pith. sign in

arxiv: 2605.24762 · v1 · pith:5LIJLHOMnew · submitted 2026-05-23 · 💻 cs.CV

4KLSDB: A Large-Scale Dataset for 4K Image Restoration and Generation

Zihao Zhu , Kuan-Ru Huang , Zhaoming Xu , Renjie Li , Bo Wu , Ruizheng Bai , Mingyang Wu , Sayak Paul
show 1 more author
Zhengzhong Tu
This is my paper

Pith reviewed 2026-06-30 13:01 UTC · model grok-4.3

classification 💻 cs.CV
keywords 4K image datasetsuper-resolutiontext-to-image diffusionimage restorationhigh-resolution benchmarksdataset curationnative 4K training
0
0 comments X

The pith

A new dataset of 129k native 4K images improves super-resolution and diffusion model results on 4K benchmarks.

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

The paper presents 4KLSDB, a collection of 129,484 curated 4K images drawn from open sources and filtered through automated and human stages to maintain quality and diversity across categories. It supplies separate validation and test sets and shows that models trained on this true 4K data achieve measurable gains on native 4K restoration and generation tasks. The work rests on the observation that current public datasets fall short in both resolution and scale for state-of-the-art training. If the claim holds, researchers gain a practical resource for pushing image fidelity higher without downsampling artifacts.

Core claim

4KLSDB supplies 129,484 carefully filtered 4K images together with 2,000 validation and 1,984 test images; when representative super-resolution and diffusion models are trained on it, performance rises on native 4K benchmarks, and a positive correlation appears between use of true 4K training data and higher fidelity in restoration.

What carries the argument

4KLSDB, the large-scale native 4K image collection assembled via multi-stage automated filtering and annotation with human and multimodal model checks.

If this is right

  • Super-resolution models trained on 4KLSDB recover finer detail on 4K inputs than models trained on lower-resolution data.
  • Text-to-image diffusion models fine-tuned on the dataset produce higher-fidelity 4K outputs on the supplied test set.
  • A direct link exists between training resolution matching the target resolution and measured fidelity gains in restoration tasks.

Where Pith is reading between the lines

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

  • Other high-resolution tasks such as inpainting or deblurring at 4K may see similar benefits if retrained on the same source.
  • The dataset size and category balance could support scaling laws studies that isolate resolution from data volume.
  • Public release of the curation pipeline would let later work test whether the same gains appear on different base models.

Load-bearing premise

The filtering and annotation steps produce a dataset free of selection bias or artifacts that would prevent the observed performance gains from generalizing.

What would settle it

Train the same super-resolution and diffusion models on a matched-scale dataset of downsampled 4K images and measure whether the 4K-native version still yields higher scores on the paper's 4K test benchmarks.

Figures

Figures reproduced from arXiv: 2605.24762 by Bo Wu, Kuan-Ru Huang, Mingyang Wu, Renjie Li, Ruizheng Bai, Sayak Paul, Zhaoming Xu, Zhengzhong Tu, Zihao Zhu.

Figure 1
Figure 1. Figure 1: Overview of the 4KLSDB filtering pipeline. An initial raw image pool is progressively refined through automated filters and a final [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Visual comparison of SeeSR on the 4KLSDB real-SR test set at [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Visual comparison between the original SANA [ [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
read the original abstract

High-resolution datasets are essential for advancing super-resolution (SR) and text-to-image (T2I) diffusion research. However, current publicly available datasets lack both the native 4K resolution and the extensive scale necessary for training state-of-the-art models. To address this gap, we introduce a 4K Large Scale Dataset and Benchmark (4KLSDB), a large-scale, diverse dataset consisting of 129,484 carefully curated 4K resolution images spanning multiple categories such as nature, urban scenes, people, food, artwork, and CGI, alongside distinct validation and test sets containing 2,000 and 1,984 images respectively. Images were sourced from established open datasets including Photo Concept Bucket, Laion2B, and PD12M. 4KLSDB underwent rigorous multi-stage automated filtering and annotation pipelines involving both human annotators and Large Multimodal Models (LMMs) to ensure high aesthetic quality and dataset consistency. We demonstrate 4KLSDB's effectiveness by training representative super-resolution and diffusion models, observing significant improvements in performance on native 4K benchmarks. Comprehensive experiments illustrate a positive correlation between training on true 4K resolution data and improved fidelity in image restoration task, especially on 4K resolution. We provide the research community a valuable resource to drive progress toward genuinely high-fidelity image synthesis and restoration by providing 4KLSDB. Our project page is available at: https://4klsdb.github.io/.

