Pith. sign in

REVIEW 2 major objections 1 minor 63 references

A fully fiber-integrated random laser produces tunable photon superbunching with second-order coherence from ~1 to ~26.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · grok-4.5

2026-07-13 20:29 UTC pith:DZQCBA3V

load-bearing objection Body is the wrong paper (image restoration); the SRFL g(2)~26 claims are unauditable from what we were given. the 2 major comments →

arxiv 2603.22025 v2 pith:DZQCBA3V submitted 2026-03-23 physics.optics

Superbunched random fiber laser

Qiang Ji , Yuxi Pang , Weiming Zhang , Xian Zhao , Zengguang Qin , Zhaojun Liu , Yanping Xu This is my paper
classification physics.optics
keywords photon superbunchingrandom fiber lasersecond-order coherencestimulated Brillouin scatteringfour-wave mixingtemporal ghost imagingParisi overlapphoton statistics
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper shows that a random fiber laser can be engineered so that ordinary Rayleigh scattering, cascaded stimulated Brillouin scattering and four-wave mixing work together to generate extreme photon bunching. Individual lines of the resulting multi-wavelength comb have a second-order coherence g(2)(0) that can be dialed continuously from the coherent value of 1 up to about 26 simply by changing pump power, which spectral line is selected, and the diffusion length of the fiber. The same system also exhibits photonic phase transitions that are tracked by a Parisi overlap order parameter; the appearance of superbunching is shown to track those transitions. When the superbunched light is used for temporal ghost imaging, high-fidelity object reconstruction is obtained with far fewer averages than thermal light requires. The result is offered as a robust, all-fiber platform for both quantum-optics experiments and correlation-based sensing.

Core claim

A fully fiber-integrated superbunched random fiber laser generates a multi-wavelength comb in which individual spectral components exhibit widely tunable photon bunching, with g(2)(0) continuously controlled from ~1 to ~26 through the cooperative action of intrinsic Rayleigh scattering, cascaded stimulated Brillouin scattering and quasi-phase-matched four-wave mixing.

What carries the argument

The cooperative Rayleigh–SBS–FWM interaction inside the random fiber laser, which shapes the photon statistics of each comb line; the Parisi overlap order parameter that quantifies the linked photonic phase transition.

Load-bearing premise

That the reported g(2)(0) values up to ~26 are produced by the intended cooperative nonlinear dynamics rather than by detector artifacts, residual pump light, spectral filtering effects or classical intensity fluctuations that can inflate measured intensity correlations.

What would settle it

A repeated intensity-correlation measurement on the same spectral lines performed with an independent beam-splitter and detector pair (and with spectral filtering that fully rejects residual pump and neighboring lines) that fails to recover g(2)(0) values above the thermal limit of 2 under identical pump power and fiber length.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Extreme photon statistics become available from a compact, alignment-free fiber platform instead of fragile free-space nonlinear optics.
  • Temporal ghost imaging can reach high reconstruction fidelity with substantially fewer ensemble averages.
  • Macroscopic disorder diagnostics (Parisi overlap) can be used as a practical control knob for microscopic photon statistics.
  • Correlation-enhanced sensing and quantum-optics experiments can be performed inside complex photonic systems without specialized free-space hardware.

Where Pith is reading between the lines

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

  • The same Rayleigh–SBS–FWM cascade could be realized in other disordered waveguides or photonic-crystal fibers to push g(2)(0) still higher or to generate multi-photon correlations.
  • If the Parisi-overlap–superbunching link is general, spin-glass-inspired order parameters may become standard diagnostics for non-classical light sources in disordered media.
  • A calibrated superbunched fiber source would allow direct benchmarking of the sample-efficiency claims of intensity-correlation imaging protocols.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 1 minor

Summary. The titled work claims a fully fiber-integrated superbunched random fiber laser (SRFL) in which Rayleigh scattering cooperates with cascaded stimulated Brillouin scattering and quasi-phase-matched four-wave mixing to produce a multi-wavelength comb. Individual comb lines are said to exhibit continuously tunable second-order coherence g^{(2)}(0) from ~1 to ~26 by varying pump power, spectral order and diffusion length; a correlation is asserted between photonic phase transitions (via a Parisi overlap order parameter) and the onset of superbunching; and the source is applied to temporal ghost imaging with reduced ensemble averaging. The supplied full-text body, however, is an unrelated computer-vision manuscript (ResFlow-Tuner) on test-time scaling of flow-matching models for real-world image restoration and contains none of the laser physics, apparatus, correlation measurements or imaging results described in the abstract.

