Pith. sign in

REVIEW 5 major objections 5 minor 50 references

DiffuPT: Class Imbalance Mitigation for Glaucoma Detection via Diffusion Based Generation and Model Pretraining

T0 review · 5 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A diffusion-based pretraining pipeline lifts glaucoma screening harmonic mean from 89.09% to 92.59% on a new Egyptian dataset.

desk verdict Useful dataset and thorough generative model comparison, but the central DiffuPT claim is confounded by a 2.7x training budget gap; worth serious review with compute-matched baselines. read the letter →

arxiv 2412.03629 v1 pith:5PJPUOWX submitted 2024-12-04 eess.IV cs.CVq-bio.QM

Youssof Nawar , Nouran Soliman , Moustafa Wassel , Mohamed ElHabebe , Noha Adly , Marwan Torki , Ahmed Elmassry , Islam Ahmed This is my paper · ORCID
classification eess.IVcs.CVq-bio.QM
keywords glaucomadetectionclassimbalancediffusionmodelslatentsyntheticdatagenerationpretrainingfundusimagesSwinTransformer
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

The paper proposes DiffuPT, a training scheme that targets glaucoma's class imbalance by generating balanced synthetic fundus images with a class-conditioned latent diffusion model, filtering them with a baseline classifier, and using them to pretrain a Swin Transformer before fine-tuning on real data. The authors collected GlaucomaEgy, an Egyptian dataset of more than 37,000 gradable fundus images in which roughly 10 percent are glaucomatous. On that dataset's test set, they report the harmonic mean of sensitivity and specificity rising from 89.09% with normal training to 92.59% with DiffuPT, with AUC reaching 98.02%. They report similar improvements on the AIROGS dataset, arguing that balanced synthetic pretraining stabilizes representations and fine-tuning then recovers discriminative accuracy.

What carries the argument

The key machinery is the DiffuPT pipeline: a class-conditioned Latent Diffusion Model (LDM) trained on both glaucoma and non-glaucoma fundus images in a compressed latent space, with a learnable class embedder providing conditioning; a DDIM sampler that generates new images; a baseline Swin Transformer classifier that filters the generated images, keeping those classified as the target class; and a two-stage schedule that pretrains a Swin Transformer on the balanced synthetic set, then fine-tunes on the real data. The LDM autoencoder is trained with LPIPS loss, and the pretraining distribution is chosen by ablation, with the 50-50 glaucoma/non-glaucoma split performing best. Filtering, the paper argues, removes noise from the generated set and improves the final harmonic mean from 91.5% to 92.59%.

What would settle it

A direct test: pretrain the same Swin Transformer on a balanced set of real glaucoma images (e.g., oversampled or additionally collected) with identical compute, learning rate, and iteration counts. If real-balanced pretraining matches or exceeds DiffuPT's 92.59% harmonic mean, then the diffusion-generated content is not the active ingredient. A second check is to measure the filter's rejection rate on generated glaucoma images and inspect what it rejects; if it systematically removes subtle or early-stage cases, the mechanism is suspect.

Watch

Extended reading notes

Core claim

The central claim is that pretraining a classifier on a balanced, diffusion-generated, baseline-filtered synthetic dataset and then fine-tuning on the real imbalanced dataset yields a more sensitive and specific glaucoma detector than standard training, weighted cross-entropy, weighted sampling, generation-based augmentation, or multi-stage decoupling. The paper explains the gain through embedding geometry: balanced synthetic pretraining lowers variance but increases overlap between classes, and fine-tuning on real data reduces that overlap, giving the best of both. The strongest reported result is a harmonic mean of 92.59% and AUC of 98.02% on the GlaucomaEgy test set, with a 1.09-point improvement attributable to the filtering stage alone.

Load-bearing premise

The method assumes that synthetic glaucoma images that survive the baseline classifier's filter are an informative stand-in for real glaucoma images during pretraining; if the filter silently drops the hard, informative cases, the pretraining set could reinforce the baseline's blind spots rather than correct them.

Editorial extensions

If this is right

  • If DiffuPT works as reported, class imbalance in glaucoma screening can be mitigated without collecting more real minority-class images or adding manual annotations.
  • The filtering stage shows that a baseline classifier can act as a quality gate for synthetic medical images, and the observed 1.09-point harmonic-mean gain indicates that which generated samples survive matters.
  • The comparison across generative models indicates that diffusion-based generation outperforms the tested GAN-based image-to-image methods in both generation quality (FID, KID) and downstream classification, making the generative model choice consequential.
  • The method's success on both GlaucomaEgy and AIROGS suggests the pretraining approach transfers across datasets and across ethnic variation in fundus appearance.
  • The ablation of pretraining distributions shows that a balanced 50-50 synthetic set gives the best AUC before fine-tuning, implying class balance in the pretraining data is important for the downstream gain.

