Pith. sign in

REVIEW 3 major objections 6 minor 102 references

RAW-Diffusion: RGB-Guided Diffusion Models for High-Fidelity RAW Image Generation

T0 review · 3 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read RAW-Diffusion is a diffusion model that reconstructs high-fidelity RAW images from RGB inputs, reports state-of-the-art performance on all four tested DSLR datasets, and matches full-data performance with only 25 training images.

desk verdict Solid RGB2RAW engineering with a real data-efficiency claim that is currently undercut by a training-protocol confound, which the authors themselves acknowledge. read the letter →

arxiv 2411.13150 v1 pith:P2T457VH submitted 2024-11-20 cs.CV

Christoph Reinders , Radu Berdan , Beril Besbinar , Junji Otsuka , Daisuke Iso This is my paper · ORCID
classification cs.CV
keywords RAWimagegenerationRGB-to-RAWreconstructiondiffusionmodelsconditionaldataefficiencyobjectdetectionreverseISPsensor-specificdatasets
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

RAW-Diffusion sets out to solve a practical bottleneck: RAW sensor images carry more information than RGB, but collecting and storing RAW datasets for each camera is expensive. The paper argues that a denoising diffusion model conditioned on an RGB image can reconstruct the corresponding RAW sensor values with higher fidelity than existing inverse-ISP networks, and that this conditioning makes the model extremely data-efficient, with 25 training images matching the full training set. If true, large existing RGB datasets could be converted into camera-specific RAW datasets, reducing the need for costly sensor-specific data collection. The paper supports this with state-of-the-art PSNR and SSIM on four DSLR datasets and with object-detection experiments using generated BDD100K-RAW and Cityscapes-RAW.

What carries the argument

The load-bearing mechanism is a conditional denoising U-Net with two novel insertion points. An RGB-guidance module, an EDSR-style stack of residual blocks without an upsampling head, extracts a guidance feature map $F_{RGB}$ from the input image. These features enter the reverse diffusion process through RGB-guided residual blocks that apply a spatially adaptive modulation to the normalized diffusion features: $F^{i+1} = \mathrm{Norm}(F^i)\cdot(1+\gamma_i(\mathrm{Down}(F_{RGB})))+\beta_i(\mathrm{Down}(F_{RGB}))$, where $\gamma_i$ and $\beta_i$ are predicted per-channel, per-pixel modulation parameters. The block is used at the bottleneck and every decoder level, so the RGB context shapes all resolutions of the reconstruction. The model directly predicts the RAW image rather than the noise, uses a hyperbolic tangent output, and is trained with a sum of MSE, L1, and logarithmic L1 losses to handle the skewed distribution of RAW values.

What would settle it

Retrain the full-data model with the same iteration-based schedule used for the 25-image runs and compare PSNR on the four test sets; if the full model clearly surpasses the 25-image model, the sample-efficiency parity claim fails. Alternatively, train a detector exclusively on RAW-Diffusion-generated RAW for a camera whose real RAW pairs were held out and check whether its AP stays within a few points of a detector trained on real RAW from that camera.

Watch

Extended reading notes

Core claim

The central claim is that a diffusion model can invert the image signal processing pipeline: given an RGB image, RAW-Diffusion iteratively denoises random noise into a RAW image whose pixel values match the original sensor readings, and it does this better than prior learned and calibration-based methods. On the test sets, the method reports PSNR of 30.05 (FiveK Nikon), 34.01 (FiveK Canon), 44.93 (NOD Nikon), and 39.17 (NOD Sony), the best among the eleven compared methods. The paper also claims that the same model trained on only 25 images performs on par with full-data training, and that RAW images generated from BDD100K and Cityscapes improve Faster R-CNN and YOLOv8 detectors more than the same datasets converted by SRISP, including zero-shot settings where the detector never sees real NOD images.

Load-bearing premise

The 25-image sample-efficiency claim assumes that the iteration-based training schedule used for the small-data runs is comparable to the epoch-based schedule used for the full-data run, since the paper notes that the iteration-based schedule improves performance.

Editorial extensions

If this is right

  • On all four DSLR test sets, reconstructed RAW images are closer to the ground-truth sensor output than those of all eleven compared methods, with the largest margin on the low-light NOD Nikon set.
  • Training with 25 images, and in some cases 10, yields PSNR within the full-data range, so a usable sensor-specific RAW generator could in principle be built from a very small paired capture session.
  • Converting large RGB datasets such as BDD100K and Cityscapes into RAW with RAW-Diffusion improves object detection on NOD more than the same conversion with SRISP, both when mixed with real NOD images and in zero-shot training.
  • DDIM sampling with as few as 6 steps matches 1000-step DDPM reconstructions, which makes large-scale dataset generation computationally practical.
  • Ablation results show direct RAW prediction, the RGB-guided residual blocks, and the logarithmic L1 loss each contribute; predicting noise instead collapses PSNR to 5.55 on NOD Nikon.

Reading between the lines

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

  • If the data-efficiency result extends to other sensors, the method could turn any large annotated RGB dataset into a RAW training set for low-light or edge vision, bypassing the per-sensor collection bottleneck that limits RAW-based computer vision.
  • A multi-sensor variant that conditions on camera identity is a natural next step; the paper lists sensor generalization as future work.
  • A strict comparison of the few-shot and full-data training protocols is needed before '25 images match the full dataset' becomes a design rule, since the paper notes that the iteration-based few-shot schedule improves performance.
  • The downstream gains from generated RAW may partly reflect distributional realism rather than pixel accuracy; measuring detector performance against PSNR at fixed sample budgets would separate these factors.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper proposes RAW-Diffusion, a denoising-diffusion model that reconstructs RAW sensor images from RGB images. The method uses an EDSR-based RGB-guidance module and RGB-guided residual blocks (SPADE-style modulation) injected into a U-Net at multiple resolutions, trains with a sum of MSE, L1, and log-L1 losses to directly predict the denoised RAW image, and uses DDIM for fast sampling. Experiments on Four DSLR datasets (FiveK Nikon/Canon, NOD Nikon/Sony) report PSNR/SSIM gains over prior RGB2RAW methods, e.g., 30.05 dB on FiveK Nikon and 44.93 dB on NOD Nikon in Table 1. The paper further claims that training on as few as 25 images matches or exceeds full-data performance (Table 2), and shows that detectors trained with BDD100K-RAW generated by RAW-Diffusion outperform those trained with SRISP-generated RAW (Tables 3-5).