Significance. If the SRFL claims were substantiated by a matching manuscript, a robust, all-fiber platform delivering g^{(2)}(0) values far above the thermal limit, continuous tunability, a disorder-physics link via Parisi overlap, and a concrete ghost-imaging demonstration would constitute a valuable resource for quantum optics and correlation imaging. The present submission supplies no data, methods or figures that would allow those claims to be assessed, so the significance of the actual scientific result cannot be evaluated.

major comments (2)
  1. The full manuscript text that accompanies the title and abstract is a completely different paper (ResFlow-Tuner: test-time scaling of FLUX.1-dev flow-matching models for real-world image restoration, with sections on UMMF, ODE-adapted TTS, Tables 1–4, Figs. 1–10, etc.). No description of the fiber laser cavity, Rayleigh–SBS–FWM interaction, Hanbury Brown–Twiss apparatus, spectral filtering, g^{(2)}( au) traces, Parisi-overlap calculation, or temporal ghost-imaging protocol appears anywhere in the body. Consequently every load-bearing experimental assertion in the abstract—tunable g^{(2)}(0) ~1–26, spectral-order/diffusion-length control, Parisi correlation, and imaging gain—is unauditable. The manuscript as submitted cannot support the central claims.
  2. Because the body is mismatched, standard experimental controls that would address the weakest assumption (detector artifacts, residual pump, filter bandwidth, classical intensity fluctuations inflating intensity correlations) are absent. No error bars, background-subtraction protocol, or comparison with a thermal reference source can be inspected. This is not a minor omission; it renders the superbunching claim scientifically uncheckable from the provided document.
minor comments (1)
  1. The abstract alone is well written and clearly states the intended contributions, but without a matching body it cannot be reviewed further.

Circularity Check

0 steps flagged

No circularity: empirical method paper with independent benchmarks and no derivation that reduces claims to fitted inputs or self-definition.

full rationale

The provided full manuscript is a computer-vision engineering paper (ResFlow-Tuner) proposing a flow-matching IR framework with UMMF conditioning and an ODE-adapted TTS search guided by a rank-based verifier ensemble. All performance claims are evaluated on external synthetic/real benchmarks (DIV2K-Val, LSDIR-Val, RealPhoto60, RealSR, DRealSR, OCR) using standard metrics (LPIPS, FID, MANIQA, MUSIQ, CLIPIQA, etc.) and ablations that vary components independently. There is no first-principles derivation chain, no parameter fitted to a subset then re-presented as a prediction of a related quantity, no uniqueness theorem imported from the authors, and no self-citation that is load-bearing for the central result. The TTS procedure (perturbation of intermediate ODE states, MSPDE rollout, ensemble ranking) is a design choice justified by SDE theory citations external to the authors; it does not define the reported quality gains by construction. The abstract/title mismatch with the laser paper is noted but does not create circularity inside the supplied text. Score 0 is therefore required.

Axiom & Free-Parameter Ledger

3 free parameters · 3 axioms · 1 invented entities

Abstract-only review of an experimental optics claim. Load-bearing content is experimental rather than axiomatic derivation. Free parameters are the experimental knobs used to tune g(2). Domain assumptions are standard nonlinear fiber optics (Rayleigh scattering, SBS, FWM). No new fundamental particles or forces; the SRFL is an engineered system. The Parisi overlap is imported from spin-glass/disorder physics as an order parameter—its applicability to this photonic system is a modeling choice that must be justified in the full paper.

