arxiv: 2604.19238 · v1 · submitted 2026-04-21 · 💻 cs.CV

Recognition: unknown

Allo{SR}²: Rectifying One-Step Super-Resolution to Stay Real via Allomorphic Generative Flows

Jie Hu, Junbo Qiao, Shaohui Lin, Wei Li, Xinghao Chen, Xudong Huang, Zihan Wang

Authors on Pith no claims yet

Pith reviewed 2026-05-10 03:38 UTC · model grok-4.3

classification 💻 cs.CV

keywords real-world super-resolutionone-step generationgenerative flow modelsprior collapsetrajectory consistencyallomorphic flowsimage restorationflow-based models

0 comments

The pith

Allo{SR}^2 rectifies one-step super-resolution trajectories using allomorphic generative flows to avoid prior collapse and trajectory drift.

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

The paper seeks to show that fine-tuning large generative flow models on small sets of low-to-high resolution image pairs causes the models to lose their ability to produce varied realistic details, a problem called prior collapse. This collapse becomes worse when the model must generate the high-resolution output in a single step rather than through many refinement steps, because errors in the generation path cannot be corrected along the way. The authors introduce three specific mechanisms: an initialization step that chooses a starting noise level based on the signal-to-noise ratio of the input, a consistency check that keeps the velocity of the generation path steady, and a matching procedure that forces the super-resolution path to stay close to the original generative path in distribution. If these mechanisms work as intended, one-step super-resolution can deliver both accurate detail recovery from the low-resolution input and the natural appearance that comes from large pre-trained models, all at low computational cost. Readers interested in fast, realistic image enhancement for photography or video would find this relevant because current one-step methods often trade off one quality for the other.

Core claim

The central claim is that one-step real-world super-resolution can be performed by treating the process as an allomorphic generative flow that shares the same underlying vector field as a pre-trained generative flow model. This is achieved by initializing the trajectory at an SNR-aligned timestep so the starting degradation matches the pre-trained model's anchoring point, by applying Flow-Anchored Trajectory Consistency to supervise velocities at intermediate states and keep the path curvature-free, and by using Allomorphic Trajectory Matching to minimize the distributional gap between the super-resolution flow and the generative flow through a self-adversarial alignment in the unified field

What carries the argument

Allomorphic generative flows, a construction that aligns the super-resolution trajectory with a pre-trained generative flow through SNR-guided initialization, velocity-level consistency supervision, and distributional trajectory matching in a shared vector field.

If this is right

One-step inference becomes practical for real-world super-resolution while preserving both pixel-level fidelity and generative detail richness.
The framework achieves leading results on standard synthetic and real-world super-resolution test sets without multi-step refinement.
Computational cost stays low enough for applications that require near real-time processing.
Prior collapse is avoided even when the pre-trained model is adapted only to small paired datasets.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same trajectory-alignment idea could be tested on other single-pass generative tasks such as image denoising or inpainting where drift from pre-trained priors is also common.
If the allomorphic matching proves stable, it may allow domain adaptation of large flow models using far fewer paired examples than currently required.
The approach points toward checking whether similar rectification can maintain consistency when extending one-step generation to video sequences.

Load-bearing premise

The assumption that SNR-guided initialization, flow-anchored velocity supervision, and allomorphic distributional alignment will together prevent prior collapse and trajectory drift on limited LR-HR data without introducing new instabilities or needing heavy retuning.

What would settle it

Running the method on a new collection of real-world low-resolution images and finding that the outputs contain more visible artifacts or lower perceptual realism scores than a multi-step baseline would show the rectification does not hold.

Figures

Figures reproduced from arXiv: 2604.19238 by Jie Hu, Junbo Qiao, Shaohui Lin, Wei Li, Xinghao Chen, Xudong Huang, Zihan Wang.