Significance. If the results hold, the paper makes a useful methodological contribution: it is, to my knowledge, the first diffusion-based RGB2RAW method to outperform inversion-based and ISP-simulation baselines on these benchmarks, and the downstream detection experiment gives a practical demonstration of generating sensor-specific RAW datasets from existing RGB corpora. The contribution is supported by careful ablations of prediction parameterization, conditioning architecture, losses, backbones, and sampling steps, and the code is released. The main caveat concerns the data-efficiency claim, which is not yet supported by a controlled comparison; with that resolved, the paper would be a solid addition.

major comments (3)
  1. [§4.4, Table 2] The data-efficiency claim is confounded by the training protocol. The full model is trained with an epoch-based schedule for 70k steps, while all subsampled models are trained with an iteration-based schedule in which the small datasets are repeated; the authors explicitly state that the iteration-based schedule 'improves performance.' Table 2 shows that 25-image training on FiveK Nikon achieves 31.29 dB PSNR versus 30.05 dB for full training, and 250-image training on FiveK Canon achieves 35.93 versus 34.01, so the protocol advantage is large enough to mask the effect of sample size. The claim that 'RAW-Diffusion achieves the same performance with 25 or fewer training samples as when training on the full training set' requires a comparison in which the full model is retrained with the same iteration-based schedule (or the few-shot models are trained epoch-based). Without that control, the parity is evidence about the training schedule, not about sample efficiency. Because this is a headline contribution, I consider this load-bearing.
  2. [§4.4] The NOD few-shot experiments do not specify how the subsampled training sets of 500, 250, 100, 50, 25, and 10 images are drawn from the 3.2k/4.0k NOD training images. For FiveK the split is fixed and follows [89], but for NOD no random-seed or selection procedure is reported, so subset composition is uncontrolled; a favorable subset could partly explain the flat or improving PSNR as training set size decreases. Please specify the selection protocol and, ideally, report results over multiple subset draws.
  3. [§4.2–§4.3, Table 1] Several baselines in Table 1 are evaluated under modified settings that are not applied uniformly. InvISP and ISPLess are retrained as 'InvISP+' and 'ISPLess+' with JPEG simulation removed and ground-truth RGB input; RISPNet is downscaled and run without ensembling; the diffusion baseline is trained on 64x64 patches; and SRISP uses the mean global feature as test-time reference. Some modifications are reasonable (e.g., avoiding JPEG artifacts in evaluation), and the reported margins are large, but the headline 'state-of-the-art' comparison is not fully uniform. I ask the authors to state clearly, for each baseline, which configuration is used and whether any modifications were needed to adapt it to the evaluation protocol, and to consider reporting results for the original configurations as well.
minor comments (6)
  1. [Table 14] The NOD Sony, AP_S column reports '.8±0.3'; the leading zero appears to be missing and should read '0.8±0.3'.
  2. [§4.5] In the third paragraph, 'RSISP [63]' should be 'SRISP [63]'.
  3. [§4.4] The sentence 'When increasing the number of training runs, the results better align with the expectations' is unclear; please rephrase to say that results become more consistent as the number of random seeds increases.
  4. [Table 2] Table 2 reports no standard deviations, while Table 1 does; adding standard deviations or confidence intervals would help readers assess whether the small-data differences are significant.
  5. [§4.4] The text says 'To maintain a consistent number of training steps, the small training sets are repeated accordingly.' Please clarify whether the total number of optimizer steps is exactly 70k for all subsampled runs, and whether the learning-rate schedule is reset per run or aligned across runs.
  6. [§4.2] For the diffusion baseline, the statement 'Diffusion is trained on 64×64 patches as described by the authors. The performance with larger patch sizes collapses' would benefit from a brief explanation, since the patch size may disadvantage the baseline relative to the 256×256 patches used for the proposed method.

Circularity Check

0 steps flagged · score 0.0 of 10

No load-bearing circularity: RGB2RAW metrics are held-out reconstructions, downstream AP is measured on real NOD test data, and the Section 4.4 training-protocol difference is an experimental-control confound rather than a circular derivation.

full rationale

I walked the paper's claimed derivation chain and found no step in which a prediction reduces by construction to a fitted input, a self-citation, or a renamed known result. The central RGB2RAW claims (Table 1) are evaluated with PSNR/SSIM on held-out test splits of FiveK and NOD, and the detection claims (Tables 3-5, 12-15) are measured as AP on the real NOD test set; no fitted constant is repackaged as a prediction. The comparison baseline SRISP [63] shares an author with this paper, but it is used as an externally measured baseline that RAW-Diffusion outperforms, not as a load-bearing justification for the method's correctness. The only concern raised by the text is in Section 4.4, where the authors state that 'training on the full dataset is performed epoch-based while training on the subsampled datasets is iteration-based which improves performance'; this is a genuine training-protocol confound for the data-efficiency headline, because the full model was not retrained with the iteration-based schedule. However, a protocol confound is not circularity under the stated criteria: the few-shot results are still empirical measurements on held-out data, and no equation or fitted parameter makes the comparison equivalent to its inputs by construction. Accordingly, no circular step is flagged, and the paper is self-contained against external benchmarks for its main SOTA and downstream claims.