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.

Referee Report

2 major / 1 minor

Summary. The paper introduces 4KLSDB, a curated dataset of 129,484 native 4K images drawn from LAION-2B, PD12M and related sources, spanning categories such as nature, urban scenes, people, food, artwork and CGI. It supplies distinct validation (2,000 images) and test (1,984 images) splits. Images are processed via multi-stage automated filtering plus human and LMM annotation to enforce aesthetic quality and consistency. Effectiveness is demonstrated by training representative super-resolution and diffusion models, with reported significant gains on native 4K benchmarks and a positive correlation between true 4K training data and restoration fidelity.

Significance. A well-curated, large-scale native 4K dataset would be a useful resource for super-resolution and high-resolution text-to-image research, filling a documented gap in publicly available training data at this resolution.

major comments (2)
  1. [Abstract] Abstract: the central effectiveness claim ('significant improvements in performance on native 4K benchmarks' and 'positive correlation between training on true 4K resolution data and improved fidelity') is stated without any quantitative metrics, baseline models, ablation tables, or statistical details, rendering it impossible to assess whether gains arise from native 4K scale or from curation-induced distribution shifts.
  2. [Abstract] Abstract (filtering pipeline): the multi-stage automated filtering plus LMM/human annotation is asserted to produce high aesthetic quality and consistency 'without introducing selection bias or artifacts,' yet no before/after aesthetic-score distributions, inter-rater agreement statistics, or ablation experiments comparing models trained on filtered versus raw source subsets of equal size are supplied; this verification is load-bearing for attributing performance gains to 4K resolution rather than selection effects.
minor comments (1)
  1. [Abstract] Clarify whether the reported 129,484-image count includes or excludes the validation and test splits.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on the abstract. We agree that strengthening the quantitative support and evidence for the filtering pipeline in the abstract will improve the manuscript. We address each point below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central effectiveness claim ('significant improvements in performance on native 4K benchmarks' and 'positive correlation between training on true 4K resolution data and improved fidelity') is stated without any quantitative metrics, baseline models, ablation tables, or statistical details, rendering it impossible to assess whether gains arise from native 4K scale or from curation-induced distribution shifts.

    Authors: We agree the abstract would benefit from explicit metrics. The full manuscript reports quantitative results from SR and diffusion model training, including PSNR/SSIM gains on native 4K benchmarks and correlation analysis between resolution and fidelity. We will revise the abstract to include representative quantitative values, baseline comparisons, and statistical details so readers can evaluate whether improvements stem from native 4K data. revision: yes

  2. Referee: [Abstract] Abstract (filtering pipeline): the multi-stage automated filtering plus LMM/human annotation is asserted to produce high aesthetic quality and consistency 'without introducing selection bias or artifacts,' yet no before/after aesthetic-score distributions, inter-rater agreement statistics, or ablation experiments comparing models trained on filtered versus raw source subsets of equal size are supplied; this verification is load-bearing for attributing performance gains to 4K resolution rather than selection effects.