**Figure 1.** Figure 1: (Left) Existing methods directly substitute Gaussian noise with LR latents, leading to a domain gap and trajectory deviation from the optimal generative path. (Right) Compared to other DM- and FM-based one-step methods like OSEDiff [46] and TSD-SR [11], our approach preserves the generative prior more effectively, recovering realistic textures in severely degraded regions where others suffer from prior col… view at source ↗

**Figure 2.** Figure 2: Overall of Allo{SR}2 . We rectify one-step SR trajectory by anchoring it to an allomorphic generative flow. (b) SNR-Guided Trajectory Initialization identifies the optimal initializing timestep to bridge the domain gap between degraded latents and the generative prior. (b) Flow-Anchored Trajectory Consistency (FATC) enforces a linear, curvature-free SR path via velocity-level supervision. (c) Allomorphic T… view at source ↗

**Figure 3.** Figure 3: Visual comparisons of different DM- and FM-based SR methods. Please zoom in for a better view. Inference Efficiency. The core advantage of Allo{SR}2 is its extreme efficiency. Compared to StableSR (200 steps) and SeeSR (50 steps), our method reduces the number of function evaluations by 50× to 200× while achieving superior or comparable perceptual quality. Even compared to current one-step SOTAs, Allo{SR}… view at source ↗

**Figure 4.** Figure 4: Visual comparisons of different DM- and FM-based SR methods. Please zoom in for a better view [PITH_FULL_IMAGE:figures/full_fig_p020_4.png] view at source ↗

read the original abstract

Real-world image super-resolution (Real-SR) has been revolutionized by leveraging the powerful generative priors of large-scale diffusion and flow-based models. However, fine-tuning these models on limited LR-HR pairs often precipitates "prior collapse" that the model sacrifices its inherent generative richness to overfit specific training degradations. This issue is further exacerbated in one-step generation, where the absence of multi-step refinement leads to significant trajectory drift and artifact generation. In this paper, we propose Allo{SR}$^2$, a novel framework that rectifies one-step SR trajectories via allomorphic generative flows to maintain high-fidelity generative realism. Specifically, we utilize Signal-to-Noise Ratio (SNR) Guided Trajectory Initialization to establish a physically grounded starting state by aligning the degradation level of LR latent features with the optimal anchoring timestep of the pre-trained flow. To ensure a stable, curvature-free path for one-step inference, we propose Flow-Anchored Trajectory Consistency (FATC), which enforces velocity-level supervision across intermediate states. Furthermore, we develop Allomorphic Trajectory Matching (ATM), a self-adversarial alignment strategy that minimizes the distributional discrepancy between the SR flow and the generative flow in a unified vector field. Extensive experiments on both synthetic and real-world benchmarks demonstrate that Allo{SR}$^2$ achieves state-of-the-art performance in one-step Real-SR, offering a superior balance between restoration fidelity and generative realism while maintaining extreme efficiency.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

Allo{SR}^2 gives a coherent set of fixes for collapse and drift in one-step flow-based real SR but the abstract leaves the actual gains unverified.

read the letter

Hi, The main point is that Allo{SR}^2 claims to deliver state-of-the-art one-step real-world super-resolution by rectifying flow trajectories with SNR-guided initialization, flow-anchored trajectory consistency for velocity supervision, and allomorphic self-adversarial trajectory matching to preserve generative realism without collapse or drift. What the paper does well is clearly articulate the failure modes from fine-tuning large pre-trained flows on small LR-HR datasets, leading to loss of inherent generative richness, and then propose targeted fixes for the one-step case where multi-step refinement isn't available. The SNR part grounds the starting latent in a physically meaningful way by matching degradation to timestep. FATC adds supervision to enforce stable paths, and ATM handles distributional alignment in the vector field. This seems like a thoughtful engineering response to practical issues in deploying these models efficiently. The approach builds directly on existing generative priors, which is fine, and the efficiency remains high as it's still one-step. Where it is soft is the absence of detailed derivations, ablation studies, or specific quantitative results in the available text. It's hard to assess whether the three components work as described or if the self-adversarial matching avoids introducing instabilities. The SOTA claim on benchmarks would need the actual numbers and comparisons to hold up under scrutiny. This work is for researchers and practitioners in image restoration and generative modeling who care about balancing quality and speed in super-resolution. A reader focused on flow-based or diffusion models for real-world applications would likely find the framework useful to explore. It deserves a serious referee because the ideas are specific, the motivation is solid, and the claims are falsifiable with experiments. Peer review would help confirm the implementation and results.