Assumptions & free parameters 6 free parameters · 3 assumptions · 0 invented entities

The central claim depends on standard diffusion-model assumptions, the learnability of the sensor-specific RGB-to-RAW mapping, and the usefulness of generated RAW data for downstream detection. No new physical entities or invented mediators are introduced.

free parameters (6)
  • Loss weights (LMSE, LL1, LlogL1) = 1, 1, 1
    Chosen by hand as equal weights; no tuning reported.
  • Diffusion noise schedule (beta1, betaT) = 0.0001, 0.02
    Standard linear schedule from prior work, not fitted to data.
  • Number of training steps = 70k
    Set by the authors; repeated for small-data runs.
  • DDIM sampling steps for generation = 24 or 6
    Used for object-detection dataset generation; shown to be robust.
  • Mixing ratio pgen = 0.95
    Probability of sampling from the generated dataset when combining with NOD images in detection training.
  • Patch size = 256x256
    Random crops; baseline Diffusion trained at 64x64, a comparison confound.
assumptions (3)
  • standard math DDPM forward and reverse process equations (Eq. 1 and 2) are correct and standard.
    The method builds on Ho et al. [32]; no new derivation is required.
  • domain assumption The RGB input processed by RawPy is an accurate representation of the camera ISP output.
    Section 4.1: paired data are generated by processing RAW with RawPy to obtain RGB; the mapping is assumed learnable.
  • domain assumption Generated RAW images are realistic enough to improve downstream detection when added to real data.
    Empirically tested in Section 4.5, but assumed as a premise for the dataset-generation contribution.

how reviews work

0 comments
Cite this review

Pith. "Pith review of RAW-Diffusion: RGB-Guided Diffusion Models for High-Fidelity RAW Image Generation." pith.science (2026). https://pith.science/paper/P2T457VH

@misc{pith2026241113150,
  author       = {Pith},
  title        = {Pith review of: RAW-Diffusion: RGB-Guided Diffusion Models for High-Fidelity RAW Image Generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/P2T457VH}},
  note         = {Machine review of arXiv:2411.13150}
}
read the original abstract

Current deep learning approaches in computer vision primarily focus on RGB data sacrificing information. In contrast, RAW images offer richer representation, which is crucial for precise recognition, particularly in challenging conditions like low-light environments. The resultant demand for comprehensive RAW image datasets contrasts with the labor-intensive process of creating specific datasets for individual sensors. To address this, we propose a novel diffusion-based method for generating RAW images guided by RGB images. Our approach integrates an RGB-guidance module for feature extraction from RGB inputs, then incorporates these features into the reverse diffusion process with RGB-guided residual blocks across various resolutions. This approach yields high-fidelity RAW images, enabling the creation of camera-specific RAW datasets. Our RGB2RAW experiments on four DSLR datasets demonstrate state-of-the-art performance. Moreover, RAW-Diffusion demonstrates exceptional data efficiency, achieving remarkable performance with as few as 25 training samples or even fewer. We extend our method to create BDD100K-RAW and Cityscapes-RAW datasets, revealing its effectiveness for object detection in RAW imagery, significantly reducing the amount of required RAW images.

Figures

Figures reproduced from arXiv: 2411.13150 by the authors.

Figure 1
Figure 1. RAW-Diffusion enables the generation of high-fidelity [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The RAW-Diffusion architecture consists of an RGB [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Qualitative results on FiveK (top) and NOD (bottom). [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Qualitative results on NOD Nikon (top) and Sony (bot [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: The results of training exclusively on the original images, i.e., pgen = 0, are highlighted by a dashed line [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Qualitative results on FiveK Canon (top), NOD Nikon (middle), and NOD Sony (bottom). The reconstructed RAW image and [PITH_FULL_IMAGE:figures/full_fig_p015_6.png]
Figure 7
Figure 7. Figure 7: Qualitative object detection results from various models on the test set of NOD Nikon (first and second row) and Sony (third and [PITH_FULL_IMAGE:figures/full_fig_p016_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

102 extracted references · 71 canonical work pages

  1. [52]

    Springer International Publishing, 2016

    2016

  2. [89]

    Smartbrush: Text and shape guided ob- ject inpainting with diffusion model

    Shaoan Xie, Zhifei Zhang, Zhe Lin, Tobias Hinz, and Kun Zhang. Smartbrush: Text and shape guided ob- ject inpainting with diffusion model. In Proceed- ings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition , pages 22428–22437, 2023

    work page 2023

  3. [1]

    A high-quality denoising dataset for smartphone cameras

    Abdelrahman Abdelhamed, Stephen Lin, and Michael S Brown. A high-quality denoising dataset for smartphone cameras. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 1692–1700, 2018

    2018

  4. [2]

    Cie xyz net: Unprocessing images for low-level computer vision tasks

    Mahmoud Afifi, Abdelrahman Abdelhamed, Abdul- lah Abuolaim, Abhijith Punnappurath, and Michael S Brown. Cie xyz net: Unprocessing images for low-level computer vision tasks. IEEE Transac- tions on Pattern Analysis and Machine Intelligence , 44(9):4688–4700, 2021

    2021

  5. [3]

    Un- processing images for learned raw denoising

    Tim Brooks, Ben Mildenhall, Tianfan Xue, Jiawen Chen, Dillon Sharlet, and Jonathan T Barron. Un- processing images for learned raw denoising. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11036–11045, 2019

    2019

  6. [4]

    Tim Brooks, Ben Mildenhall, Tianfan Xue, Jiawen Chen, Dillon Sharlet, and Jonathan T. Barron. Un- processing Images for Learned Raw Denoising. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition, pages 11036– 11045, 2019

    2019

  7. [5]

    Reconfiguring the imaging pipeline for com- puter vision

    Mark Buckler, Suren Jayasuriya, and Adrian Samp- son. Reconfiguring the imaging pipeline for com- puter vision. In Proceedings of the IEEE Interna- tional Conference on Computer Vision , pages 975– 984, 2017

    2017

  8. [6]

    Learning photographic global tonal adjustment with a database of input / output image pairs

    Vladimir Bychkovsky, Sylvain Paris, Eric Chan, and Fr´edo Durand. Learning photographic global tonal adjustment with a database of input / output image pairs. In The Twenty-Fourth IEEE Conference on Computer Vision and Pattern Recognition, 2011

    2011

Show all 102 references
  1. [7]

    Training Vi- sion Transformers with Only 2040 Images

    Yun-Hao Cao, Hao Yu, and Jianxin Wu. Training Vi- sion Transformers with Only 2040 Images. In Com- puter Vision - ECCV 2022, 2022

    2022

  2. [8]

    End-to-end object detection with trans- formers

    Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey Zagoruyko. End-to-end object detection with trans- formers. In Andrea Vedaldi, Horst Bischof, Thomas Brox, and Jan-Michael Frahm, editors, Computer Vi- sion – ECCV 2020 , pages 2...