Reading between the lines

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

  • I suspect part of the gain is a generic regularization effect: any large, balanced, slightly noisy pretraining set may stabilize features, so a controlled comparison against real-image oversampling is needed to separate synthetic-content effects from balance effects.
  • The filter's threshold is a hidden hyperparameter; varying it likely trades precision against recall in the pretraining set, and the paper's single fixed filter leaves this trade-off unexplored.
  • The same pipeline could apply to other imbalanced medical imaging tasks, but the main risk is circularity: if the baseline classifier's false negatives are exactly the subtle cases the generator is asked to produce, the filter may systematically remove the most informative synthetic examples.
  • If the generated images were released, they could serve as a privacy-preserving proxy for rare disease classes, though the paper does not test this.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 5 minor

Summary. The paper proposes DiffuPT, a training scheme for glaucoma classification that uses a latent diffusion model trained on imbalanced real fundus images to generate a balanced synthetic dataset, filters the generated samples with the baseline classifier, pretrains a Swin Transformer on the filtered synthetic data, and then fine-tunes on the real, imbalanced data. The authors introduce GlaucomaEgy, a large Egyptian fundus dataset (33,633 gradable images), and report that DiffuPT improves the harmonic mean from 89.09% to 92.59% on its test set, with similar gains on the AIROGS dataset. Extensive comparisons are presented against GAN-based and diffusion-based generation methods and against standard class-imbalance techniques.

Significance. Class imbalance is an important problem in medical image classification, and the idea of using diffusion-generated data for pretraining is timely and potentially useful. The main contribution is a large national dataset and a clearly described empirical framework. However, the current evidence is weakened by a training-budget confound (DiffuPT receives 160k iterations versus 60k for the baseline), a filtering step whose effect is unexplained, anomalous AUC values for two baseline methods, and the absence of repeated-run statistics. If the authors provide compute-matched baselines, a mechanistic analysis of the filter, and confidence intervals, the approach would represent a meaningful, practical contribution. The paper does not include code or reproducibility artifacts, and the single-run results should not be taken as definitive.

major comments (5)
  1. [§5.1 vs §5.3; Tables 7, 8] The central comparison is confounded by unequal training budgets. Section 5.1 states that baseline classifiers are trained for 60k iterations, while Section 5.3 states that DiffuPT uses 100k pretraining iterations followed by 60k fine-tuning iterations, for a total of 160k optimizer steps. The tables that report the headline results (Tables 7 and 8) do not specify the iteration counts of the other imbalance-mitigation methods, but these methods are described in the context of the baseline setup and thus likely also receive 60k steps. If so, DiffuPT receives 2.67x more training than every comparator, and the reported gains could be explained by additional training alone. The filtering ablation in Table 9 does not resolve this because both arms use the same 160k budget. To support the claim that diffusion-based balanced pretraining is beneficial, the authors must add a compute-matched control, e.g., training the baseline and each comparator for 160k iterations, or pretraining for 100k iterations on randomly shuffled real data and then fine-tuning for 60k.
  2. [§4.2; Table 9] The filtering stage is load-bearing but unexplained. Section 4.2 states that generated images are "filtered using our already trained baseline classifier" to determine whether they belong to the desired class, and Table 9 shows that filtering improves the harmonic mean from 91.5 to 92.59. However, the manuscript never reports the filtering threshold, the fraction of generated samples removed, or what types of samples are removed. If the filter prunes hard, informative positives, the pretraining set becomes easy and reinforces the baseline's blind spots, which would invalidate the proposed mechanism. At minimum, the authors should report the filtering statistics, and ideally compare filtering by the baseline against filtering by an oracle or by another classifier to establish that the filter removes noise rather than informative positives.
  3. [Tables 7 and 8] The AUC values reported for the weighted-CE baselines are anomalous and undermine the comparison. In Table 7, Weighted CE achieves AUC 90.27 and Weighted CE + Sampler achieves 92.37, while every other method (including normal training) achieves AUC above 97. The same pattern appears in Table 8, with AUC 89.17 and 90.09 for those two baselines. Since class weighting changes the learned model and should not collapse AUC to this degree, these results suggest an implementation or evaluation issue (for example, a learning-rate mismatch or a bug in how predictions are aggregated). The authors must explain these anomalies or re-run the experiments; otherwise the comparisons against those baselines are not reliable.
  4. [§5.3; Tables 7-10] No statistical significance or repeated runs are reported. All tables show a single value per metric, and the margins between DiffuPT and the best comparator are small on the test set (92.59 vs 92.17 harmonic mean for the augmentation method, 98.02 vs 97.40 AUC for the weighted sampler). Given the stochasticity of deep-network training, these differences may not be reproducible. The authors should run at least three to five independent seeds for each condition and report means and standard deviations or confidence intervals, particularly for the DiffuPT versus augmentation comparison that is closest.
  5. [§5.3] The pretraining checkpoint selection protocol may introduce another confound. Section 5.3 states: "We used the same validation set to determine the best initial weights." This means the validation set is used to select among pretrained checkpoints before fine-tuning, a selection step that the compared methods may not have. If the comparators are not given the same validation-based selection, the comparison is not fair. The authors should either apply the same selection protocol to all methods or show that the result is robust to removing this selection step.
minor comments (5)
  1. [Figure 1] Figure 1 presents embedding analyses for 'normal training', 'pretraining', and 'fine-tuned' models, but the figure has no axis labels, no description of which layer's embeddings are visualized, and no details on the dimensionality-reduction technique; please add these details so the figure is interpretable.
  2. [§5.2.2 and Table 5] The text says LDM is chosen because it had the best 'generation scores', however DDPM has the best Inception Score (188.39 vs 153.74) while LDM has better FID and KID; the choice of LDM should be justified explicitly in light of this trade-off.
  3. [§4.1] The autoencoder used in the LDM was trained 'on a collection of AIROGS and GlaucomaEgy datasets', but the manuscript does not state whether this training used the same train/validation/test splits or whether the test sets of either dataset were included; if the autoencoder saw test images, this would be a data-leakage risk that must be clarified.
  4. [Table 7 header] The header 'GlacomaEgy Validation' contains a typo; it should read 'GlaucomaEgy'.
  5. [§5.4 and Table 10] The AIROGS experiments do not specify how many generated samples were used for pretraining, whether the baseline classifier filtering was applied, or what the class distribution of the pretraining set was; please report these details so the setup is reproducible.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the pipeline is empirical and held-out test evaluations carry the claims; the baseline-filter self-reference and minor same-author citations do not force any reported result.