free parameters (3)
  • pump power
    Primary experimental control stated to tune g(2)(0) continuously; specific values and calibration not given in abstract.
  • spectral order (comb line index)
    Which multi-wavelength comb component is measured is said to control bunching; selection criteria not specified.
  • diffusion length
    Stated tuning parameter for photon statistics; physical definition and range not given in abstract.
axioms (3)
  • domain assumption Intrinsic Rayleigh scattering, cascaded stimulated Brillouin scattering, and quasi-phase-matched four-wave mixing cooperatively determine the photon statistics of the SRFL.
    Central mechanistic claim of the abstract; treated as the physical origin of superbunching without derivation available here.
  • domain assumption Second-order coherence g(2)(0) is a faithful measure of photon bunching for the reported multi-wavelength emission under the experimental detection conditions.
    Standard quantum-optics assumption; validity depends on spectral filtering, detector response, and classical vs quantum intensity fluctuations not detailed in the abstract.
  • ad hoc to paper A Parisi overlap order parameter meaningfully quantifies photonic phase transitions in this random fiber laser and correlates with the emergence of superbunching.
    Import of a spin-glass concept into this photonic system; abstract asserts a direct correlation but does not supply the mapping or independent validation.
invented entities (1)
  • Superbunched random fiber laser (SRFL) as a cooperative Rayleigh–SBS–FWM platform no independent evidence
    purpose: Provide a fully fiber-integrated source of extreme, tunable photon bunching and a testbed linking disorder physics to photon statistics.
    Named system that is the paper’s experimental object. Independent evidence would be external replication of g(2)(0) and the Parisi–superbunching correlation; not available from the abstract alone.

pith-pipeline@v1.1.0-grok45 · 24117 in / 2861 out tokens · 33906 ms · 2026-07-13T20:29:28.787900+00:00 · methodology

0 comments
read the original abstract

Photon superbunching, distinguished by second-order coherence values far exceeding the Gaussian thermal limit, represents a highly desirable resource for quantum optics and correlation-based imaging technologies. However, existing approaches typically rely on fragile experimental platforms, inefficient nonlinear conversion processes, or mechanically complex optical architectures. Here, we demonstrate a fully fiber-integrated superbunched random fiber laser (SRFL) in which intrinsic Rayleigh scattering cooperatively interacts with cascaded stimulated Brillouin scattering and quasi-phase-matched four-wave mixing to tailor extreme photon statistics. The SRFL generates a multi-wavelength comb, in which individual spectral components exhibit widely tunable photon bunching, with the second-order coherence g(2)(0) continuously controlled from ~1 to ~26 by tuning the pump power, spectral order and diffusion length. Moreover, we establish a direct correlation between photonic phase transitions (quantified by the Parisi overlap order parameter) and the emergence of superbunching, thereby bridging macroscopic disorder physics and microscopic photon statistics. Finally, we employ the superbunched emission for temporal ghost imaging, realizing high-fidelity temporal object reconstruction with a substantial reduction in required ensemble averaging. These findings validate random fiber lasers as a robust, scalable, and integrated platform for generating extreme photon statistics and unlock new avenues for correlation-enhanced photonic sensing and quantum optics investigations in complex photonic systems.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

63 extracted references · 18 linked inside Pith

  1. [1]

    In: CVPRW

    Agustsson, E., Timofte, R.: Ntire 2017 challenge on single image super-resolution: Dataset and study. In: CVPRW. pp. 126–135 (2017)

    2017

  2. [2]

    NIPS37, 55443–55469 (2024)

    Ai, Y., Zhou, X., Huang, H., Han, X., Chen, Z., You, Q., Yang, H.: Dream- clear: High-capacity real-world image restoration with privacy-safe dataset cura- tion. NIPS37, 55443–55469 (2024)

    2024

  3. [3]

    arXiv preprint arXiv:2303.08797 (2023)

    Albergo,M.S.,Boffi,N.M.,Vanden-Eijnden,E.:Stochasticinterpolants:Aunifying framework for flows and diffusions. arXiv preprint arXiv:2303.08797 (2023)

    Pith/arXiv arXiv 2023

  4. [4]

    arXiv preprint arXiv:2502.13923 (2025)