Referee Report

0 major / 0 minor

Summary. The manuscript presents Allo{SR}^2, a framework for one-step real-world super-resolution that rectifies trajectories in generative flows to avoid prior collapse and drift. It uses SNR-guided initialization to match LR degradation to flow timesteps, FATC for enforcing consistent velocity in one-step paths, and ATM for self-adversarial matching of distributions in a shared vector field. Experiments on synthetic and real benchmarks are said to show SOTA performance balancing fidelity and realism with high efficiency.

Significance. This work could be significant if validated, as one-step Real-SR is important for practical applications. By building on pre-trained flows and adding targeted supervision without full retraining, it offers a way to maintain generative capabilities while adapting to real degradations. The efficiency of one-step inference is a strong point, and the approach may generalize to other generative tasks.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive assessment of Allo{SR}^2 and the recommendation for minor revision. The feedback highlights the practical importance of one-step Real-SR and the efficiency of our approach, which aligns with our goals.

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The paper's core claims rest on three explicitly proposed mechanisms—SNR-guided trajectory initialization to align LR latents with a pre-trained flow's anchoring timestep, FATC for velocity-level supervision enforcing curvature-free paths, and ATM for self-adversarial distributional matching in a unified vector field—introduced to mitigate prior collapse and trajectory drift in one-step inference. These are defined functionally and independently on top of existing pre-trained models rather than by construction from fitted parameters, self-referential equations, or load-bearing self-citations. No equations reduce the claimed SOTA balance of fidelity and realism to the inputs themselves, and the approach remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only the abstract is available, so the complete ledger cannot be extracted. The framework assumes pre-trained flow models retain usable generative richness after adaptation and introduces new trajectory concepts whose parameters (e.g., anchoring timestep, velocity weights) are not specified.

pith-pipeline@v0.9.0 · 5582 in / 1206 out tokens · 44495 ms · 2026-05-10T03:38:00.906215+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

64 extracted references · 13 canonical work pages · 7 internal anchors

[1]

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

Agustsson, E., Timofte, R.: Ntire 2017 challenge on single image super-resolution: Dataset and study. In: Proceedings of the IEEE conference on computer vision and pattern recognition workshops. pp. 126–135 (2017)

2017
[2]

Ad- vances in Neural Information Processing Systems37, 55443–55469 (2024)

Ai, Y., Zhou, X., Huang, H., Han, X., Chen, Z., You, Q., Yang, H.: Dreamclear: High-capacity real-world image restoration with privacy-safe dataset curation. Ad- vances in Neural Information Processing Systems37, 55443–55469 (2024)

2024
[3]

Stochastic Interpolants: A Unifying Framework for Flows and Diffusions

Albergo,M.S.,Boffi,N.M.,Vanden-Eijnden,E.:Stochasticinterpolants:Aunifying framework for flows and diffusions, 2023. URL https://arxiv. org/abs/2303.08797 3, 1 (2023)

work page internal anchor Pith review arXiv 2023
[4]

Building Normalizing Flows with Stochastic Interpolants

Albergo, M.S., Vanden-Eijnden, E.: Building normalizing flows with stochastic interpolants. arXiv preprint arXiv:2209.15571 (2022)

work page internal anchor Pith review arXiv 2022
[5]

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

Cai, J., Zeng, H., Yong, H., Cao, Z., Zhang, L.: Toward real-world single im- age super-resolution: A new benchmark and a new model. In: Proceedings of the IEEE/CVF international conference on computer vision. pp. 3086–3095 (2019)

2019
[6]

Twinflow: Realizing one-step generation on large models with self-adversarial flows.arXiv preprint arXiv:2512.05150, 2025

