1, I′G,t−1); please define it once and use it consistently.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: DPI's masking and corrector are explicitly constructed and evaluated on external benchmarks, with no load-bearing self-citation or fitted-parameter-as-prediction.

full rationale

The derivation chain of DPI is self-contained in the direction claimed. The condition masks in Eqs. 8-9 are explicit algebraic combinations of posterior samples and are not defined in terms of the final SR output. The CRT is trained with a supervised pixel loss (Eqs. 15 and 17) on FFHQ ground truth under the degradation model of Eq. 19, so its correction is a learned function rather than a fitted constant or a renamed version of the target metric. The face-recognition consistency results in Tables 3 and 6 are computed on held-out CelebA1000 and FFHQ1000 test sets using the external DeepFace model; although the LR inputs are generated with the same degradation family used to train CRT, this is in-distribution evaluation, not circular prediction. The time-step separation assumption is imported from DR2 (Wang et al. 2023d), an external prior work, and is not a self-citation; even if that assumption fails to transfer to real-world degradations, the failure would be a generalization risk, not a circular derivation. No equation reduces to its own input, and no fitted parameter is renamed as a prediction. The paper's empirical claims are therefore supported by independent evaluation rather than by definitional equivalence.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central claim rests on standard DDPM Gaussianity plus three domain assumptions: the validity of the pre-trained FFHQ prior, the time-step-separation property imported from Wang et al. 2023d, and the transferability of CRT trained on a synthetic degradation model to real-world faces. Four scalar hyperparameters (tau, s, omega, k) are hand-set per scale.

free parameters (4)
  • tau = 100 (x4), 300 (x8), 500 (x16), 500 (real)
    Divides sampling into fixed-mask and adaptive-mask stages; chosen per scale/dataset.
  • s = 1.4 (x4), 1.2 (x8), 1.0 (x16/real)
    Controls probability for random adaptive mask edge guidance.
  • omega = 500, 750, 1000 per scale
    Sets time-dependent weight w=t/omega for the adaptive mask condition.
  • k = 2
    Sparsity of the condition mask, fixed for all experiments.
assumptions (4)
  • standard math The DDPM reverse process is Gaussian with mean and variance from the denoiser (Eq. 3).
    Standard DDPM assumption used throughout the method.
  • domain assumption There exists a time step gamma such that for t > gamma, the conditional posterior q(xt|x) is close to the true posterior q(xt).
    Invoked from Wang et al. 2023d without proof in this paper, and used to justify the CRT approximation in Eq. 16.
  • domain assumption The pre-trained FFHQ DDPM provides a valid face prior for CelebA and real-world faces.
    The method relies on the pre-trained DPS model generalizing to test faces.
  • domain assumption CRT trained on Eq. 19 degradations transfers to real-world LFW, WebPhoto, and WIDER.
    The real-world evaluation assumes the synthetic degradation model spans real degradations.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Diffusion Prior Interpolation for Flexibility Real-World Face Super-Resolution." pith.science (2026). https://pith.science/paper/6Q3UYQCJ

@misc{pith2026241216552,
  author       = {Pith},
  title        = {Pith review of: Diffusion Prior Interpolation for Flexibility Real-World Face Super-Resolution},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6Q3UYQCJ}},
  note         = {Machine review of arXiv:2412.16552}
}
read the original abstract

Diffusion models represent the state-of-the-art in generative modeling. Due to their high training costs, many works leverage pre-trained diffusion models' powerful representations for downstream tasks, such as face super-resolution (FSR), through fine-tuning or prior-based methods. However, relying solely on priors without supervised training makes it challenging to meet the pixel-level accuracy requirements of discrimination task. Although prior-based methods can achieve high fidelity and high-quality results, ensuring consistency remains a significant challenge. In this paper, we propose a masking strategy with strong and weak constraints and iterative refinement for real-world FSR, termed Diffusion Prior Interpolation (DPI). We introduce conditions and constraints on consistency by masking different sampling stages based on the structural characteristics of the face. Furthermore, we propose a condition Corrector (CRT) to establish a reciprocal posterior sampling process, enhancing FSR performance by mutual refinement of conditions and samples. DPI can balance consistency and diversity and can be seamlessly integrated into pre-trained models. In extensive experiments conducted on synthetic and real datasets, along with consistency validation in face recognition, DPI demonstrates superiority over SOTA FSR methods. The code is available at \url{https://github.com/JerryYann/DPI}.

Figures

Figures reproduced from arXiv: 2412.16552 by the authors.