    2020

  3. [9]

    An empirical camera model for internet color vision

    Ayan Chakrabarti, Daniel Scharstein, and Todd E Zickler. An empirical camera model for internet color vision. In British Machine Vision Conference, 2009

    2009

  4. [10]

    Modeling radiometric uncertainty for vi- sion with tone-mapped color images

    Ayan Chakrabarti, Ying Xiong, Baochen Sun, Trevor Darrell, Daniel Scharstein, Todd Zickler, and Kate Saenko. Modeling radiometric uncertainty for vi- sion with tone-mapped color images. IEEE Transac- tions on Pattern Analysis and Machine Intelligence , 36(11):2185–2198, 2014

    2014

  5. [11]

    Zhang, and Jian Sun

    Liangyu Chen, Xiaojie Chu, X. Zhang, and Jian Sun. Simple baselines for image restoration. In European Conference on Computer Vision, 2022

    2022

  6. [12]

    Model-based image signal processors via learnable dictionaries

    Marcos V Conde, Steven McDonagh, Matteo Mag- gioni, Ales Leonardis, and Eduardo P ´erez-Pellitero. Model-based image signal processors via learnable dictionaries. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 481–489, 2022

    2022

  7. [13]

    Reversed image signal processing and RAW reconstruction

    Marcos V Conde, Radu Timofte, Yibin Huang, Jingyang Peng, Chang Chen, Cheng Li, Eduardo P´erez-Pellitero, Fenglong Song, Furui Bai, Shuai Liu, and others. Reversed image signal processing and RAW reconstruction. AIM 2022 challenge re- port. In Proceedings of the European Confe...

    2022

  8. [14]

    The cityscapes dataset for semantic urban scene un- derstanding

    Marius Cordts, Mohamed Omran, Sebastian Ramos, Timo Rehfeld, Markus Enzweiler, Rodrigo Benen- son, Uwe Franke, Stefan Roth, and Bernt Schiele. The cityscapes dataset for semantic urban scene un- derstanding. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (C...

    2016

  9. [15]

    Diffusion models in vi- sion: A survey

    Florinel-Alin Croitoru, Vlad Hondru, Radu Tudor Ionescu, and Mubarak Shah. Diffusion models in vi- sion: A survey. IEEE Transactions on Pattern Anal- ysis and Machine Intelligence, 2023

    2023

  10. [16]

    Diffuseraw: End-to-end generative raw image processing for low-light images

    Rishit Dagli. Diffuseraw: End-to-end generative raw image processing for low-light images. arXiv preprint arXiv:2402.18575, 2023

    2023 arXiv

  11. [17]

    Jaiswal, and Peter A

    Gourav Datta, Zeyu Liu, Zihan Yin, Linyu Sun, Akhilesh R. Jaiswal, and Peter A. Beerel. En- abling ISPless Low-Power Computer Vision. In2023 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pages 2429–2438. IEEE, 2023

    2023

  12. [18]

    Recovering high dynamic range radiance maps from photographs

    Paul E Debevec and Jitendra Malik. Recovering high dynamic range radiance maps from photographs. In Seminal Graphics Papers: Pushing the Boundaries, Volume 2, pages 643–652. Association for Comput- ing Machinery, 2023

    2023

  13. [19]

    Diffusion models beat gans on image synthesis

    Prafulla Dhariwal and Alexander Nichol. Diffusion models beat gans on image synthesis. Advances in neural information processing systems , 34:8780– 8794, 2021

    2021

  14. [20]

    Diffusion models beat GANs on image synthesis

    Prafulla Dhariwal and Alexander Nichol. Diffusion models beat GANs on image synthesis. In M. Ran- zato, A. Beygelzimer, Y . Dauphin, P.S. Liang, and 9 J. Wortman Vaughan, editors, Advances in Neural Information Processing Systems , volume 34, pages 8780–8794. Curran Associates...

    2021

  15. [21]

    RISPNet: A network for Re- versed image signal processing

    Xiaoyi Dong, Yu Zhu, Chenghua Li, Peisong Wang, and Jian Cheng. RISPNet: A network for Re- versed image signal processing. In Computer Vision – ECCV 2022 Workshops, pages 445–457, 2023

    2022

  16. [22]

    An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, and Neil Houlsby. An Image is Worth 16x16 Words: Transformers for Image Recognition at...