Cheng, Z., Sun, P., Li, J., Lin, T.: Twinflow: Realizing one-step generation on large models with self-adversarial flows. arXiv preprint arXiv:2512.05150 (2025)

work page arXiv 2025
[7]

(No Title) (2014)

Diederik, K.: Adam: A method for stochastic optimization. (No Title) (2014)

2014
[8]

IEEE transactions on pattern analysis and machine intelligence44(5), 2567–2581 (2020)

Ding, K., Ma, K., Wang, S., Simoncelli, E.P.: Image quality assessment: Unify- ing structure and texture similarity. IEEE transactions on pattern analysis and machine intelligence44(5), 2567–2581 (2020)

2020
[9]

In: European conference on computer vision

Dong, C., Loy, C.C., He, K., Tang, X.: Learning a deep convolutional network for image super-resolution. In: European conference on computer vision. pp. 184–199. Springer (2014)

2014
[10]

IEEE transactions on pattern analysis and machine intelligence 38(2), 295–307 (2015)

Dong, C., Loy, C.C., He, K., Tang, X.: Image super-resolution using deep convo- lutional networks. IEEE transactions on pattern analysis and machine intelligence 38(2), 295–307 (2015)

2015
[11]

In: Proceedings of the Computer Vision and Pattern Recognition Con- ference

Dong, L., Fan, Q., Guo, Y., Wang, Z., Zhang, Q., Chen, J., Luo, Y., Zou, C.: Tsd-sr: One-step diffusion with target score distillation for real-world image super- resolution. In: Proceedings of the Computer Vision and Pattern Recognition Con- ference. pp. 23174–23184 (2025)

2025
[12]

Advances in neural in- formation processing systems27(2014)

Goodfellow, I.J., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., Bengio, Y.: Generative adversarial nets. Advances in neural in- formation processing systems27(2014)

2014
[13]

Advances in neural information processing systems30(2017)

Heusel,M.,Ramsauer,H.,Unterthiner,T.,Nessler,B.,Hochreiter,S.:Ganstrained by a two time-scale update rule converge to a local nash equilibrium. Advances in neural information processing systems30(2017)

2017
[14]

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)

2020
[15]

Iclr1(2), 3 (2022)

Hu, E.J., Shen, Y., Wallis, P., Allen-Zhu, Z., Li, Y., Wang, S., Wang, L., Chen, W., et al.: Lora: Low-rank adaptation of large language models. Iclr1(2), 3 (2022)

2022
[16]

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

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

2019
[17]

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

Ke, J., Wang, Q., Wang, Y., Milanfar, P., Yang, F.: Musiq: Multi-scale image quality transformer. In: Proceedings of the IEEE/CVF international conference on computer vision. pp. 5148–5157 (2021)

2021
[18]

Consistency Trajectory Models: Learning Probability Flow ODE Trajectory of Diffusion, March 2024

Kim, D., Lai, C.H., Liao, W.H., Murata, N., Takida, Y., Uesaka, T., He, Y., Mitsu- fuji, Y., Ermon, S.: Consistency trajectory models: Learning probability flow ode trajectory of diffusion. arXiv preprint arXiv:2310.02279 (2023) 16 Z. Wang et al

work page arXiv 2023
[19]

Labs, B.F.: Flux.https://github.com/black- forest- labs/flux(2024), ac- cessed: 2024-11-14

2024
[20]

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

Ledig, C., Theis, L., Huszár, F., Caballero, J., Cunningham, A., Acosta, A., Aitken, A., Tejani, A., Totz, J., Wang, Z., et al.: Photo-realistic single image super- resolution using a generative adversarial network. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 4681–4690 (2017)

2017
[21]

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

Li, Y., Zhang, K., Liang, J., Cao, J., Liu, C., Gong, R., Zhang, Y., Tang, H., Liu, Y., Demandolx, D., et al.: Lsdir: A large scale dataset for image restoration. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 1775–1787 (2023)

