Pith. sign in

REVIEW 2 major objections 2 minor 49 references

SAM-Flow confines velocity updates to attention-masked semantic regions while anchoring the rest of the latent trajectory to the source image.

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 ↗

SAM-Flow introduces source-anchored masked flows with dynamic soft masks for localized training-free semantic editing in diffusion and flow models.

T0 review reviewed 2026-06-28 challenge →

load-bearing objection SAM-Flow adds source-anchored masking to differential flow updates using scout-image attention maps, which targets background leakage but leaves mask precision as an open question. the 2 major comments →

arxiv 2606.06228 v1 pith:6E5VZZZ7 submitted 2026-06-04 cs.CV

SAM-Flow: Source-Anchored Masked Flow for Training-Free Image Editing

Haowang Cui , Rui Chen , Tao Luo , Tao Guo , Zheng Qin , Jiaze Wang This is my paper
classification cs.CV
keywords training-free image editingflow matchingmasked flowsemantic editingbackground preservationattention mapsdifferential velocitydiffusion models
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 reading

The paper introduces a source-anchored masked flow approach to prevent editing effects from spreading beyond intended areas in training-free diffusion and flow-matching pipelines. It first generates token-grounded attention maps from a scout image to mark editable semantic zones, then restricts differential velocity changes to those zones while the background follows the original source path. Time-varying projections, dynamic soft masks, transition regions, and temporal accumulation keep boundaries stable. A sympathetic reader would care because global transport in existing methods routinely alters non-target pixels, making clean localized edits impractical without retraining. The method runs plug-and-play on backbones such as Stable Diffusion 3 and FLUX.

Core claim

SAM-Flow achieves accurate semantic editing while significantly improving background preservation by first using a scout image and token-grounded attention maps to localize editable semantic regions, then applying differential velocity updates only inside those regions while anchoring the remaining latent areas to the source-image trajectory through a time-varying source-anchored projection mechanism that incorporates dynamic soft masks, transition regions, and temporal mask accumulation.

What carries the argument

Source-anchored masked flow, which localizes regions via scout-image attention maps and applies differential updates only inside dynamic soft masks while the exterior remains fixed to the source latent path.

Load-bearing premise

Token-grounded attention maps from a scout image can reliably mark exactly the semantic regions that should receive edits without missing parts or including background pixels.

What would settle it

Quantitative or visual evidence that background pixels outside the intended masks show measurable changes or artifacts after the editing process completes.

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

If this is right

  • Differential velocity updates remain confined to the localized regions, eliminating the global transport that causes background leakage in prior inversion and flow methods.
  • The time-varying projection with soft masks and temporal accumulation produces spatially stable edits and natural boundaries without extra post-processing.
  • The framework integrates directly with existing flow-matching models such as Stable Diffusion 3 and FLUX without any fine-tuning or additional training.
  • Accurate semantic changes occur inside the marked regions while the source trajectory preserves the rest of the image.

Where Pith is reading between the lines

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

  • The same attention-based localization could reduce reliance on explicit user masks in other generative editing tasks such as inpainting.
  • Temporal mask accumulation might transfer to video editing pipelines if frame-to-frame consistency of the scout attention is verified.
  • If the masks prove robust across diverse scenes, the approach could simplify interfaces by letting users specify edits via text or clicks rather than pixel-level drawing.
Share X Bluesky LinkedIn Reddit HN

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 / 2 minor

Summary. The paper introduces SAM-Flow, a source-anchored masked flow framework for training-free image editing. It localizes editable semantic regions via token-grounded attention maps computed on a scout image, applies differential velocity updates only inside those regions, and anchors the background to the source-image latent trajectory. A time-varying source-anchored projection with dynamic soft masks, transition regions, and temporal mask accumulation is added for spatial stability. The method is presented as plug-and-play with flow-matching backbones such as Stable Diffusion 3 and FLUX, with the central empirical claim being accurate semantic edits accompanied by significantly improved background preservation.