    2021

  17. [23]

    HDR image recon- struction from a single exposure using deep CNNs

    Eilertsen, Gabriel, Joel Kronander, Gyorgy Denes, Rafał Mantiuk, and Jonas Unger. HDR image recon- struction from a single exposure using deep CNNs. ACM Transactions on Graphics (TOG), 36(6), 2017

    2017

  18. [24]

    Sigmoid-Weighted Linear Units for Neural Network Function Approximation in Reinforcement Learning

    Stefan Elfwing, Eiji Uchibe, and Kenji Doya. Sigmoid-Weighted Linear Units for Neural Network Function Approximation in Reinforcement Learning. Neural Networks, 107:3–11, 2018

    2018

  19. [25]

    Data augmentation for object detection via controllable diffusion mod- els

    Haoyang Fang, Boran Han, Shuai Zhang, Su Zhou, Cuixiong Hu, and Wen-Ming Ye. Data augmentation for object detection via controllable diffusion mod- els. In Proceedings of the IEEE/CVF Winter Con- ference on Applications of Computer Vision , pages 1257–1266, 2024

    2024

  20. [26]

    Deep joint demosaicking and denoising

    Micha ¨el Gharbi, Gaurav Chaurasia, Sylvain Paris, and Fr ´edo Durand. Deep joint demosaicking and denoising. ACM Transactions on Graphics (ToG) , 35(6):1–12, 2016

    2016

  21. [27]

    Rich feature hierarchies for accurate object detection and semantic segmentation

    Ross Girshick, Jeff Donahue, Trevor Darrell, and Ji- tendra Malik. Rich feature hierarchies for accurate object detection and semantic segmentation. In Pro- ceedings of the IEEE conference on computer vision and pattern recognition, pages 580–587, 2014

    2014

  22. [28]

    Determin- ing the camera response from images: What is know- able? IEEE Transactions on pattern analysis and machine intelligence, 25(11):1455–1467, 2003

    Michael D Grossberg and Shree K Nayar. Determin- ing the camera response from images: What is know- able? IEEE Transactions on pattern analysis and machine intelligence, 25(11):1455–1467, 2003

    2003

  23. [29]

    Masked Au- toencoders Are Scalable Vision Learners

    Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Doll ´ar, and Ross Girshick. Masked Au- toencoders Are Scalable Vision Learners. In 2022 IEEE/CVF Conference on Computer Vision and Pat- tern Recognition (CVPR), 2022

    2022

  24. [30]

    Deep Residual Learning for Image Recog- nition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep Residual Learning for Image Recog- nition. In 2016 IEEE Conference on Computer Vi- sion and Pattern Recognition (CVPR) , pages 770– 778, June 2016

    2016

  25. [31]

    Gaussian Error Linear Units (GELUs)

    Dan Hendrycks and Kevin Gimpel. Gaussian Error Linear Units (GELUs). arXiv preprint arXiv:1606.08415, 2016

    2016 arXiv

  26. [32]

    Denois- ing diffusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denois- ing diffusion probabilistic models. In H. Larochelle, M. Ranzato, R. Hadsell, M.F. Balcan, and H. Lin, editors, Advances in Neural Information Processing Systems, volume 33, pages 6840–6851. Curran As- sociates, Inc., 2020

    2020

  27. [33]

    Cascaded diffusion models for high fidelity image generation

    Jonathan Ho, Chitwan Saharia, William Chan, David J Fleet, Mohammad Norouzi, and Tim Sali- mans. Cascaded diffusion models for high fidelity image generation. The Journal of Machine Learning Research, 23(1):2249–2281, 2022

    2022

  28. [34]

    Classifier-free diffu- sion guidance

    Jonathan Ho and Tim Salimans. Classifier-free diffu- sion guidance. In NeurIPS 2021 Workshop on Deep Generative Models and Downstream Applications , 2021

    2021

  29. [35]

    Crafting object detection in very low light

    Yang Hong, Kaixuan Wei, Linwei Chen, and Ying Fu. Crafting object detection in very low light. In British Machine Vision Conference (BMVC), 2021

    2021

  30. [36]

    Ultra- lytics YOLOv8

    Glenn Jocher, Ayush Chaurasia, and Jing Qiu. Ultra- lytics YOLOv8. 2023

    2023

  31. [37]

    Diffusion models in medical imaging: A comprehensive sur- vey

    Amirhossein Kazerouni, Ehsan Khodapanah Agh- dam, Moein Heidari, Reza Azad, Mohsen Fayyaz, Ilker Hacihaliloglu, and Dorit Merhof. Diffusion models in medical imaging: A comprehensive sur- vey. Medical Image Analysis, 88, 2023

    2023

  32. [38]

    A new in-camera imaging model for color com- puter vision and its application

    Seon Joo Kim, Hai Ting Lin, Zheng Lu, Sabine S¨usstrunk, Stephen Lin, and Michael S Brown. A new in-camera imaging model for color com- puter vision and its application. IEEE Transac- tions on Pattern Analysis and Machine Intelligence , 34(12):2289–2302, 2012

    2012

  33. [39]

    DiffWave: A versatile dif- fusion model for audio synthesis

    Zhifeng Kong, Wei Ping, Jiaji Huang, Kexin Zhao, and Bryan Catanzaro. DiffWave: A versatile dif- fusion model for audio synthesis. In International Conference on Learning Representations, 2021

    2021

  34. [40]

    Alex Krizhevsky, Ilya Sutskever, and Geoffrey E. Hinton. ImageNet classification with deep convolu- tional neural networks. In Advances in Neural Infor- mation Processing Systems, 2012

    2012

  35. [41]

    Ro- bust reflection removal with flash-only cues in the wild

    Chenyang Lei, Xudong Jiang, and Qifeng Chen. Ro- bust reflection removal with flash-only cues in the wild. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023

    2023

  36. [42]

    Salman Asif, and Zhan Ma

    Zhihao Li, Ming Lu, Xu Zhang, Xin Feng, M. Salman Asif, and Zhan Ma. Efficient visual com- 10 puting with camera RAW snapshots. IEEE Transac- tions on Pattern Analysis and Machine Intelligence , pages 1–18, 2024

    2024

  37. [43]

    Cameranet: A two-stage framework for ef- fective camera isp learning

    Zhetong Liang, Jianrui Cai, Zisheng Cao, and Lei Zhang. Cameranet: A two-stage framework for ef- fective camera isp learning. IEEE Transactions on Image Processing, 30:2248–2262, 2021

    2021

  38. [44]

    Enhanced deep resid- ual networks for single image super-resolution

    Bee Lim, Sanghyun Son, Heewon Kim, Seungjun Nah, and Kyoung Mu Lee. Enhanced deep resid- ual networks for single image super-resolution. In The IEEE Conference on Computer Vision and Pat- tern Recognition (CVPR) Workshops, 2017