    Bai, S., Chen, K., Liu, X., Wang, J., Ge, W., Song, S., Dang, K., Wang, P., Wang, S., Tang, J., Zhong, H., Zhu, Y., Yang, M., Li, Z., Wan, J., Wang, P., Ding, W., Fu, Z., Xu, Y., Ye, J., Zhang, X., Xie, T., Cheng, Z., Zhang, H., Yang, Z., Xu, H., Lin, J.: Qwen2.5-vl technical report. arXiv preprint arXiv:2502.13923 (2025)

    Pith/arXiv arXiv 2025

  5. [5]

    In: ICCV

    Cai, J., Zeng, H., Yong, H., Cao, Z., Zhang, L.: Toward real-world single image super-resolution: A new benchmark and a new model. In: ICCV. pp. 3086–3095 (2019)

    2019

  6. [6]

    arXiv preprint arXiv:2310.00426 (2023)

    Chen, J., Yu, J., Ge, C., Yao, L., Xie, E., Wu, Y., Wang, Z., Kwok, J., Luo, P., Lu, H., et al.: Pixart-�: Fast training of diffusion transformer for photorealistic text-to-image synthesis. arXiv preprint arXiv:2310.00426 (2023)

    Pith/arXiv arXiv 2023

  7. [7]

    In: CVPR

    Chen, J., Pan, J., Dong, J.: Faithdiff: Unleashing diffusion priors for faithful image super-resolution. In: CVPR. pp. 28188–28197 (2025)

    2025

  8. [8]

    In: CVPR

    Chen, Z., Wu, Z., Zamfir, E., Zhang, K., Zhang, Y., Timofte, R., Yang, X., Yu, H., Wan, C., Hong, Y., et al.: Ntire 2024 challenge on image super-resolution (x4): Methods and results. In: CVPR. pp. 6108–6132 (2024)

    2024

  9. [9]

    TPAMI44(5), 2567–2581 (2020)

    Ding, K., Ma, K., Wang, S., Simoncelli, E.P.: Image quality assessment: Unifying structure and texture similarity. TPAMI44(5), 2567–2581 (2020)

    2020

  10. [10]

    In: ICML (2024)

    Esser, P., Kulal, S., Blattmann, A., Entezari, R., Müller, J., Saini, H., Levi, Y., Lorenz, D., Sauer, A., Boesel, F., et al.: Scaling rectified flow transformers for high-resolution image synthesis. In: ICML (2024)

    2024

  11. [11]

    In: ICCVW

    Gu, S., Lugmayr, A., Danelljan, M., Fritsche, M., Lamour, J., Timofte, R.: Div8k: Diverse 8k resolution image dataset. In: ICCVW. pp. 3512–3516. IEEE (2019)

    2019

  12. [12]

    arXiv preprint arXiv:2505.17618 (2025)

    He, H., Liang, J., Wang, X., Wan, P., Zhang, D., Gai, K., Pan, L.: Scaling image and video generation via test-time evolutionary search. arXiv preprint arXiv:2505.17618 (2025)

    Pith/arXiv arXiv 2025

  13. [13]

    In: EMNLP

    Hessel, J., Holtzman, A., Forbes, M., Le Bras, R., Choi, Y.: Clipscore: A reference- free evaluation metric for image captioning. In: EMNLP. pp. 7514–7528 (2021)

    2021

  14. [14]

    NIPS30 (2017) Title Suppressed Due to Excessive Length 19

    Heusel,M.,Ramsauer,H.,Unterthiner,T.,Nessler,B.,Hochreiter,S.:Ganstrained by a two time-scale update rule converge to a local nash equilibrium. NIPS30 (2017) Title Suppressed Due to Excessive Length 19

    2017

  15. [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. [16]

    In: ICCV

    Ke, J., Wang, Q., Wang, Y., Milanfar, P., Yang, F.: Musiq: Multi-scale image quality transformer. In: ICCV. pp. 5148–5157 (2021)

    2021

  17. [17]

    arXiv preprint arXiv:2503.19385 (2025)