Significance. If the localization and anchoring mechanism demonstrably confines velocity updates without residual leakage, the approach would address a recurring limitation of global latent transport in training-free editing pipelines. The plug-and-play integration with widely used models and public code release would increase practical utility. Significance is currently limited by the absence of explicit quantification of the leakage-prevention claim.

major comments (2)
  1. [Abstract / §3] Abstract and §3 (method description): the load-bearing assumption that token-grounded attention maps produce masks precise enough to confine differential velocity updates without boundary leakage or artifacts is stated but not accompanied by any controlled measurement (e.g., masked-region PSNR, trajectory divergence outside the editable region, or misalignment ablation). Without such evidence the reported background-preservation gain over global differential-flow baselines cannot be verified.
  2. [§4] §4 (experiments): no table or figure reports quantitative background-preservation metrics (PSNR, LPIPS, or flow-consistency scores restricted to non-editable regions) under controlled mask misalignment; the qualitative examples alone are insufficient to substantiate the “significantly improving background preservation” claim.
minor comments (2)
  1. [Abstract] Abstract: the acronym “SAM” is not expanded on first use; clarify whether it denotes the Segment Anything Model or a different component.
  2. [§3] §3: the precise formulation of the time-varying projection operator and the schedule for temporal mask accumulation should be given as an equation or pseudocode to allow reproduction.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback highlighting the need for stronger quantitative support of the localization and anchoring claims. We address each major comment below and will incorporate additional metrics in the revision.

read point-by-point responses

  1. Referee: [Abstract / §3] Abstract and §3 (method description): the load-bearing assumption that token-grounded attention maps produce masks precise enough to confine differential velocity updates without boundary leakage or artifacts is stated but not accompanied by any controlled measurement (e.g., masked-region PSNR, trajectory divergence outside the editable region, or misalignment ablation). Without such evidence the reported background-preservation gain over global differential-flow baselines cannot be verified.

    Authors: We agree that the manuscript would benefit from explicit controlled measurements of leakage prevention. The current quantitative results report overall PSNR/LPIPS improvements and qualitative background preservation, but do not isolate non-editable regions or include misalignment ablations. In the revised version we will add a dedicated ablation table with region-restricted PSNR, LPIPS, and flow-consistency scores on non-editable areas, plus a mask-misalignment study. revision: yes

  2. Referee: [§4] §4 (experiments): no table or figure reports quantitative background-preservation metrics (PSNR, LPIPS, or flow-consistency scores restricted to non-editable regions) under controlled mask misalignment; the qualitative examples alone are insufficient to substantiate the “significantly improving background preservation” claim.

    Authors: We acknowledge that the experiments section lacks the requested region-specific quantitative metrics under controlled misalignment. While overall metrics and qualitative results support the claim, these do not directly quantify leakage. We will add the suggested tables and figures reporting PSNR/LPIPS/flow-consistency restricted to non-editable regions, together with a controlled misalignment ablation, in the revised manuscript. revision: yes

Circularity Check

0 steps flagged

No significant circularity; purely algorithmic method with no derivation chain

full rationale

The paper describes an algorithmic procedure (scout-image attention maps for region localization, differential velocity updates inside masks, source-anchored projection outside) integrated with existing flow-matching backbones. No equations, fitted parameters, or predictions are presented that could reduce to their own inputs by construction. No self-citation load-bearing steps or uniqueness theorems appear in the text. The method is self-contained as a plug-and-play technique whose validity rests on empirical results rather than any closed-form derivation.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review supplies no equations, parameters, or explicit assumptions; the ledger is therefore empty.

reviewed 2026-06-28 · how reviews work

0 comments
Cite this review

Pith. "Pith review of SAM-Flow: Source-Anchored Masked Flow for Training-Free Image Editing." pith.science (2026). https://pith.science/paper/6E5VZZZ7