Figure 1
Figure 1. Visualization of FSR results using different types of methods, including those based on discriminative models (c [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Graphical model of Diffusion Prior Interpolation. ⊙ represents element-wise matrix multiplication. xT and yT correspond to the initial random noise and the initial condition respectively. We use the scalar τ to divide the sampling process into two stages. CRT is the Corrector function that applies Eq. 16 to CMs correction. yt represents the intermediate condition. After posterior sampling, y ′ t is multiplied by mf … view at source ↗
Figure 10
Figure 10. The specific definition of mf is as follows: mf (i, j) =  1, if i, j mod k = 0 0, otherwise (10) where mf ∈ RH×W . When k is set to 2, the form of mf resembles a grid mask (Chen et al. 2020) with a grid size of 1 pixel. We project I bc L onto mf to generate an initial condition, denoted as yT : yT (i, j) =    I bc L (  i k  ,  j k  ), if mf (i, j) = 1 0, otherwise (11) where I bc L ∈ R H k × W k and yT ∈ RH×… view at source ↗
Figures from the paper (13 more)
Figure 3
Figure 3. Figure 3: Condition Refinement. We present the initial con￾dition yT along with the refined conditions in the intermedi￾ate stages (y500, y50). Alongside the diffusion sampling, the conditions are updated with CRT. Refinement of the condi￾tions contributes to a more effective sa…
Figure 4
Figure 4. Figure 4: From top to bottom are the FSR results of DPI and DDPM-based methods on [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Qualitative comparisons on WebPhoto dataset. Please zoom in for best view. w=0.0 w=0.3 w=0.6 w=0.9 Input CodeFormer DPI Sample 1 DPI Sample 2 DPI Sample 3 DPI Sample 4 [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Diversity Comparison in WIDER dataset. DPI allows users to adjust hyperparameters to generate diverse results. Note that DPI can also be configured with hyperparameters to ensure consistency. Please zoom in for best view. Datasets LFW WebPhoto WIDER Degradation mild me…
Figure 7
Figure 7. Figure 7: Qualitative comparison on heavy degradation. [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 8
Figure 8. Figure 8: Visual comparisons of the impact of hyperparame [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]
Figure 9
Figure 9. Figure 9: Illustrative architecture of Corrector. tial moving average (EMA) decay rate of 0.9999. The Py￾Torch (Paszke et al. 2019) framework is employed, and the training is conducted in parallel on 3080Ti GPUs with a batch size of 32. For all experiments, CRT is trained on the…
Figure 10
Figure 10. Figure 10: Quantitative comparison at different steps on [PITH_FULL_IMAGE:figures/full_fig_p012_10.png]
Figure 11
Figure 11. Figure 11: FSR visualizations of DPI and other diffusion-based methods on [PITH_FULL_IMAGE:figures/full_fig_p013_11.png]
Figure 12
Figure 12. Figure 12: Visualizations of Ablation Studies. inal models. The differences in their complexity primarily arise from the following aspects: • Conditioning Mechanism: The introduction of condition￾ing adds extra computational and memory overhead dur￾ing sampling steps. The comple…
Figure 13
Figure 13. Figure 13: Qualitative comparisons on LFW dataset. Input DR2 GFPGAN CodeFormer DiffFace DPI [PITH_FULL_IMAGE:figures/full_fig_p014_13.png]
Figure 14
Figure 14. Figure 14: Qualitative comparisons on WIDER dataset. ×4 ×8 ×16 Methods CS↓ ACC↑ CS↓ ACC↑ CS↓ ACC↑ Bicubic 0.1295 85.1 0.3226 74.2 0.5844 11.3 CodeFormer 0.0806 93.2 0.1935 91.9 0.3109 80.5 DDRM 0.1420 88.8 0.2383 84.6 0.3414 72.5 SR3 0.0627 94.2 0.1797 92.0 0.3222 70.1 ILVR 0.09…
Figure 15
Figure 15. Figure 15: Visualization of general image super-resolution. The top and bottom images correspond to the LR and SR images, [PITH_FULL_IMAGE:figures/full_fig_p015_15.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

60 extracted references · 26 canonical work pages

  1. [1]

    Bartlett, M. S. 1978. An introduction to stochastic processes: with special reference to methods and applications. CUP Archive

    work page 1978

  2. [2]

    Blau, Y.; and Michaeli, T. 2018. The perception-distortion tradeoff. In Proceedings of the IEEE conference on computer vision and pattern recognition, 6228--6237

    work page 2018

  3. [3]

    Bredell, G.; Flouris, K.; Chaitanya, K.; Erdil, E.; and Konukoglu, E. 2023. Explicitly Minimizing the Blur Error of Variational Autoencoders. arXiv preprint arXiv:2304.05939

    arXiv 2023

  4. [4]