    Kim, J., Yoon, T., Hwang, J., Sung, M.: Inference-time scaling for flow models via stochastic generation and rollover budget forcing. arXiv preprint arXiv:2503.19385 (2025)

    arXiv 2025

  18. [18]

    arXiv preprint arXiv:2501.05803 (2025)

    Kim,S.,Kim,M.,Park,D.:Test-timealignmentofdiffusionmodelswithoutreward over-optimization. arXiv preprint arXiv:2501.05803 (2025)

    Pith/arXiv arXiv 2025

  19. [19]

    Labs, B.F.: Flux.�����������������������������������������(2024)

    2024

  20. [20]

    arXiv preprint arXiv:2206.03001 (2022)

    Li, C., Liu, W., Guo, R., Yin, X., Jiang, K., Du, Y., Du, Y., Zhu, L., Lai, B., Hu, X., et al.: Pp-ocrv3: More attempts for the improvement of ultra lightweight ocr system. arXiv preprint arXiv:2206.03001 (2022)

    Pith/arXiv arXiv 2022

  21. [21]

    arXiv preprint arXiv:2503.02039 (2025)

    Li, X., Uehara, M., Su, X., Scalia, G., Biancalani, T., Regev, A., Levine, S., Ji, S.: Dynamic search for inference-time alignment in diffusion models. arXiv preprint arXiv:2503.02039 (2025)

    Pith/arXiv arXiv 2025

  22. [22]

    arXiv preprint arXiv:2408.08252 (2024)

    Li, X., Zhao, Y., Wang, C., Scalia, G., Eraslan, G., Nair, S., Biancalani, T., Ji, S., Regev, A., Levine, S., et al.: Derivative-free guidance in continuous and discrete diffusion models with soft value-based decoding. arXiv preprint arXiv:2408.08252 (2024)

    Pith/arXiv arXiv 2024

  23. [23]

    In: CVPRW

    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: CVPRW. pp. 1775–1787 (2023)

    2023

  24. [24]

    In: ECCV

    Liang, J., Zeng, H., Zhang, L.: Efficient and degradation-adaptive network for real- world image super-resolution. In: ECCV. pp. 574–591. Springer (2022)

    2022

  25. [25]

    In: CVPRW

    Lim, B., Son, S., Kim, H., Nah, S., Mu Lee, K.: Enhanced deep residual networks for single image super-resolution. In: CVPRW. pp. 136–144 (2017)

    2017

  26. [26]

    In: ECCV

    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: ECCV. pp. 430–448. Springer (2024)

    2024

  27. [27]

    arXiv preprint arXiv:2210.02747 (2022)

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

    Pith/arXiv arXiv 2022

  28. [28]

    arXiv preprint arXiv:2209.03003 (2022)

    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)

    Pith/arXiv arXiv 2022

  29. [29]

    In: ECCV

    Ma, N., Goldstein, M., Albergo, M.S., Boffi, N.M., Vanden-Eijnden, E., Xie, S.: Sit: Exploring flow and diffusion-based generative models with scalable interpolant transformers. In: ECCV. pp. 23–40. Springer (2024)

    2024

  30. [30]

    In: ICLR (2025)

    Martin, S., Gagneux, A., Hagemann, P., Steidl, G.: Pnp-flow: Plug-and-play image restoration with flow matching. In: ICLR (2025)

    2025

  31. [31]

    arXiv preprint arXiv:2306.06101 (2023)

    Mishchenko, K., Defazio, A.: Prodigy: An expeditiously adaptive parameter-free learner. arXiv preprint arXiv:2306.06101 (2023)

    Pith/arXiv arXiv 2023

  32. [32]

    arXiv preprint arXiv:2501.19252 (2025)

    Oshima, Y., Suzuki, M., Matsuo, Y., Furuta, H.: Inference-time text-to-video align- ment with diffusion latent beam search. arXiv preprint arXiv:2501.19252 (2025)

    arXiv 2025

  33. [33]

    arXiv preprint arXiv:2412.11024 (2024)

    Patel, Z., DeLoye, J., Mathias, L.: Exploring diffusion and flow matching under generator matching. arXiv preprint arXiv:2412.11024 (2024)