@misc{pith2026260606228,
  author       = {Pith},
  title        = {Pith review of: SAM-Flow: Source-Anchored Masked Flow for Training-Free Image Editing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6E5VZZZ7}},
  note         = {Machine review of arXiv:2606.06228}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Training-free image editing has recently attracted increasing attention due to its ability to modify real images using powerful pre-trained diffusion and flow-matching models without additional training. However, existing inversion-based and differential-flow-based methods usually perform global latent transport, which inevitably propagates editing effects to non-target regions and leads to background leakage. To address this problem, we propose SAM-Flow, a source-anchored masked flow framework for localized training-free image editing. Instead of updating the whole latent representation, SAM-Flow first uses a scout image and token-grounded attention maps to localize the editable semantic regions. It then applies differential velocity updates only within these regions, while anchoring the remaining areas to the source-image latent trajectory. To further improve spatial stability and boundary naturalness, we introduce a time-varying source-anchored projection mechanism with dynamic soft masks, transition regions, and temporal mask accumulation. The proposed method is plug-and-play and can be integrated with mainstream flow-matching backbones such as Stable Diffusion 3 and FLUX without any fine-tuning. Extensive qualitative and quantitative experiments demonstrate that SAM-Flow achieves accurate semantic editing while significantly improving background preservation, providing a simple and general localized editing paradigm for training-free image editing. Code is available at: https://github.com/chwbob/Sam-Flow.

Figures

Figures reproduced from arXiv: 2606.06228 by Haowang Cui, Jiaze Wang, Rui Chen, Tao Guo, Tao Luo, Zheng Qin.

Figure 1
Figure 1. Figure 1: SAM-Flow achieves excellent semantic editing results while preserving the background of the source image. The red text indicates [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Comparison between SAM-Flow and two mainstream paradigms of training-free image editing. [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: The detailed pipeline of SAM-Flow. the source image is often incomplete. In object replacement, the new object may differ in shape, scale, and position from the original object. In object addition, the target object does not exist in the source image. In object removal, the target prompt may no longer explicitly contain the removed object. Therefore, SAM-Flow first uses a scout image to estimate the possib… view at source ↗
Figure 4
Figure 4. Figure 4: Qualitative comparison between SAM-Flow and other methods. The red text indicates the source prompt, and the green text indicates the target [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 6
Figure 6. Figure 6: Ablation study visualized by the trade-off between DINO and [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗

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

49 extracted references · 3 canonical work pages

  1. [1]

    Denoising diffusion probabilistic models,

    J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” inAdvances in Neural Information Processing Systems, vol. 33, 2020, pp. 6840–6851

    2020

  2. [2]

    Score-based generative modeling through stochastic differ- ential equations,

    Y . Song, J. Sohl-Dickstein, D. P. Kingma, A. Kumar, S. Ermon, and B. Poole, “Score-based generative modeling through stochastic differ- ential equations,” inInternational Conference on Learning Representa- tions, 2021, pp. 1–36

    2021

  3. [3]

    Flow matching for generative modeling,

    Y . Lipman, R. T. Q. Chen, H. Ben-Hamu, M. Nickel, and M. Le, “Flow matching for generative modeling,” inThe Eleventh International Conference on Learning Representations, 2023, pp. 1–28

    2023

  4. [4]

    Flow straight and fast: Learning to generate and transfer data with rectified flow,

    X. Liu, C. Gong, and Q. Liu, “Flow straight and fast: Learning to generate and transfer data with rectified flow,” inThe Eleventh International Conference on Learning Representations, 2023, pp. 1–33

    2023

  5. [5]

    Scaling rectified flow transformers for high-resolution image synthesis,

    P. Esser, S. Kulal, A. Blattmann, R. Entezari, J. M ¨uller, H. Saini, Y . Levi, D. Lorenz, A. Sauer, F. Boeselet al., “Scaling rectified flow transformers for high-resolution image synthesis,” inForty-First International Conference on Machine Learning, 2024, pp. 1–28

    2024

  6. [6]

    (2024) Flux.1

    Black Forest Labs. (2024) Flux.1. [Online]. Available: https:// blackforestlabs.ai/

    2024

  7. [7]

    Sdedit: Guided image synthesis and editing with stochastic differential equations,

    C. Meng, Y . He, Y . Song, J. Song, J. Wu, J.-Y . Zhu, and S. Ermon, “Sdedit: Guided image synthesis and editing with stochastic differential equations,” inInternational Conference on Learning Representations, 2022, pp. 1–33

    2022

  8. [8]

    Imagic: Text-based real image editing with diffusion models,

    B. Kawar, S. Zada, O. Lang, O. Tov, H. Chang, T. Dekel, I. Mosseri, and M. Irani, “Imagic: Text-based real image editing with diffusion models,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 6007–6017

    2023

  9. [9]