    Authors: The manuscript details the multi-stage pipeline (automated filters, LMMs, and human annotation) in Section 3 to enforce quality. We acknowledge that before/after distributions, inter-rater statistics, and equal-size filtered-vs-raw ablations would further isolate resolution effects from curation. We will revise the abstract to qualify the claim and add references to any existing pipeline statistics; if these specific ablations are absent, we will consider adding a concise analysis. revision: partial

Circularity Check

0 steps flagged

No circularity: dataset paper with empirical demonstration only.

full rationale

The paper introduces a curated 4K image dataset and reports standard training experiments on SR and diffusion models to show performance gains on 4K benchmarks. No equations, parameters, or derivations are present. No self-definitional steps, fitted inputs renamed as predictions, or load-bearing self-citations for theorems exist. The central claim reduces to empirical results on held-out test sets rather than any construction that equates output to input by definition. Lack of ablations on curation is a verification gap, not circularity.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The paper rests on the domain assumption that images from the cited open datasets can be filtered into a high-quality 4K collection; no free parameters or invented entities are introduced.

axioms (1)
  • domain assumption Images sourced from Photo Concept Bucket, Laion2B, and PD12M can be filtered via automated and human/LMM pipelines to produce consistent high-aesthetic-quality 4K data.
    The entire dataset construction depends on this premise about source suitability and filtering efficacy.

pith-pipeline@v0.9.1-grok · 5824 in / 1250 out tokens · 41082 ms · 2026-06-30T13:01:31.420336+00:00 · methodology

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

60 extracted references · 7 canonical work pages · 3 internal anchors

  1. [1]

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

    Eirikur Agustsson and Radu Timofte. Ntire 2017 challenge on single image super-resolution: Dataset and study. InThe IEEE Conference on Computer Vision and Pattern Recogni- tion (CVPR) Workshops, 2017. 2, 4

    2017

  2. [2]

    Hitsr: A hierarchical transformer for reference-based super- resolution.arXiv preprint arXiv:2408.16959, 2024

    Masoomeh Aslahishahri, Jordan Ubbens, and Ian Stavness. Hitsr: A hierarchical transformer for reference-based super- resolution.arXiv preprint arXiv:2408.16959, 2024. 5

    work page arXiv 2024

  3. [3]

    A non- local algorithm for image denoising

    Antoni Buades, Bartomeu Coll, and J-M Morel. A non- local algorithm for image denoising. In2005 IEEE computer society conference on computer vision and pattern recognition (CVPR’05), pages 60–65. Ieee, 2005. 2

    2005

  4. [4]

    Toward real-world single image super-resolution: A new benchmark and a new model

    Jianrui Cai, Hui Zeng, Hongwei Yong, Zisheng Cao, and Lei Zhang. Toward real-world single image super-resolution: A new benchmark and a new model. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 3086–3095, 2019. 2

    2019

  5. [5]

    Pixart- σ: Weak-to-strong training of diffusion transformer for 4k text-to-image generation

    Junsong Chen, Chongjian Ge, Enze Xie, Yue Wu, Lewei Yao, Xiaozhe Ren, Zhongdao Wang, Ping Luo, Huchuan Lu, and Zhenguo Li. Pixart- σ: Weak-to-strong training of diffusion transformer for 4k text-to-image generation. InEuropean Conference on Computer Vision, pages 74–91. Springer, 2024. 2

    2024

  6. [6]

    Image quality assessment: Unifying structure and texture similarity.IEEE transactions on pattern analysis and machine intelligence, 44(5):2567–2581, 2020

    Keyan Ding, Kede Ma, Shiqi Wang, and Eero P Simoncelli. Image quality assessment: Unifying structure and texture similarity.IEEE transactions on pattern analysis and machine intelligence, 44(5):2567–2581, 2020. 5, 7

    2020

  7. [7]

    Scale guided hypernetwork for blind super-resolution image quality assessment.arXiv preprint arXiv:2306.02398,

    Jun Fu. Scale guided hypernetwork for blind super-resolution image quality assessment.arXiv preprint arXiv:2306.02398,

    work page arXiv

  8. [8]

    Vefx-bench: A holistic bench- mark for generic video editing and visual effects, 2026