    2017

  39. [45]

    Determining the ra- diometric response function from a single grayscale image

    Stephen Lin and Lei Zhang. Determining the ra- diometric response function from a single grayscale image. In 2005 IEEE Computer Society Confer- ence on Computer Vision and Pattern Recognition (CVPR’05), volume 2, pages 66–73. IEEE, 2005

    2005

  40. [46]

    Feature pyramid networks for object detection

    Tsung-Yi Lin, Piotr Doll ´ar, Ross Girshick, Kaiming He, Bharath Hariharan, and Serge Belongie. Feature pyramid networks for object detection. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 2117–2125, 2017

    2017

  41. [47]

    Focal loss for dense ob- ject detection

    Tsung-Yi Lin, Priya Goyal, Ross Girshick, Kaim- ing He, and Piotr Doll ´ar. Focal loss for dense ob- ject detection. In Proceedings of the IEEE interna- tional conference on computer vision , pages 2980– 2988, 2017

    2017

  42. [48]

    Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Doll´ar, and C

    Tsung-Yi Lin, Michael Maire, Serge J. Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Doll´ar, and C. Lawrence Zitnick. Microsoft COCO: Common objects in context. InEuropean Conference on Computer Vision, 2014

    2014

  43. [49]

    Deep-flexisp: A three-stage framework for night photography ren- dering

    Shuai Liu, Chaoyu Feng, Xiaotao Wang, Hao Wang, Ran Zhu, Yongqiang Li, and Lei Lei. Deep-flexisp: A three-stage framework for night photography ren- dering. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 1211–1220, 2022

    2022

  44. [50]

    Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection

    Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Qing Jiang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, and Lei Zhang. Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection. In Computer Vision – ECCV 2024, 2024

    2024

  45. [51]

    Wei Liu, Dragomir Anguelov, Dumitru Erhan, Chris- tian Szegedy, Scott Reed, Cheng-Yang Fu, and Alexander C. Berg. SSD: Single shot MultiBox de- tector. In Computer Vision – ECCV 2016, pages 21–

    2016

  46. [53]

    Invert- ing Image Signal Processing Pipeline with Diffusion Models

    Xinman Liu, Xuanchi Ren, and Ziyi Wu. Invert- ing Image Signal Processing Pipeline with Diffusion Models. Technical report, University of Toronto, 2022

    2022

  47. [54]

    Raw or cooked? object detection on raw images

    William Ljungbergh, Joakim Johnander, Christoffer Petersson, and Michael Felsberg. Raw or cooked? object detection on raw images. In Scandina- vian Conference on Image Analysis, pages 374–385. Springer, 2023

    2023

  48. [55]

    Decoupled weight decay regularization

    Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization. In International Conference on Learning Representations, 2019

    2019

  49. [56]

    Repaint: Inpainting using denoising diffusion probabilistic models

    Andreas Lugmayr, Martin Danelljan, Andres Romero, Fisher Yu, Radu Timofte, and Luc Van Gool. Repaint: Inpainting using denoising diffusion probabilistic models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11461–11471, 2022

    2022

  50. [57]

    Understanding diffusion models: A uni- fied perspective

    Calvin Luo. Understanding diffusion models: A uni- fied perspective. arXiv preprint arXiv:2208.11970 , 2022

    2022 arXiv

  51. [58]

    Radiomet- ric self calibration

    Tomoo Mitsunaga and Shree K Nayar. Radiomet- ric self calibration. In Proceedings. 1999 IEEE com- puter society conference on computer vision and pat- tern recognition (Cat. No PR00149), volume 1, pages 374–380. IEEE, 1999

    1999

  52. [59]

    GenISP: Neu- ral ISP for low-light machine cognition

    Igor Morawski, Yu-An Chen, Yu-Sheng Lin, Shusil Dangi, Kai He, and Winston H Hsu. GenISP: Neu- ral ISP for low-light machine cognition. In Proceed- ings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 630–639, 2022

    2022

  53. [60]

    T2I-Adapter: Learning Adapters to Dig out More Controllable Ability for Text-to-Image Diffu- sion Models

    Chong Mou, Xintao Wang, Liangbin Xie, Yanze Wu, Jian Zhang, Zhongang Qi, Ying Shan, and Xiaohu Qie. T2I-Adapter: Learning Adapters to Dig out More Controllable Ability for Text-to-Image Diffu- sion Models. In AAAI Conference on Artificial Intel- ligence (AAAI), 2024

    2024

  54. [61]

    Im- proved denoising diffusion probabilistic models

    Alexander Quinn Nichol and Prafulla Dhariwal. Im- proved denoising diffusion probabilistic models. In Marina Meila and Tong Zhang, editors, Proceed- ings of the 38th International Conference on Ma- chine Learning, volume 139 of Proceedings of Ma- chine Learning Research, pages...

    2021

  55. [62]

    GLIDE: Towards photorealistic image generation and edit- ing with text-guided diffusion models

    Alexander Quinn Nichol, Prafulla Dhariwal, Aditya Ramesh, Pranav Shyam, Pamela Mishkin, Bob Mc- Grew, Ilya Sutskever, and Mark Chen. GLIDE: Towards photorealistic image generation and edit- ing with text-guided diffusion models. In Kama- lika Chaudhuri, Stefanie Jegelka, Le So...

    2022

  56. [63]

    Maxime Oquab, Timoth ´ee Darcet, Th ´eo Moutakanni, Huy V . V o, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel HAZIZA, Francisco Massa, Alaaeldin El-Nouby, Mido As- sran, Nicolas Ballas, Wojciech Galuba, Russell Howes, Po-Yao Huang, Shang-Wen Li, Ishan Misra, Micha...