2023
[22]

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

Liang, J., Zeng, H., Zhang, L.: Details or artifacts: A locally discriminative learning approach to realistic image super-resolution. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 5657–5666 (2022)

2022
[23]

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

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

2021
[24]

In: European conference on computer vision

Lin, X., He, J., Chen, Z., Lyu, Z., Dai, B., Yu, F., Qiao, Y., Ouyang, W., Dong, C.: Diffbir: Toward blind image restoration with generative diffusion prior. In: European conference on computer vision. pp. 430–448. Springer (2024)

2024
[25]

Flow Matching for Generative Modeling

Lipman, Y., Chen, R.T., Ben-Hamu, H., Nickel, M., Le, M.: Flow matching for generative modeling. arXiv preprint arXiv:2210.02747 (2022)

work page internal anchor Pith review Pith/arXiv arXiv 2022
[26]

Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow

Liu, X., Gong, C., Liu, Q.: Flow straight and fast: Learning to generate and transfer data with rectified flow. arXiv preprint arXiv:2209.03003 (2022)

work page internal anchor Pith review arXiv 2022
[27]

In: The Twelfth Interna- tional Conference on Learning Representations (2023)

Liu, X., Zhang, X., Ma, J., Peng, J., et al.: Instaflow: One step is enough for high-quality diffusion-based text-to-image generation. In: The Twelfth Interna- tional Conference on Learning Representations (2023)

2023
[28]

Latent Consistency Models: Synthesizing High-Resolution Images with Few-Step Inference

Luo, S., Tan, Y., Huang, L., Li, J., Zhao, H.: Latent consistency mod- els: Synthesizing high-resolution images with few-step inference. arXiv preprint arXiv:2310.04378 (2023)

work page internal anchor Pith review arXiv 2023
[29]

In: Proceedings of the IEEE/CVF confer- ence on computer vision and pattern recognition

Meng, C., Rombach, R., Gao, R., Kingma, D., Ermon, S., Ho, J., Salimans, T.: On distillation of guided diffusion models. In: Proceedings of the IEEE/CVF confer- ence on computer vision and pattern recognition. pp. 14297–14306 (2023)

2023
[30]

In: International conference on machine learning

Nichol, A.Q., Dhariwal, P.: Improved denoising diffusion probabilistic models. In: International conference on machine learning. pp. 8162–8171. PMLR (2021)

2021
[31]

Peebles,W.,Xie,S.:Scalablediffusionmodelswithtransformers.In:Proceedingsof the IEEE/CVF international conference on computer vision. pp. 4195–4205 (2023)

2023
[32]

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)

2022
[33]

Progressive Distillation for Fast Sampling of Diffusion Models

Salimans, T., Ho, J.: Progressive distillation for fast sampling of diffusion models. arXiv preprint arXiv:2202.00512 (2022)

work page internal anchor Pith review arXiv 2022
[34]

In: Proceed- ings of the 40th International Conference on Machine Learning

Song, Y., Dhariwal, P., Chen, M., Sutskever, I.: Consistency models. In: Proceed- ings of the 40th International Conference on Machine Learning. pp. 32211–32252 (2023)

2023
[35]

Advances in neural information processing systems32(2019)

Song, Y., Ermon, S.: Generative modeling by estimating gradients of the data distribution. Advances in neural information processing systems32(2019)

2019
[36]

Score-Based Generative Modeling through Stochastic Differential Equations

Song, Y., Sohl-Dickstein, J., Kingma, D.P., Kumar, A., Ermon, S., Poole, B.: Score- based generative modeling through stochastic differential equations. arXiv preprint arXiv:2011.13456 (2020) Allo{SR}2 17

work page internal anchor Pith review Pith/arXiv arXiv 2011
[37]

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

Sun,L.,Wu,R.,Ma,Z.,Liu,S.,Yi,Q.,Zhang,L.:Pixel-levelandsemantic-levelad- justable super-resolution: A dual-lora approach. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 2333–2343 (2025)

2025
[38]