    Xiangbo Gao, Sicong Jiang, Bangya Liu, Xinghao Chen, Minglai Yang, Siyuan Yang, Mingyang Wu, Jiongze Yu, Qi Zheng, Haozhi Wang, Jiayi Zhang, Jie Yang, Zihan Wang, Qing Yin, and Zhengzhong Tu. Vefx-bench: A holistic bench- mark for generic video editing and visual effects, 2026. 2

    2026

  9. [9]

    Geneval: An object-focused framework for evaluating text- to-image alignment

    Dhruba Ghosh, Hannaneh Hajishirzi, and Ludwig Schmidt. Geneval: An object-focused framework for evaluating text- to-image alignment. InAdvances in Neural Information Pro- cessing Systems, 2023. 3

    2023

  10. [10]

    Div8k: Diverse 8k resolution image dataset

    Shuhang Gu, Andreas Lugmayr, Martin Danelljan, Manuel Fritsche, Julien Lamour, and Radu Timofte. Div8k: Diverse 8k resolution image dataset. In2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW), pages 3512–3516. IEEE, 2019. 2, 4, 5

    2019

  11. [11]

    Mambair: A simple baseline for image restoration with state-space model

    Hang Guo, Jinmin Li, Tao Dai, Zhihao Ouyang, Xudong Ren, and Shu-Tao Xia. Mambair: A simple baseline for image restoration with state-space model. InEuropean conference on computer vision, pages 222–241. Springer, 2024. 5

    2024

  12. [12]

    Clipscore: A reference-free evaluation metric for image captioning

    Jack Hessel, Ari Holtzman, Maxwell Forbes, Ronan Le Bras, and Yejin Choi. Clipscore: A reference-free evaluation metric for image captioning. InProceedings of the 2021 conference on empirical methods in natural language processing, pages 7514–7528, 2021. 5

    2021

  13. [13]

    Gans trained by a two time-scale update rule converge to a local nash equilibrium

    Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bern- hard Nessler, and Sepp Hochreiter. Gans trained by a two time-scale update rule converge to a local nash equilibrium. Advances in neural information processing systems, 30, 2017. 5, 7

    2017

  14. [14]

    Denoising diffu- sion probabilistic models.Advances in neural information processing systems, 33:6840–6851, 2020

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffu- sion probabilistic models.Advances in neural information processing systems, 33:6840–6851, 2020. 2

    2020

  15. [15]

    T2i-compbench: A comprehensive benchmark for open- world compositional text-to-image generation

    Kaiyi Huang, Kaiyue Sun, Enze Xie, Zhenguo Li, and Xihui Liu. T2i-compbench: A comprehensive benchmark for open- world compositional text-to-image generation. InAdvances in Neural Information Processing Systems, 2023. 3

    2023

  16. [16]

    Hq-edit: A high-quality dataset for instruction-based image editing, 2024

    Mude Hui, Siwei Yang, Bingchen Zhao, Yichun Shi, Heng Wang, Peng Wang, Yuyin Zhou, and Cihang Xie. Hq-edit: A high-quality dataset for instruction-based image editing. arXiv preprint arXiv:2404.09990, 2024. 2

    work page arXiv 2024

  17. [17]

    Blind decon- volution using a normalized sparsity measure

    Dilip Krishnan, Terence Tay, and Rob Fergus. Blind decon- volution using a normalized sparsity measure. InCVPR 2011, pages 233–240. IEEE, 2011. 2

    2011

  18. [18]

    Photo- realistic single image super-resolution using a generative ad- versarial network

    Christian Ledig, Lucas Theis, Ferenc Huszár, Jose Caballero, Andrew Cunningham, Alejandro Acosta, Andrew Aitken, Alykhan Tejani, Johannes Totz, Zehan Wang, et al. Photo- realistic single image super-resolution using a generative ad- versarial network. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 4681–4690,

  19. [19]

    Playground v2.5: Three insights to- wards enhancing aesthetic quality in text-to-image generation,

    Daiqing Li, Aleks Kamko, Ehsan Akhgari, Ali Sabet, Linmiao Xu, and Suhail Doshi. Playground v2.5: Three insights to- wards enhancing aesthetic quality in text-to-image generation,