    Instructpix2pix: Learning to follow image editing instructions,

    T. Brooks, A. Holynski, and A. A. Efros, “Instructpix2pix: Learning to follow image editing instructions,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 18 392–18 402

    2023

  10. [10]

    Prompt-to-prompt image editing with cross-attention control,

    A. Hertz, R. Mokady, J. Tenenbaum, K. Aberman, Y . Pritch, and D. Cohen-Or, “Prompt-to-prompt image editing with cross-attention control,” inThe Eleventh International Conference on Learning Rep- resentations, 2023, pp. 1–36

    2023

  11. [11]

    Plug-and-play diffusion features for text-driven image-to-image translation,

    N. Tumanyan, M. Geyer, S. Bagon, and T. Dekel, “Plug-and-play diffusion features for text-driven image-to-image translation,” inPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 1921–1930

    2023

  12. [12]

    Null- text inversion for editing real images using guided diffusion models,

    R. Mokady, A. Hertz, K. Aberman, Y . Pritch, and D. Cohen-Or, “Null- text inversion for editing real images using guided diffusion models,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 6038–6047

    2023

  13. [13]

    Pnp inversion: Boosting diffusion-based editing with 3 lines of code,

    X. Ju, A. Zeng, Y . Bian, S. Liu, and Q. Xu, “Pnp inversion: Boosting diffusion-based editing with 3 lines of code,” inThe Twelfth Interna- tional Conference on Learning Representations, 2024, pp. 1–28

    2024

  14. [14]

    Inversion-free image editing with language-guided diffusion models,

    S. Xu, Y . Huang, J. Pan, Z. Ma, and J. Chai, “Inversion-free image editing with language-guided diffusion models,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 9452–9461

    2024

  15. [15]

    An edit friendly ddpm noise space: Inversion and manipulations,

    I. Huberman-Spiegelglas, V . Kulikov, and T. Michaeli, “An edit friendly ddpm noise space: Inversion and manipulations,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 12 469–12 478

    2024

  16. [16]

    Edict: Exact diffusion inversion via coupled transformations,

    B. Wallace, A. Gokul, and N. Naik, “Edict: Exact diffusion inversion via coupled transformations,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2023, pp. 22 532–22 541

    2023

  17. [17]

    Direct inversion: Optimization-free text-driven real image editing with diffusion models,

    A. Elarabawy, H. Kamath, and S. Denton, “Direct inversion: Optimization-free text-driven real image editing with diffusion models,” arXiv preprint arXiv:2211.07825, 2022

    work page arXiv 2022

  18. [18]

    Flowedit: Inversion-free text-based editing using pre-trained flow mod- els,

    V . Kulikov, M. Kleiner, I. Huberman-Spiegelglas, and T. Michaeli, “Flowedit: Inversion-free text-based editing using pre-trained flow mod- els,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2025, pp. 19 721–19 730

    2025

  19. [19]

    What the daam: Interpreting stable diffusion using cross attention,

    R. Tang, L. Liu, A. Pandey, Z. Jiang, G. Yang, K. Kumar, P. Stenetorp, J. Lin, and F. T ¨ure, “What the daam: Interpreting stable diffusion using cross attention,” inProceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2023, pp. 5644–5659

    2023

  20. [20]

    Attend-and- excite: Attention-based semantic guidance for text-to-image diffusion models,

    H. Chefer, Y . Alaluf, Y . Vinker, L. Wolf, and D. Cohen-Or, “Attend-and- excite: Attention-based semantic guidance for text-to-image diffusion models,”ACM transactions on Graphics (TOG), vol. 42, no. 4, pp. 1– 10, 2023

    2023

  21. [21]

    Masactrl: Tuning-free mutual self-attention control for consistent image synthesis and editing,

    M. Cao, X. Wang, Z. Qi, Y . Shan, X. Qie, and Y . Zheng, “Masactrl: Tuning-free mutual self-attention control for consistent image synthesis and editing,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 22 560–22 570

    2023

  22. [22]

    High- resolution image synthesis with latent diffusion models,

    R. Rombach, A. Blattmann, D. Lorenz, P. Esser, and B. Ommer, “High- resolution image synthesis with latent diffusion models,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion, 2022, pp. 10 684–10 695