    2024

  57. [64]

    Self-Supervised Reversed Image Signal Pro- cessing via Reference-Guided Dynamic Parameter Selection

    Junji Otsuka, Masakazu Yoshimura, and Takeshi Ohashi. Self-Supervised Reversed Image Signal Pro- cessing via Reference-Guided Dynamic Parameter Selection. arXiv preprint arXiv:2303.13916, 2023

    2023 arXiv

  58. [65]

    Semantic Image Synthesis with Spatially-Adaptive Normalization

    Taesung Park, Ming-Yu Liu, Ting-Chun Wang, and Jun-Yan Zhu. Semantic Image Synthesis with Spatially-Adaptive Normalization. In 2019 IEEE/CVF Conference on Computer Vision and Pat- tern Recognition (CVPR), pages 2332–2341, 2019

    2019

  59. [66]

    Prajit Ramachandran, Barret Zoph, and Quoc V . Le. Searching for Activation Functions. In 6th Inter- national Conference on Learning Representations, ICLR 2018, Workshop Track Proceedings, 2018

    2018

  60. [67]

    Hierarchical text- conditional image generation with clip latents

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

    2022 arXiv

  61. [68]

    You only look once: Unified, real- time object detection

    Joseph Redmon, Santosh Divvala, Ross Girshick, and Ali Farhadi. You only look once: Unified, real- time object detection. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 779–788, 2016

    2016

  62. [69]

    Object recognition from very few training examples for enhancing bicycle maps

    Christoph Reinders, Hanno Ackermann, Michael Ying Yang, and Bodo Rosenhahn. Object recognition from very few training examples for enhancing bicycle maps. In 2018 IEEE Intelligent Vehicles Symposium (IV), 2018

    2018

  63. [70]

    Learning convolutional neural networks for object detection with very little training data

    Christoph Reinders, Hanno Ackermann, Michael Ying Yang, and Bodo Rosenhahn. Learning convolutional neural networks for object detection with very little training data. Multimodal Scene Understanding, 2019

    2019

  64. [71]

    ChimeraMix: Image classification on small datasets via masked feature mixing

    Christoph Reinders, Frederik Schubert, and Bodo Rosenhahn. ChimeraMix: Image classification on small datasets via masked feature mixing. In Pro- ceedings of the Thirty-First International Joint Con- ference on Artificial Intelligence (IJCAI), 2022

    2022

  65. [72]

    Two worlds in one network: Fusing deep learning and random forests for classification and ob- ject detection

    Christoph Reinders, Michael Ying Yang, and Bodo Rosenhahn. Two worlds in one network: Fusing deep learning and random forests for classification and ob- ject detection. Volunteered Geographic Information, 2024

    2024

  66. [73]

    Faster R-CNN: Towards real-time object de- tection with region proposal networks

    Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun. Faster R-CNN: Towards real-time object de- tection with region proposal networks. In C. Cortes, N. Lawrence, D. Lee, M. Sugiyama, and R. Garnett, editors, Advances in Neural Information Processing Systems, volume 28. Curra...

    2015

  67. [74]

    High- resolution image synthesis with latent diffusion mod- els

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High- resolution image synthesis with latent diffusion mod- els. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 10684–10695, 2022

    2022

  68. [75]

    Deepisp: Toward learning an end-to-end image pro- cessing pipeline

    Eli Schwartz, Raja Giryes, and Alex M Bronstein. Deepisp: Toward learning an end-to-end image pro- cessing pipeline. IEEE Transactions on Image Pro- cessing, 28(2):912–923, 2018

    2018

  69. [76]

    Overfeat: Integrated recognition, localization and de- tection using convolutional networks

    Pierre Sermanet, David Eigen, Xiang Zhang, Micha¨el Mathieu, Rob Fergus, and Yann LeCun. Overfeat: Integrated recognition, localization and de- tection using convolutional networks. arXiv preprint arXiv:1312.6229, 2013

    2013 arXiv

  70. [77]

    A benchmark dataset and evaluation for non-lambertian and uncalibrated photometric stereo

    Boxin Shi, Zhe Wu, Zhipeng Mo, Dinglong Duan, Sai-Kit Yeung, and Ping Tan. A benchmark dataset and evaluation for non-lambertian and uncalibrated photometric stereo. In Proceedings of the IEEE Con- ference on Computer Vision and Pattern Recognition, pages 3707–3716, 2016

    2016

  71. [78]

    Deep unsuper- vised learning using nonequilibrium thermodynam- ics

    Jascha Sohl-Dickstein, Eric Weiss, Niru Mah- eswaranathan, and Surya Ganguli. Deep unsuper- vised learning using nonequilibrium thermodynam- ics. In Francis Bach and David Blei, editors, Pro- ceedings of the 32nd International Conference on Machine Learning, volume 37 ofProcee...

    2015

  72. [79]

    Denoising diffusion implicit models

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denoising diffusion implicit models. In Inter- national Conference on Learning Representations , 2021

    2021

  73. [80]

    Ef- ficientdet: Scalable and efficient object detection

    Mingxing Tan, Ruoming Pang, and Quoc V Le. Ef- ficientdet: Scalable and efficient object detection. In Proceedings of the IEEE/CVF conference on com- puter vision and pattern recognition , pages 10781– 10790, 2020

    2020

  74. [81]

    Fcos: Fully convolutional one-stage object detection

    Zhi Tian, Chunhua Shen, Hao Chen, and Tong He. Fcos: Fully convolutional one-stage object detection. In Proceedings of the IEEE/CVF international con- ference on computer vision, pages 9627–9636, 2019. 12

    2019

  75. [82]

    Effective data augmen- tation with diffusion models

    Brandon Trabucco, Kyle Doherty, Max A Gurinas, and Ruslan Salakhutdinov. Effective data augmen- tation with diffusion models. In The Twelfth Inter- national Conference on Learning Representations , 2024

    2024

  76. [83]

    Exploiting diffusion prior for real-world image super-resolution

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

    2024

  77. [84]