  20. [20]

    A sys- tematic survey of deep learning-based single-image super- resolution.ACM Computing Surveys, 56(10):1–40, 2024

    Juncheng Li, Zehua Pei, Wenjie Li, Guangwei Gao, Long- guang Wang, Yingqian Wang, and Tieyong Zeng. A sys- tematic survey of deep learning-based single-image super- resolution.ACM Computing Surveys, 56(10):1–40, 2024. 2

    2024

  21. [21]

    Lsdir: A large scale dataset for image restoration

    Yawei Li, Kai Zhang, Jingyun Liang, Jiezhang Cao, Ce Liu, Rui Gong, Yulun Zhang, Hao Tang, Yun Liu, Denis Deman- dolx, et al. Lsdir: A large scale dataset for image restoration. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 1775–1787, 2023. 2, 4

    2023

  22. [22]

    Swinir: Image restoration using swin transformer

    Jingyun Liang, Jiezhang Cao, Guolei Sun, Kai Zhang, Luc Van Gool, and Radu Timofte. Swinir: Image restoration using swin transformer. InProceedings of the IEEE/CVF international conference on computer vision, pages 1833– 1844, 2021. 2, 5

    2021

  23. [23]

    Public domain 12m: A highly aesthetic image-text dataset with novel governance mechanisms, 2024

    Jordan Meyer, Nick Padgett, Cullen Miller, and Laura Exline. Public domain 12m: A highly aesthetic image-text dataset with novel governance mechanisms, 2024. 3

    2024

  24. [24]

    completely blind

    Anish Mittal, Rajiv Soundararajan, and Alan C Bovik. Mak- ing a “completely blind” image quality analyzer.IEEE Signal processing letters, 20(3):209–212, 2012. 5

    2012

  25. [25]

    completely blind

    Anish Mittal, Rajiv Soundararajan, and Alan C. Bovik. Mak- ing a “completely blind” image quality analyzer.IEEE Signal Processing Letters, 20(3):209–212, 2013. 7

    2013

  26. [26]

    Deep multi-scale convolutional neural network for dynamic scene deblurring

    Seungjun Nah, Tae Hyun Kim, and Kyoung Mu Lee. Deep multi-scale convolutional neural network for dynamic scene deblurring. InProceedings of the IEEE conference on com- puter vision and pattern recognition, pages 3883–3891, 2017. 2

    2017

  27. [27]

    Rethinking image super-resolution from training data perspectives

    Go Ohtani, Ryu Tadokoro, Ryosuke Yamada, Yuki M Asano, Iro Laina, Christian Rupprecht, Nakamasa Inoue, Rio Yokota, Hirokatsu Kataoka, and Yoshimitsu Aoki. Rethinking image super-resolution from training data perspectives. InEuropean Conference on Computer Vision, pages 19–36. Springer, 2024. 4

    2024

  28. [28]

    Unveiling hidden details: A raw data-enhanced paradigm for real-world super-resolution.arXiv preprint arXiv:2411.10798, 2024

    Long Peng, Wenbo Li, Jiaming Guo, Xin Di, Haoze Sun, Yong Li, Renjing Pei, Yang Wang, Yang Cao, and Zheng- Jun Zha. Unveiling hidden details: A raw data-enhanced paradigm for real-world super-resolution.arXiv preprint arXiv:2411.10798, 2024. 2

    work page arXiv 2024

  29. [29]

    Sdxl: Improving latent diffusion models for high-resolution image synthesis, 2023

    Dustin Podell, Zion English, Kyle Lacey, Andreas Blattmann, Tim Dockhorn, Jonas Müller, Joe Penna, and Robin Rombach. Sdxl: Improving latent diffusion models for high-resolution image synthesis, 2023. 3

    2023

  30. [30]