    Brock, A.; Donahue, J.; and Simonyan, K. 2018. Large Scale GAN Training for High Fidelity Natural Image Synthesis. In International Conference on Learning Representations

    work page 2018

  5. [5]

    C.; Xu, X.; Wang, X.; Gu, J.; and Loy, C

    Chan, K. C.; Xu, X.; Wang, X.; Gu, J.; and Loy, C. C. 2022. GLEAN: Generative latent bank for image super-resolution and beyond. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(3): 3154--3168

    work page 2022

  6. [6]

    Chen, C.; Li, X.; Yang, L.; Lin, X.; Zhang, L.; and Wong, K.-Y. K. 2021. Progressive semantic-aware style transformation for blind face restoration. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 11896--11905

    2021

  7. [7]

    Chen, P.; Liu, S.; Zhao, H.; and Jia, J. 2020. Gridmask data augmentation. arXiv preprint arXiv:2001.04086

    arXiv 2020

  8. [8]

    Chen, Y.; Tai, Y.; Liu, X.; Shen, C.; and Yang, J. 2018. Fsrnet: End-to-end learning face super-resolution with facial priors. In Proceedings of the IEEE conference on computer vision and pattern recognition, 2492--2501

    2018

Show all 60 references
  1. [9]

    Choi, J.; Kim, S.; Jeong, Y.; Gwon, Y.; and Yoon, S. 2021. ILVR: Conditioning Method for Denoising Diffusion Probabilistic Models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 14367--14376

    2021

  2. [10]

    T.; Klasky, M

    Chung, H.; Kim, J.; Mccann, M. T.; Klasky, M. L.; and Ye, J. C. 2022. Diffusion Posterior Sampling for General Noisy Inverse Problems. In The Eleventh International Conference on Learning Representations

    2022

  3. [11]

    Dhariwal, P.; and Nichol, A. 2021. Diffusion models beat gans on image synthesis. Advances in neural information processing systems, 34: 8780--8794

    2021

  4. [12]

    Fei, B.; Lyu, Z.; Pan, L.; Zhang, J.; Yang, W.; Luo, T.; Zhang, B.; and Dai, B. 2023. Generative Diffusion Prior for Unified Image Restoration and Enhancement. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 9935--9946

    2023

  5. [13]

    Gao, G.; Xu, Z.; Li, J.; Yang, J.; Zeng, T.; and Qi, G.-J. 2023. Ctcnet: A cnn-transformer cooperation network for face image super-resolution. IEEE Transactions on Image Processing, 32: 1978--1991

    2023

  6. [14]

    Gu, Y.; Wang, X.; Xie, L.; Dong, C.; Li, G.; Shan, Y.; and Cheng, M.-M. 2022. VQFR: Blind face restoration with vector-quantized dictionary and parallel decoder. In European Conference on Computer Vision, 126--143. Springer

    2022

  7. [15]

    Ho, J.; Jain, A.; and Abbeel, P. 2020. Denoising diffusion probabilistic models. Advances in neural information processing systems, 33: 6840--6851

    2020

  8. [16]

    B.; Mattar, M.; Berg, T.; and Learned-Miller, E

    Huang, G. B.; Mattar, M.; Berg, T.; and Learned-Miller, E. 2008. Labeled faces in the wild: A database forstudying face recognition in unconstrained environments. In Workshop on faces in'Real-Life'Images: detection, alignment, and recognition

    2008

  9. [17]

    Jiang, J.; Wang, C.; Liu, X.; and Ma, J. 2021. Deep learning-based face super-resolution: A survey. ACM Computing Surveys (CSUR), 55(1): 1--36

    2021

  10. [18]

    Karras, T.; Laine, S.; and Aila, T. 2019. A style-based generator architecture for generative adversarial networks. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 4401--4410

    2019

  11. [19]

    Kawar, B.; Elad, M.; Ermon, S.; and Song, J. 2022. Denoising diffusion restoration models. Advances in Neural Information Processing Systems, 35: 23593--23606

    2022

  12. [20]

    Kim, K.; Kim, Y.; Cho, S.; Seo, J.; Nam, J.; Lee, K.; Kim, S.; and Lee, K. 2022. Diffface: Diffusion-based face swapping with facial guidance. arXiv preprint arXiv:2212.13344

    2022 arXiv

  13. [21]

    P.; and Ba, J

    Kingma, D. P.; and Ba, J. 2014. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980

    2014 arXiv

  14. [22]

    Kirichenko, P.; Izmailov, P.; and Wilson, A. G. 2020. Why normalizing flows fail to detect out-of-distribution data. Advances in neural information processing systems, 33: 20578--20589