In: Proceedings of the AAAI conference on artificial intelligence

Wang, J., Chan, K.C., Loy, C.C.: Exploring clip for assessing the look and feel of images. In: Proceedings of the AAAI conference on artificial intelligence. vol. 37, pp. 2555–2563 (2023)

2023
[39]

International Journal of Computer Vision 132(12), 5929–5949 (2024)

Wang, J., Yue, Z., Zhou, S., Chan, K.C., Loy, C.C.: Exploiting diffusion prior for real-world image super-resolution. International Journal of Computer Vision 132(12), 5929–5949 (2024)

2024
[40]

In: Proceedings of the IEEE/CVF in- ternational conference on computer vision

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

1905
[41]

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

Wang, X., Yu, K., Dong, C., Tang, X., Loy, C.C.: Deep network interpolation for continuous imagery effect transition. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 1692–1701 (2019)

2019
[42]

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

Wang, Y., Yang, W., Chen, X., Wang, Y., Guo, L., Chau, L.P., Liu, Z., Qiao, Y., Kot, A.C., Wen, B.: Sinsr: diffusion-based image super-resolution in a single step. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 25796–25805 (2024)

2024
[43]

Advances in neural information processing systems36, 8406–8441 (2023)

Wang, Z., Lu, C., Wang, Y., Bao, F., Li, C., Su, H., Zhu, J.: Prolificdreamer: High-fidelity and diverse text-to-3d generation with variational score distillation. Advances in neural information processing systems36, 8406–8441 (2023)

2023
[44]

IEEE transactions on image processing 13(4), 600–612 (2004)

Wang, Z., Bovik, A.C., Sheikh, H.R., Simoncelli, E.P.: Image quality assessment: from error visibility to structural similarity. IEEE transactions on image processing 13(4), 600–612 (2004)

2004
[45]

In: European conference on computer vision

Wei, P., Xie, Z., Lu, H., Zhan, Z., Ye, Q., Zuo, W., Lin, L.: Component divide- and-conquer for real-world image super-resolution. In: European conference on computer vision. pp. 101–117. Springer (2020)

2020
[46]

Advances in Neural Information Processing Systems 37, 92529–92553 (2024)

Wu, R., Sun, L., Ma, Z., Zhang, L.: One-step effective diffusion network for real- world image super-resolution. Advances in Neural Information Processing Systems 37, 92529–92553 (2024)

2024
[47]

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

Wu, R., Yang, T., Sun, L., Zhang, Z., Li, S., Zhang, L.: Seesr: Towards semantics- aware real-world image super-resolution. In: Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition. pp. 25456–25467 (2024)

2024
[48]

arXiv preprint arXiv:2508.08227 (2025)

Wu, Z., Sun, Z., Zhou, T., Fu, B., Cong, J., Dong, Y., Zhang, H., Tang, X., Chen, M., Wei, X.: Omgsr: You only need one mid-timestep guidance for real-world image super-resolution. arXiv preprint arXiv:2508.08227 (2025)

work page arXiv 2025
[49]

arXiv preprint arXiv:2307.02457 (2023)

Xie, L., Wang, X., Chen, X., Li, G., Shan, Y., Zhou, J., Dong, C.: Desra: detect and delete the artifacts of gan-based real-world super-resolution models. arXiv preprint arXiv:2307.02457 (2023)

work page arXiv 2023
[50]

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

Xu, J., Li, W., Sun, H., Li, F., Wang, Z., Peng, L., Ren, J., Yang, H., Hu, X., Pei, R., et al.: Fast image super-resolution via consistency rectified flow. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 11755–11765 (2025)

2025
[51]

arXiv preprint arXiv:2407.02398 , year=

Yang, L., Zhang, Z., Zhang, Z., Liu, X., Xu, M., Zhang, W., Meng, C., Ermon, S., Cui, B.: Consistency flow matching: Defining straight flows with velocity consis- tency. arXiv preprint arXiv:2407.02398 (2024)

work page arXiv 2024
[52]