    Hierarchical Text-Conditional Image Generation with CLIP Latents

    Aditya Ramesh, Prafulla Dhariwal, Alex Nichol, Casey Chu, and Mark Chen. Hierarchical text-conditional image genera- tion with clip latents.arXiv preprint arXiv:2204.06125, 1(2): 3, 2022. 2

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  31. [31]

    Benjamin Erichson, Junyi Guo, Shashank Subra- manian, Omer San, Zarija Lukic, and Michael W

    Pu Ren, N. Benjamin Erichson, Junyi Guo, Shashank Subra- manian, Omer San, Zarija Lukic, and Michael W. Mahoney. Superbench: A super-resolution benchmark dataset for sci- entific machine learning.Data-centric Machine Learning Research, 2(8):1–45, 2025. 2

    2025

  32. [32]

    High-resolution image synthesis with latent diffusion models.CVPR, 2022

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Björn Ommer. High-resolution image synthesis with latent diffusion models.CVPR, 2022. 2

    2022

  33. [33]

    Sara Mahdavi, Rapha Gontijo Lopes, Tim Salimans, Jonathan Ho, David J

    Chitwan Saharia, William Chan, Saurabh Saxena, Lala Li, Jay Whang, Emily Denton, Seyed Kamyar Seyed Ghasemipour, Burcu Karagol Ayan, S. Sara Mahdavi, Rapha Gontijo Lopes, Tim Salimans, Jonathan Ho, David J. Fleet, and Mohammad Norouzi. Photorealistic text-to-image diffusion models with deep language understanding. InAdvances in Neural Infor- mation Proces...

    2022

  34. [34]

    Fleet, and Mohammad Norouzi

    Chitwan Saharia, Jonathan Ho, William Chan, Tim Sali- mans, David J. Fleet, and Mohammad Norouzi. Image super- resolution via iterative refinement.IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(4):4713–4726,

  35. [35]

    Laion-5b: An open large-scale dataset for training next gen- eration image-text models.Advances in neural information processing systems, 35:25278–25294, 2022

    Christoph Schuhmann, Romain Beaumont, Richard Vencu, Cade Gordon, Ross Wightman, Mehdi Cherti, Theo Coombes, Aarush Katta, Clayton Mullis, Mitchell Wortsman, et al. Laion-5b: An open large-scale dataset for training next gen- eration image-text models.Advances in neural information processing systems, 35:25278–25294, 2022. 3

    2022

  36. [36]

    Scaling vision pre-training to 4k resolution

    Baifeng Shi, Boyi Li, Han Cai, Yao Lu, Sifei Liu, Marco Pavone, Jan Kautz, Song Han, Trevor Darrell, Pavlo Molchanov, et al. Scaling vision pre-training to 4k resolution. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 9631–9640, 2025. 3

    2025

  37. [37]

    Journeydb: A benchmark for generative image understanding, 2023

    Keqiang Sun, Junting Pan, Yuying Ge, Hao Li, Haodong Duan, Xiaoshi Wu, Renrui Zhang, Aojun Zhou, Zequan Qin, Yi Wang, Jifeng Dai, Yu Qiao, and Hongsheng Li. Journeydb: A benchmark for generative image understanding, 2023. 3

    2023

  38. [38]

    Jianyi Wang, Zongsheng Yue, Shangchen Zhou, Kelvin C. K. Chan, and Chen Change Loy. Exploiting diffusion prior for real-world image super-resolution.International Journal of Computer Vision, 2024. 2

    2024

  39. [39]

    Qwen2-vl: Enhancing vision-language model’s perception of the world at any resolution, 2024

    Peng Wang, Shuai Bai, Sinan Tan, Shijie Wang, Zhihao Fan, Jinze Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Yang Fan, Kai Dang, Mengfei Du, Xuancheng Ren, Rui Men, Dayiheng Liu, Chang Zhou, Jingren Zhou, and Junyang Lin. Qwen2-vl: Enhancing vision-language model’s perception of the world at any resolution, 2024. 3

    2024

  40. [40]