    Pith/arXiv arXiv 2024

  34. [34]

    von Platen, P., Patil, S., Lozhkov, A., Cuenca, P., Lambert, N., Rasul, K., Davaadorj, M., Nair, D., Paul, S., Berman, W., Xu, Y., Liu, S., Wolf, T.: Diffusers: State-of-the-art diffusion models.���������������������������������������� (2022) 20 Authors Suppressed Due to Excessive Length

    2022

  35. [35]

    arXiv preprint arXiv:2307.01952 (2023)

    Podell, D., English, Z., Lacey, K., Blattmann, A., Dockhorn, T., Müller, J., Penna, J., Rombach, R.: Sdxl: Improving latent diffusion models for high-resolution image synthesis. arXiv preprint arXiv:2307.01952 (2023)

    Pith/arXiv arXiv 2023

  36. [36]

    In: CVPR

    Ren, B., Guo, H., Sun, L., Wu, Z., Timofte, R., Li, Y.: The tenth ntire 2025 efficient super-resolution challenge report. In: CVPR. pp. 917–966 (2025)

    2025

  37. [37]

    NIPS35, 25278– 25294 (2022)

    Schuhmann, C., Beaumont, R., Vencu, R., Gordon, C., Wightman, R., Cherti, M., Coombes, T., Katta, A., Mullis, C., Wortsman, M., et al.: Laion-5b: An open large- scale dataset for training next generation image-text models. NIPS35, 25278– 25294 (2022)

    2022

  38. [38]

    arXiv preprint arXiv:2502.00968 (2025)

    Singh, A., Mukherjee, S., Beirami, A., Jamali-Rad, H.: Code: Blockwise control for denoising diffusion models. arXiv preprint arXiv:2502.00968 (2025)

    Pith/arXiv arXiv 2025

  39. [39]

    arXiv preprint arXiv:2410.02217 (2024)

    Singh, S., Fischer, I.: Stochastic sampling from deterministic flow models. arXiv preprint arXiv:2410.02217 (2024)

    Pith/arXiv arXiv 2024

  40. [40]

    arXiv preprint arXiv:2011.13456 (2020)

    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)

    Pith/arXiv arXiv 2011

  41. [41]

    NIPS 33, 3008–3021 (2020)

    Stiennon, N., Ouyang, L., Wu, J., Ziegler, D., Lowe, R., Voss, C., Radford, A., Amodei, D., Christiano, P.F.: Learning to summarize with human feedback. NIPS 33, 3008–3021 (2020)

    2020

  42. [42]

    In: CVPR

    Sun, L., Guo, H., Ren, B., Van Gool, L., Timofte, R., Li, Y.: The tenth ntire 2025 image denoising challenge report. In: CVPR. pp. 1342–1369 (2025)

    2025

  43. [43]

    In: ICCV

    Tan, Z., Liu, S., Yang, X., Xue, Q., Wang, X.: Ominicontrol: Minimal and universal control for diffusion transformer. In: ICCV. pp. 14940–14950 (2025)

    2025

  44. [44]

    ACM TOG43(4), 1–12 (2024)

    Tang, L., Ruiz, N., Chu, Q., Li, Y., Holynski, A., Jacobs, D.E., Hariharan, B., Pritch, Y., Wadhwa, N., Aberman, K., et al.: Realfill: Reference-driven generation for authentic image completion. ACM TOG43(4), 1–12 (2024)

    2024

  45. [45]

    In: ICDAR2024 Workshops (2024)

    Tom, G., Mathew, M., Mondal, A., Karatzas, D., Jawahar, C.V., Weinman, J.: Icdar2024 challenge on occluded roadtext. In: ICDAR2024 Workshops (2024)

    2024

  46. [46]

    arXiv preprint arXiv:2307.09288 (2023)

    Touvron, H., Martin, L., Stone, K., Albert, P., Almahairi, A., Babaei, Y., Bash- lykov, N., Batra, S., Bhargava, P., Bhosale, S., et al.: Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288 (2023)