    2022

  23. [23]

    Denoising diffusion implicit models,

    J. Song, C. Meng, and S. Ermon, “Denoising diffusion implicit models,” inInternational Conference on Learning Representations, 2021, pp. 1– 20

    2021

  24. [24]

    Negative-prompt inversion: Fast image inversion for editing with text-guided diffusion models,

    D. Miyake, A. Iohara, Y . Saito, and T. Tanaka, “Negative-prompt inversion: Fast image inversion for editing with text-guided diffusion models,” in2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2025, pp. 2063–2072

    2025

  25. [25]

    Taming rectified flow for inversion and editing,

    J. Wang, J. Pu, Z. Qi, J. Guo, Y . Ma, N. Huang, Y . Chen, X. Li, and Y . Shan, “Taming rectified flow for inversion and editing,” in International Conference on Machine Learning. PMLR, 2025, pp. 64 044–64 058

    2025

  26. [26]

    Stable flow: Vital layers for training- free image editing,

    O. Avrahami, O. Patashnik, O. Fried, E. Nemchinov, K. Aberman, D. Lischinski, and D. Cohen-Or, “Stable flow: Vital layers for training- free image editing,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, June 2025, pp. 7877–7888

    2025

  27. [27]

    Reflex: Text-guided editing of real images in rectified flow via mid-step feature extraction and attention adaptation,

    J. Kim, J. Park, Y . Song, N. Kwak, and W. Rhee, “Reflex: Text-guided editing of real images in rectified flow via mid-step feature extraction and attention adaptation,” inProceedings of the IEEE/CVF International Conference on Computer Vision, October 2025, pp. 15 939–15 948

    2025

  28. [28]

    Flowalign: Trajectory- regularized, inversion-free flow-based image editing,

    J. Kim, Y . Hong, J. Park, and J. C. Ye, “Flowalign: Trajectory- regularized, inversion-free flow-based image editing,” inThe Fourteenth International Conference on Learning Representations, 2026

    2026

  29. [29]

    Blended diffusion for text- driven editing of natural images,

    O. Avrahami, D. Lischinski, and O. Fried, “Blended diffusion for text- driven editing of natural images,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, June 2022, pp. 18 208–18 218

    2022

  30. [30]

    Blended latent diffusion,

    O. Avrahami, O. Fried, and D. Lischinski, “Blended latent diffusion,” ACM Transactions on Graphics, vol. 42, no. 4, pp. 1–11, 2023

    2023

  31. [31]

    Glide: Towards photorealistic image gen- eration and editing with text-guided diffusion models,

    A. Q. Nichol, P. Dhariwal, A. Ramesh, P. Shyam, P. Mishkin, B. Mcgrew, I. Sutskever, and M. Chen, “Glide: Towards photorealistic image gen- eration and editing with text-guided diffusion models,” inInternational Conference on Machine Learning. PMLR, 2022, pp. 16 784–16 804

    2022

  32. [32]

    Repaint: Inpainting using denoising diffusion probabilistic models,

    A. Lugmayr, M. Danelljan, A. Romero, F. Yu, R. Timofte, and L. Van Gool, “Repaint: Inpainting using denoising diffusion probabilistic models,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, June 2022, pp. 11 461–11 471

    2022

  33. [33]

    Diffedit: Diffusion- based semantic image editing with mask guidance,

    G. Couairon, J. Verbeek, H. Schwenk, and M. Cord, “Diffedit: Diffusion- based semantic image editing with mask guidance,” inThe Eleventh International Conference on Learning Representations, 2023, pp. 1–22

    2023

  34. [34]

    Paint by example: Exemplar-based image editing with diffusion models,

    B. Yang, S. Gu, B. Zhang, T. Zhang, X. Chen, X. Sun, D. Chen, and F. Wen, “Paint by example: Exemplar-based image editing with diffusion models,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, June 2023, pp. 18 381–18 391

    2023

  35. [35]

    Inpaint anything: Segment anything meets image inpainting.arXiv preprint arXiv:2304.06790, 2023