    Semantic Image Synthesis via Diffusion Models

    Weilun Wang, Jianmin Bao, Wengang Zhou, Dong- dong Chen, Dong Chen, Lu Yuan, and Houqiang Li. Semantic Image Synthesis via Diffusion Models. arXiv preprint arXiv:2207.00050, 2022

    2022

  78. [85]

    ESRGAN: Enhanced super-resolution gener- ative adversarial networks

    Xintao Wang, Ke Yu, Shixiang Wu, Jinjin Gu, Yi- hao Liu, Chao Dong, Yu Qiao, and Chen Change Loy. ESRGAN: Enhanced super-resolution gener- ative adversarial networks. In Laura Leal-Taix ´e and Stefan Roth, editors, The European Conference on Computer Vision Workshops (ECCVW), 2019

    2019

  79. [86]

    Bovik, H.R

    Zhou Wang, A.C. Bovik, H.R. Sheikh, and E.P. Si- moncelli. Image quality assessment: From error vis- ibility to structural similarity. IEEE Transactions on Image Processing, 13(4):600–612, 2004

    2004

  80. [87]

    Group normalization

    Yuxin Wu and Kaiming He. Group normalization. International Journal of Computer Vision, 128:742– 755, 2018

    2018

  81. [88]

    In- vertible grayscale

    Menghan Xia, Xueting Liu, and Tien-Tsin Wong. In- vertible grayscale. ACM Transactions on Graphics (SIGGRAPH Asia 2018 issue) , 37(6):246:1–246:10, Nov. 2018

    2018

  82. [90]

    Invert- ible Image Signal Processing

    Yazhou Xing, Zian Qian, and Qifeng Chen. Invert- ible Image Signal Processing. 2021 IEEE/CVF Con- ference on Computer Vision and Pattern Recognition (CVPR), pages 6283–6292, 2021

    2021

  83. [91]

    GeoDiff: A geometric diffusion model for molecular conformation genera- tion

    Minkai Xu, Lantao Yu, Yang Song, Chence Shi, Ste- fano Ermon, and Jian Tang. GeoDiff: A geometric diffusion model for molecular conformation genera- tion. In International Conference on Learning Rep- resentations, 2022

    2022

  84. [92]

    Dynamicisp: dynamically con- trolled image signal processor for image recognition

    Masakazu Yoshimura, Junji Otsuka, Atsushi Irie, and Takeshi Ohashi. Dynamicisp: dynamically con- trolled image signal processor for image recognition. In Proceedings of the IEEE/CVF International Con- ference on Computer Vision , pages 12866–12876, 2023

    2023

  85. [93]

    Rawgment: noise-accounted raw augmentation enables recognition in a wide variety of environments

    Masakazu Yoshimura, Junji Otsuka, Atsushi Irie, and Takeshi Ohashi. Rawgment: noise-accounted raw augmentation enables recognition in a wide variety of environments. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recog- nition, pages 14007–14017, 2023

    2023

  86. [94]

    BDD100K: A diverse driving dataset for heterogeneous multitask learning

    Fisher Yu, Haofeng Chen, Xin Wang, Wenqi Xian, Yingying Chen, Fangchen Liu, Vashisht Madhavan, and Trevor Darrell. BDD100K: A diverse driving dataset for heterogeneous multitask learning. 2020 IEEE/CVF Conference on Computer Vision and Pat- tern Recognition (CVPR), pages 2633–...

    2020

  87. [95]

    Small object detection via coarse-to-fine proposal generation and imitation learning

    Xiang Yuan, Gong Cheng, Kebing Yan, Qinghua Zeng, and Junwei Han. Small object detection via coarse-to-fine proposal generation and imitation learning. In Proceedings of the IEEE/CVF Interna- tional Conference on Computer Vision (ICCV), pages 6317–6327, 2023

    2023

  88. [96]

    Cycleisp: Real image restora- tion via improved data synthesis

    Syed Waqas Zamir, Aditya Arora, Salman Khan, Munawar Hayat, Fahad Shahbaz Khan, Ming-Hsuan Yang, and Ling Shao. Cycleisp: Real image restora- tion via improved data synthesis. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 2696–2705, 2020

    2020

  89. [97]

    Dauphin, and David Lopez-Paz

    Hongyi Zhang, Moustapha Ciss ´e, Yann N. Dauphin, and David Lopez-Paz. Mixup: Beyond Empirical Risk Minimization. In International Conference on Learning Representations, 2018

    2018

  90. [98]

    Zhang, A

    L. Zhang, A. Rao, and M. Agrawala. Adding condi- tional control to text-to-image diffusion models. In 2023 IEEE/CVF International Conference on Com- puter Vision (ICCV), pages 3813–3824, Los Alami- tos, CA, USA, Oct. 2023. IEEE Computer Society

    2023

  91. [99]

    Zoom to learn, learn to zoom

    Xuaner Zhang, Qifeng Chen, Ren Ng, and Vladlen Koltun. Zoom to learn, learn to zoom. In Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 3762–3770, 2019

    2019

  92. [100]

    Progressive end-to-end ob- ject detection in crowded scenes

    Anlin Zheng, Yuang Zhang, Xiangyu Zhang, Xiao- juan Qi, and Jian Sun. Progressive end-to-end ob- ject detection in crowded scenes. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages 857–866, June 2022

    2022

  93. [101]

    Object detection in 20 years: A survey

    Zhengxia Zou, Keyan Chen, Zhenwei Shi, Yuhong Guo, and Jieping Ye. Object detection in 20 years: A survey. Proceedings of the IEEE, 2023. 13 RA W-Diffusion: RGB-Guided Diffusion Models for High-Fidelity RA W Image Generation (Supplementary Material) In the supplementary materi...

    2023

  94. [102]

    and YOLOv8 [36]. Faster R-CNN has a ResNet-50 backbone pretrained on ImageNet, and it is trained with RGB and RAW images normalized using the correspond- ing mean and standard deviation of the dataset. We apply random flip, random resize, and cropping as data augmenta- tion, u...

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.