    Esrgan: En- hanced super-resolution generative adversarial networks

    Xintao Wang, Ke Yu, Shixiang Wu, Jinjin Gu, Yihao Liu, Chao Dong, Yu Qiao, and Chen Change Loy. Esrgan: En- hanced super-resolution generative adversarial networks. In Proceedings of the European conference on computer vision (ECCV) workshops, pages 0–0, 2018. 2

    2018

  41. [41]

    Real-esrgan: Training real-world blind super-resolution with pure synthetic data

    Xintao Wang, Liangbin Xie, Chao Dong, and Ying Shan. Real-esrgan: Training real-world blind super-resolution with pure synthetic data. InProceedings of the IEEE/CVF Inter- national Conference on Computer Vision Workshops, pages 1905–1914, 2021. 2

    1905

  42. [42]

    Bovik, H.R

    Zhou 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, 13(4): 600–612, 2004. 7

    2004

  43. [43]

    Image quality assessment: from error visibility to structural similarity.IEEE transactions on image processing, 13(4):600–612, 2004

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Simoncelli. Image quality assessment: from error visibility to structural similarity.IEEE transactions on image processing, 13(4):600–612, 2004. 5

    2004

  44. [44]

    Diffusiondb: A large-scale prompt gallery dataset for text-to-image genera- tive models

    Zijie J Wang, Evan Montoya, David Munechika, Haoyang Yang, Benjamin Hoover, and Duen Horng Chau. Diffusiondb: A large-scale prompt gallery dataset for text-to-image genera- tive models. InProceedings of the 61st annual meeting of the association for computational linguistics (volume 1: Long papers), pages 893–911, 2023. 2, 4

    2023

  45. [45]

    Component divide-and- conquer for real-world image super-resolution

    Pengxu Wei, Ziwei Xie, Hannan Lu, Zongyuan Zhan, Qixiang Ye, Wangmeng Zuo, and Liang Lin. Component divide-and- conquer for real-world image super-resolution. InEuropean Conference on Computer Vision, pages 101–117. Springer,

  46. [46]

    Q-Align: Teaching LMMs for Visual Scoring via Discrete Text-Defined Levels

    Haoning Wu, Zicheng Zhang, Weixia Zhang, Chaofeng Chen, Liang Liao, Chunyi Li, Yixuan Gao, Annan Wang, Erli Zhang, Wenxiu Sun, et al. Q-align: Teaching lmms for vi- sual scoring via discrete text-defined levels.arXiv preprint arXiv:2312.17090, 2023. 4

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  47. [47]

    One-step effective diffusion network for real-world image super-resolution.Advances in Neural Information Processing Systems, 37:92529–92553, 2024

    Rongyuan Wu, Lingchen Sun, Zhiyuan Ma, and Lei Zhang. One-step effective diffusion network for real-world image super-resolution.Advances in Neural Information Processing Systems, 37:92529–92553, 2024. 5

    2024

  48. [48]

    Seesr: Towards semantics-aware real-world image super-resolution

    Rongyuan Wu, Tao Yang, Lingchen Sun, Zhengqiang Zhang, Shuai Li, and Lei Zhang. Seesr: Towards semantics-aware real-world image super-resolution. InProceedings of the IEEE/CVF conference on computer vision and pattern recog- nition, pages 25456–25467, 2024. 5

    2024

  49. [49]

    SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformers

    Enze Xie, Junsong Chen, Junyu Chen, Han Cai, Haotian Tang, Yujun Lin, Zhekai Zhang, Muyang Li, Ligeng Zhu, Yao Lu, et al. Sana: Efficient high-resolution image synthesis with lin- ear diffusion transformers.arXiv preprint arXiv:2410.10629,

    work page internal anchor Pith review Pith/arXiv arXiv

  50. [50]

    Fanjiang Ye, Zepeng Zhao, Yi Mu, Jucheng Shen, Renjie Li, Kaijian Wang, Saurabh Agarwal, Myungjin Lee, Triston Cao, Aditya Akella, Arvind Krishnamurthy, T. S. Eugene Ng, Zhengzhong Tu, and Yuke Wang. Supergen: An efficient ultra-high-resolution video generation system with sketching and tiling, 2025. 2