    2020

  15. [23]

    Leng, J.; and Wang, Y. 2022. RCNet: Recurrent collaboration network guided by facial priors for face super-resolution. In 2022 IEEE International Conference on Multimedia and Expo (ICME), 01--06. IEEE

    2022

  16. [24]

    Li, H.; Yang, Y.; Chang, M.; Chen, S.; Feng, H.; Xu, Z.; Li, Q.; and Chen, Y. 2022. Srdiff: Single image super-resolution with diffusion probabilistic models. Neurocomputing, 479: 47--59

    2022

  17. [25]

    Li, X.; Chen, C.; Zhou, S.; Lin, X.; Zuo, W.; and Zhang, L. 2020. Blind face restoration via deep multi-scale component dictionaries. In European conference on computer vision, 399--415. Springer

    2020

  18. [26]

    Liu, Z.; Luo, P.; Wang, X.; and Tang, X. 2015. Deep learning face attributes in the wild. In Proceedings of the IEEE international conference on computer vision, 3730--3738

    2015

  19. [27]

    Liu, Z.-S.; Siu, W.-C.; and Wang, L.-W. 2021. Variational autoencoder for reference based image super-resolution. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 516--525

    2021

  20. [28]

    Lu, T.; Wang, Y.; Zhang, Y.; Jiang, J.; Wang, Z.; and Xiong, Z. 2022. Rethinking prior-guided face super-resolution: a new paradigm with facial component prior. IEEE Transactions on Neural Networks and Learning Systems

    2022

  21. [29]

    Lugmayr, A.; Danelljan, M.; Romero, A.; Yu, F.; Timofte, R.; and Van Gool, L. 2022. Repaint: Inpainting using denoising diffusion probabilistic models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 11461--11471

    2022

  22. [30]

    Lugmayr, A.; Danelljan, M.; Van Gool, L.; and Timofte, R. 2020. Srflow: Learning the super-resolution space with normalizing flow. In Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part V 16, 715--732. Springer

    2020

  23. [31]

    G.; and Patel, V

    Mei, K.; Nair, N. G.; and Patel, V. M. 2022. Bi-Noising Diffusion: Towards Conditional Diffusion Models with Generative Restoration Priors. arXiv preprint arXiv:2212.07352

    2022 arXiv

  24. [32]

    Menon, S.; Damian, A.; Hu, S.; Ravi, N.; and Rudin, C. 2020. Pulse: Self-supervised photo upsampling via latent space exploration of generative models. In Proceedings of the ieee/cvf conference on computer vision and pattern recognition, 2437--2445

    2020

  25. [33]

    Moser, B.; Frolov, S.; Raue, F.; Palacio, S.; and Dengel, A. 2023. DWA: Differential Wavelet Amplifier for Image Super-Resolution. arXiv preprint arXiv:2304.01994

    2023 arXiv

  26. [34]

    Q.; and Dhariwal, P

    Nichol, A. Q.; and Dhariwal, P. 2021. Improved denoising diffusion probabilistic models. In International Conference on Machine Learning, 8162--8171. PMLR

    2021

  27. [35]

    Paszke, A.; Gross, S.; Massa, F.; Lerer, A.; Bradbury, J.; Chanan, G.; Killeen, T.; Lin, Z.; Gimelshein, N.; Antiga, L.; et al. 2019. Pytorch: An imperative style, high-performance deep learning library. Advances in neural information processing systems, 32

    2019

  28. [36]

    Ronneberger, O.; Fischer, P.; and Brox, T. 2015. U-net: Convolutional networks for biomedical image segmentation. In Medical Image Computing and Computer-Assisted Intervention--MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part II...

    2015

  29. [37]

    J.; and Norouzi, M

    Saharia, C.; Ho, J.; Chan, W.; Salimans, T.; Fleet, D. J.; and Norouzi, M. 2023. Image Super-Resolution via Iterative Refinement. IEEE Transactions on Pattern Analysis & Machine Intelligence, 45(04): 4713--4726

    2023

  30. [38]

    S.; Scholkopf, B.; and Hirsch, M

    Sajjadi, M. S.; Scholkopf, B.; and Hirsch, M. 2017. Enhancenet: Single image super-resolution through automated texture synthesis. In Proceedings of the IEEE international conference on computer vision, 4491--4500

    2017

  31. [39]

    I.; and Ozpinar, A

    Serengil, S. I.; and Ozpinar, A. 2020. Lightface: A hybrid deep face recognition framework. In 2020 innovations in intelligent systems and applications conference (ASYU), 1--5. IEEE

    2020

  32. [40]