In: Proceedings of the IEEE/CVF conference on computer vision and pattern recog- nition

Yang, S., Wu, T., Shi, S., Lao, S., Gong, Y., Cao, M., Wang, J., Yang, Y.: Maniqa: Multi-dimension attention network for no-reference image quality assessment. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recog- nition. pp. 1191–1200 (2022) 18 Z. Wang et al

2022
[53]

In: European conference on computer vision

Yang, T., Wu, R., Ren, P., Xie, X., Zhang, L.: Pixel-aware stable diffusion for real- istic image super-resolution and personalized stylization. In: European conference on computer vision. pp. 74–91. Springer (2024)

2024
[54]

Advances in neural information processing systems37, 47455–47487 (2024)

Yin, T., Gharbi, M., Park, T., Zhang, R., Shechtman, E., Durand, F., Freeman, B.: Improved distribution matching distillation for fast image synthesis. Advances in neural information processing systems37, 47455–47487 (2024)

2024
[55]

One-step diffusion with distribution matching distillation.arXiv preprint arXiv:2311.18828, 2023

Yin, T., Gharbi, M., Zhang, R., Shechtman, E., Durand, F., Freeman, W.T., Park, T.: One-step diffusion with distribution matching distillation, 2024. URL https://arxiv. org/abs/2311.18828

work page arXiv 2024
[56]

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

You, W., Zhang, M., Zhang, L., Zhou, X., Shi, K., Gu, S.: Consistency trajec- tory matching for one-step generative super-resolution. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 12747–12756 (2025)

2025
[57]

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

Yu,F.,Gu,J.,Li,Z.,Hu,J.,Kong,X.,Wang,X.,He,J.,Qiao,Y.,Dong,C.:Scaling up to excellence: Practicing model scaling for photo-realistic image restoration in the wild. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 25669–25680 (2024)

2024
[58]

Advances in neural information processing systems 36, 13294–13307 (2023)

Yue, Z., Wang, J., Loy, C.C.: Resshift: Efficient diffusion model for image super- resolution by residual shifting. Advances in neural information processing systems 36, 13294–13307 (2023)

2023
[59]

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

Zhang, K., Liang, J., Van Gool, L., Timofte, R.: Designing a practical degradation model for deep blind image super-resolution. In: Proceedings of the IEEE/CVF international conference on computer vision. pp. 4791–4800 (2021)

2021
[60]

In: Proceedings of the Computer Vision and Pat- tern Recognition Conference

Zhang, L., You, W., Shi, K., Gu, S.: Uncertainty-guided perturbation for image super-resolution diffusion model. In: Proceedings of the Computer Vision and Pat- tern Recognition Conference. pp. 17980–17989 (2025)

2025
[61]

IEEE Transactions on Image Processing24(8), 2579–2591 (2015)

Zhang, L., Zhang, L., Bovik, A.C.: A feature-enriched completely blind image qual- ity evaluator. IEEE Transactions on Image Processing24(8), 2579–2591 (2015)

2015
[62]

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

Zhang, R., Isola, P., Efros, A.A., Shechtman, E., Wang, O.: The unreasonable effectiveness of deep features as a perceptual metric. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 586–595 (2018)

2018
[63]

In: European Conference on Computer Vision

Zhang, Y., Ji, B., Hao, J., Yao, A.: Perception-distortion balanced admm opti- mization for single-image super-resolution. In: European Conference on Computer Vision. pp. 108–125. Springer (2022)

2022
[64]

In: Proceedings of the IEEE interna- tional conference on computer vision

Zhu, J.Y., Park, T., Isola, P., Efros, A.A.: Unpaired image-to-image translation using cycle-consistent adversarial networks. In: Proceedings of the IEEE interna- tional conference on computer vision. pp. 2223–2232 (2017) Allo{SR}2 19 A Algorithm of Allo{SR} 2 The pseudo-code of Allo{SR}2 training Scheme is summarized as Algorithm 1. Algorithm 1Training S...

2017