    T. Yu, R. Feng, R. Feng, J. Liu, X. Jin, W. Zeng, and Z. Chen, “Inpaint anything: Segment anything meets image inpainting,”arXiv preprint arXiv:2304.06790, 2023

    work page arXiv 2023

  36. [36]

    Adding conditional control to text-to-image diffusion models,

    L. Zhang, A. Rao, and M. Agrawala, “Adding conditional control to text-to-image diffusion models,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 3836–3847

    2023

  37. [37]

    T2i- adapter: Learning adapters to dig out more controllable ability for text- to-image diffusion models,

    C. Mou, X. Wang, L. Xie, Y . Wu, J. Zhang, Z. Qi, and Y . Shan, “T2i- adapter: Learning adapters to dig out more controllable ability for text- to-image diffusion models,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 38, no. 5, 2024, pp. 4296–4304

    2024

  38. [38]

    Gligen: Open-set grounded text-to-image generation,

    Y . Li, H. Liu, Q. Wu, F. Mu, J. Yang, J. Gao, C. Li, and Y . J. Lee, “Gligen: Open-set grounded text-to-image generation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 22 511–22 521

    2023

  39. [39]

    Segment anything,

    A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.-Y . Lo, P. Dollar, and R. Girshick, 10 “Segment anything,” inProceedings of the IEEE/CVF International Conference on Computer Vision, October 2023, pp. 4015–4026

    2023

  40. [40]

    Auto-encoding variational bayes,

    D. P. Kingma and M. Welling, “Auto-encoding variational bayes,” in International Conference on Learning Representations, 2014, pp. 1–14

    2014

  41. [41]

    Scalable diffusion models with transformers,

    W. Peebles and S. Xie, “Scalable diffusion models with transformers,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 4195–4205

    2023

  42. [42]

    Exploring the limits of transfer learning with a unified text-to-text transformer,

    C. Raffel, N. Shazeer, A. Roberts, K. Lee, S. Narang, M. Matena, Y . Zhou, W. Li, and P. J. Liu, “Exploring the limits of transfer learning with a unified text-to-text transformer,”Journal of Machine Learning Research, vol. 21, no. 140, pp. 1–67, 2020

    2020

  43. [43]

    Learning transferable visual models from natural language supervision,

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clarket al., “Learning transferable visual models from natural language supervision,” inInternational Conference on Machine Learning. PMLR, 2021, pp. 8748–8763

    2021

  44. [44]

    Ntire 2017 challenge on single image super-resolution: Dataset and study,

    E. Agustsson and R. Timofte, “Ntire 2017 challenge on single image super-resolution: Dataset and study,” inProceedings of the IEEE Con- ference on Computer Vision and Pattern Recognition Workshops, 2017, pp. 126–135

    2017

  45. [45]

    Text-to-image rectified flow as plug-and-play priors,

    X. Yang, C. Cheng, X. Yang, F. Liu, and G. Lin, “Text-to-image rectified flow as plug-and-play priors,” inThe Thirteenth International Conference on Learning Representations, 2025, pp. 1–25

    2025

  46. [46]

    Semantic image inversion and editing using rectified stochastic differential equations,

    L. Rout, Y . Chen, N. Ruiz, C. Caramanis, S. Shakkottai, and W.-S. Chu, “Semantic image inversion and editing using rectified stochastic differential equations,” inThe Thirteenth International Conference on Learning Representations, 2025, pp. 1–36

    2025

  47. [47]

    Emerging properties in self-supervised vision transformers,

    M. Caron, H. Touvron, I. Misra, H. J ´egou, J. Mairal, P. Bojanowski, and A. Joulin, “Emerging properties in self-supervised vision transformers,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 9650–9660

    2021

  48. [48]

    Image quality assessment: From error visibility to structural similarity,

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

    2004

  49. [49]

    The unreasonable effectiveness of deep features as a perceptual metric,

    R. Zhang, P. Isola, A. A. Efros, E. Shechtman, and O. Wang, “The unreasonable effectiveness of deep features as a perceptual metric,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018, pp. 586–595. A-1 APPENDIXA BACKBONE-SPECIFICTOKEN-GROUNDEDATTENTIONEXTRACTION The main paper describes the general scout-and-union loca...

    work page arXiv 2018

This paper was first reviewed by grok-4.3 on June 28, 2026.