    Shang, S.; Shan, Z.; Liu, G.; and Zhang, J. 2023. Resdiff: Combining cnn and diffusion model for image super-resolution. arXiv preprint arXiv:2303.08714

    2023 arXiv

  33. [41]

    Song, J.; Meng, C.; and Ermon, S. 2020. Denoising Diffusion Implicit Models. In International Conference on Learning Representations

    2020

  34. [42]

    P.; Kumar, A.; Ermon, S.; and Poole, B

    Song, Y.; Sohl-Dickstein, J.; Kingma, D. P.; Kumar, A.; Ermon, S.; and Poole, B. 2020. Score-based generative modeling through stochastic differential equations. arXiv preprint arXiv:2011.13456

    2020 arXiv

  35. [43]

    Wang, C.; Jiang, J.; Zhong, Z.; and Liu, X. 2022 a . Propagating facial prior knowledge for multitask learning in face super-resolution. IEEE Transactions on Circuits and Systems for Video Technology, 32(11): 7317--7331

    2022

  36. [44]

    Wang, C.; Jiang, J.; Zhong, Z.; and Liu, X. 2023 a . Spatial-Frequency Mutual Learning for Face Super-Resolution. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 22356--22366

    2023

  37. [45]

    C.; and Loy, C

    Wang, J.; Yue, Z.; Zhou, S.; Chan, K. C.; and Loy, C. C. 2023 b . Exploiting Diffusion Prior for Real-World Image Super-Resolution. arXiv preprint arXiv:2305.07015

    2023 arXiv

  38. [46]

    C.; and Loy, C

    Wang, J.; Yue, Z.; Zhou, S.; Chan, K. C.; and Loy, C. C. 2024. Exploiting diffusion prior for real-world image super-resolution. International Journal of Computer Vision, 1--21

    2024

  39. [47]

    Wang, X.; Li, Y.; Zhang, H.; and Shan, Y. 2021. Towards real-world blind face restoration with generative facial prior. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 9168--9178

    2021

  40. [48]

    Wang, Y.; Hu, Y.; Yu, J.; and Zhang, J. 2023 c . Gan prior based null-space learning for consistent super-resolution. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, 2724--2732

    2023

  41. [49]

    Wang, Y.; Yu, J.; and Zhang, J. 2022. Zero-shot image restoration using denoising diffusion null-space model. arXiv preprint arXiv:2212.00490

    2022 arXiv

  42. [50]

    Wang, Z.; Zhang, J.; Chen, R.; Wang, W.; and Luo, P. 2022 b . Restoreformer: High-quality blind face restoration from undegraded key-value pairs. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 17512--17521

    2022

  43. [51]

    Wang, Z.; Zhang, Z.; Zhang, X.; Zheng, H.; Zhou, M.; Zhang, Y.; and Wang, Y. 2023 d . DR2: Diffusion-based Robust Degradation Remover for Blind Face Restoration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 1704--1713

    2023

  44. [52]

    Wei, M.; Shen, Y.; Wang, Y.; Xie, H.; and Wang, F. L. 2023. RainDiffusion: When Unsupervised Learning Meets Diffusion Models for Real-world Image Deraining. arXiv preprint arXiv:2301.09430

    2023 arXiv

  45. [53]

    G.; and Milanfar, P

    Whang, J.; Delbracio, M.; Talebi, H.; Saharia, C.; Dimakis, A. G.; and Milanfar, P. 2022. Deblurring via stochastic refinement. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 16293--16303

    2022

  46. [54]

    Yang, P.; Zhou, S.; Tao, Q.; and Loy, C. C. 2024. PGDiff: Guiding diffusion models for versatile face restoration via partial guidance. Advances in Neural Information Processing Systems, 36

    2024

  47. [55]

    Yang, S.; Luo, P.; Loy, C.-C.; and Tang, X. 2016. Wider face: A face detection benchmark. In Proceedings of the IEEE conference on computer vision and pattern recognition, 5525--5533

    2016

  48. [56]

    Zhang, L.; Rao, A.; and Agrawala, M. 2023. Adding conditional control to text-to-image diffusion models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 3836--3847

    2023

  49. [57]

    Zhou, S.; Chan, K.; Li, C.; and Loy, C. C. 2022. Towards robust blind face restoration with codebook lookup transformer. Advances in Neural Information Processing Systems, 35: 30599--30611

    2022

  50. [58]

    W.; and Yuen, P

    Zou, W. W.; and Yuen, P. C. 2011. Very low resolution face recognition problem. IEEE Transactions on image processing, 21(1): 327--340

    2011

  51. [59]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all...

  52. [60]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

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