    2025

  51. [51]

    Agent banana: High-fidelity image editing with agentic thinking and tooling, 2026

    Ruijie Ye, Jiayi Zhang, Zhuoxin Liu, Zihao Zhu, Siyuan Yang, Li Li, Tianfu Fu, Franck Dernoncourt, Yue Zhao, Jiacheng Zhu, Ryan Rossi, Wenhao Chai, and Zhengzhong Tu. Agent banana: High-fidelity image editing with agentic thinking and tooling, 2026. 2

    2026

  52. [52]

    Sparkvsr: Interactive video super-resolution via sparse keyframe propa- gation, 2026

    Jiongze Yu, Xiangbo Gao, Pooja Verlani, Akshay Gadde, Yilin Wang, Balu Adsumilli, and Zhengzhong Tu. Sparkvsr: Interactive video super-resolution via sparse keyframe propa- gation, 2026. 2

    2026

  53. [53]

    Pku-aigiqa-4k: A perceptual qual- ity assessment database for both text-to-image and image-to- image ai-generated images

    Jiquan Yuan, Jihe Li, Fanyi Yang, Xinyan Cao, Jinming Che, Jinlong Lin, and Xixin Cao. Pku-aigiqa-4k: A perceptual qual- ity assessment database for both text-to-image and image-to- image ai-generated images. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 3331– 3340, 2025. 3

    2025

  54. [54]

    Restormer: Efficient transformer for high-resolution image restoration

    Syed Waqas Zamir, Aditya Arora, Salman Khan, Mu- nawar Hayat, Fahad Shahbaz Khan, and Ming-Hsuan Yang. Restormer: Efficient transformer for high-resolution image restoration. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 5728–5739,

  55. [55]

    Diffusion-4k: Ultra-high-resolution image synthesis with latent diffusion models

    Jinjin Zhang, Qiuyu Huang, Junjie Liu, Xiefan Guo, and Di Huang. Diffusion-4k: Ultra-high-resolution image synthesis with latent diffusion models. InProceedings of the Computer Vision and Pattern Recognition Conference, pages 23464– 23473, 2025. 2, 4

    2025

  56. [56]

    Beyond a gaussian denoiser: Residual learning of deep cnn for image denoising.IEEE transactions on image processing, 26(7):3142–3155, 2017

    Kai Zhang, Wangmeng Zuo, Yunjin Chen, Deyu Meng, and Lei Zhang. Beyond a gaussian denoiser: Residual learning of deep cnn for image denoising.IEEE transactions on image processing, 26(7):3142–3155, 2017. 2

    2017

  57. [57]

    Ffdnet: Toward a fast and flexible solution for cnn-based image denoising

    Kai Zhang, Wangmeng Zuo, and Lei Zhang. Ffdnet: Toward a fast and flexible solution for cnn-based image denoising. IEEE Transactions on Image Processing, 27(9):4608–4622,

  58. [58]

    Designing a practical degradation model for deep blind image super-resolution

    Kai Zhang, Jingyun Liang, Luc Van Gool, and Radu Timo- fte. Designing a practical degradation model for deep blind image super-resolution. InProceedings of the IEEE/CVF in- ternational conference on computer vision, pages 4791–4800,

  59. [59]

    The unreasonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shechtman, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 586–595, 2018. 5, 7

    2018

  60. [60]

    Wang, James Zou, Xiaoyu Wang, Ming-Hsuan Yang, and Zhengzhong Tu

    Yushen Zuo, Qi Zheng, Mingyang Wu, Xinrui Jiang, Ren- jie Li, Jian Wang, Yide Zhang, Gengchen Mai, Lihong V . Wang, James Zou, Xiaoyu Wang, Ming-Hsuan Yang, and Zhengzhong Tu. 4kagent: Agentic any image to 4k super- resolution, 2025. 2

    2025