    Pith/arXiv arXiv 2023

  47. [47]

    In: CVPR

    Tsao, L.Y., Lo, Y.C., Chang, C.C., Chen, H.W., Tseng, R., Feng, C., Lee, C.Y.: Boosting flow-based generative super-resolution models via learned prior. In: CVPR. pp. 26005–26015 (2024)

    2024

  48. [48]

    In: AAAI

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

    2023

  49. [49]

    IJCV132(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. IJCV132(12), 5929–5949 (2024)

    2024

  50. [50]

    In: ICCVW

    Wang, X., Xie, L., Dong, C., Shan, Y.: Real-esrgan: Training real-world blind super-resolution with pure synthetic data. In: ICCVW. pp. 1905–1914 (2021)

    1905

  51. [51]

    In: ECCV

    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: ECCV. pp. 101–117. Springer (2020)

    2020

  52. [52]

    NIPS37, 92529–92553 (2024)

    Wu, R., Sun, L., Ma, Z., Zhang, L.: One-step effective diffusion network for real- world image super-resolution. NIPS37, 92529–92553 (2024)

    2024

  53. [53]

    In: CVPR

    Wu, R., Yang, T., Sun, L., Zhang, Z., Li, S., Zhang, L.: Seesr: Towards semantics- aware real-world image super-resolution. In: CVPR. pp. 25456–25467 (2024)

    2024

  54. [54]

    In: ICCV

    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: ICCV. pp. 11755–11765 (2025) Title Suppressed Due to Excessive Length 21

    2025

  55. [55]

    NIPS36, 15903–15935 (2023)

    Xu, J., Liu, X., Wu, Y., Tong, Y., Li, Q., Ding, M., Tang, J., Dong, Y.: Imagere- ward: Learning and evaluating human preferences for text-to-image generation. NIPS36, 15903–15935 (2023)

    2023

  56. [56]

    In: CVPR

    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: CVPR. pp. 1191–1200 (2022)

    2022

  57. [57]

    In: ECCV

    Yang, T., Wu, R., Ren, P., Xie, X., Zhang, L.: Pixel-aware stable diffusion for realistic image super-resolution and personalized stylization. In: ECCV. pp. 74–91. Springer (2024)

    2024

  58. [58]

    In: CVPR

    Ying, Z., Niu, H., Gupta, P., Mahajan, D., Ghadiyaram, D., Bovik, A.: From patches to pictures (paq-2-piq): Mapping the perceptual space of picture quality. In: CVPR. pp. 3575–3585 (2020)

    2020

  59. [59]

    In: ECCV

    You,Z.,Li,Z.,Gu,J.,Yin,Z.,Xue,T.,Dong,C.:Depictingbeyondscores:Advanc- ing image quality assessment through multi-modal language models. In: ECCV. pp. 259–276. Springer (2024)

    2024

  60. [60]

    In: CVPR

    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: CVPR. pp. 25669–25680 (2024)

    2024

  61. [61]

    In: ICCV

    Zhang, K., Liang, J., Van Gool, L., Timofte, R.: Designing a practical degradation model for deep blind image super-resolution. In: ICCV. pp. 4791–4800 (2021)

    2021

  62. [62]

    Zhang, Q., Lyu, F., Sun, Z., Wang, L., Zhang, W., Hua, W., Wu, H., Guo, Z., Wang, Y., Muennighoff, N., et al.: A survey on test-time scaling in large language models: What, how, where, and how well? arXiv preprint arXiv:2503.24235 (2025)

    Pith/arXiv arXiv 2025

  63. [63]

    In: CVPR

    Zhang, R., Isola, P., Efros, A.A., Shechtman, E., Wang, O.: The unreasonable effectiveness of deep features as a perceptual metric. In: CVPR. pp. 586–595 (2018) 22 Authors Suppressed Due to Excessive Length LQ Input w/o TTS w/o Text Prompt, w/ TTS w/o Reward Ensemble, w/ TTS ResFlow-Tuner (Ours) GT LQ Input w/o TTS w/o Text Prompt, w/ TTS w/o Reward Ensem...

    2018