full rationale

The paper makes no first-principles derivation whose output is, by construction, equal to an input. The central claim is an empirical comparison on held-out test sets (Tables 7, 8, and 10), with the generative model, filter, pretraining, and fine-tuning all trained on train/validation data and evaluated on test data that is not used for fitting. The one self-referential element is Section 4.2, where generated samples are filtered by the already trained baseline classifier; however, this filter is an ablation variable evaluated in Table 9, not a fitted parameter later reported as a prediction, and the improved test harmonic mean is not derivable from the filter definition alone. Minor same-author citations such as [47] for transformer-based glaucoma classification and [8,9] for prior retinal work exist, but they are background and not load-bearing for the DiffuPT result. The unequal training budget concern (60k baseline steps versus 100k pretraining plus 60k fine-tuning for DiffuPT) is a legitimate experimental confound and correctness risk, but it is not circularity: it does not make any reported metric equivalent to an input by construction.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The paper's claim is empirical and method-oriented, so the main free parameters are hand-chosen design choices: guidance scale, pretraining distribution, synthetic set size, filtering threshold, and augmentation peak. The core domain assumptions are synthetic data fidelity and label correctness. No new physical or theoretical entities are introduced.

free parameters (5)
  • Guidance scale w = 3.0
    Used for class-conditional sampling (Appendix A, Eq. 5); chosen by the authors to control generation quality with no reported sensitivity analysis.
  • Pretraining class distribution = 50-50 glaucoma/non-glaucoma
    Selected from five candidate distributions (30-70 up to 70-30) based on validation AUC and harmonic mean (Table 6); this choice directly affects the final DiffuPT result.
  • Number of generated pretraining samples = 96,307 total, 43,550 glaucoma
    The generation budget and balance used in pretraining; the authors do not report a sensitivity analysis over the total synthetic set size.
  • Filtering criterion = Not specified (baseline classifier score threshold)
    Generated samples are kept only when the baseline classifier assigns the desired class, but no threshold or confidence value is given (Section 4.2, Figure 3).
  • Augmentation sample count peak = 2,000-3,000 synthetic glaucoma samples
    The best augmentation result was found by scanning 1,000 to 5,000 samples (Section 5.2.4); this peak value is used for the augmentation baseline in the main comparison.
assumptions (4)
  • domain assumption The latent diffusion autoencoder trained on AIROGS and GlaucomaEgy with LPIPS loss reconstructs fundus images faithfully (SSIM = 92.53).
    Used to justify the latent space as perceptually equivalent (Section 4.1); no independent verification that reconstruction quality translates into downstream classifier benefit.
  • domain assumption Adjudicated labels of GlaucomaEgy and AIROGS are correct ground truth.
    All metrics rely on these labels; grading mismatch percentages are high for grades 1 and 2 (Table 2), so label noise is a real but unquantified factor.
  • ad hoc to paper The baseline classifier's filtering removes noise rather than informative hard positives.
    Filtering improves the harmonic mean from 91.5 to 92.59 (Table 9), but the mechanism, whether it removes noise or merely easy-to-classify samples, is not tested.
  • ad hoc to paper The observed improvement is caused by the pretraining scheme, not by the extra compute or the validation-based configuration selection.
    DiffuPT uses 100k pretraining iterations plus 60k fine-tuning, while baselines use 60k; compute is not matched, so part of the gain may be due to longer training.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DiffuPT: Class Imbalance Mitigation for Glaucoma Detection via Diffusion Based Generation and Model Pretraining." pith.science (2026). https://pith.science/paper/5PJPUOWX

@misc{pith2026241203629,
  author       = {Pith},
  title        = {Pith review of: DiffuPT: Class Imbalance Mitigation for Glaucoma Detection via Diffusion Based Generation and Model Pretraining},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5PJPUOWX}},
  note         = {Machine review of arXiv:2412.03629}
}
read the original abstract

Glaucoma is a progressive optic neuropathy characterized by structural damage to the optic nerve head and functional changes in the visual field. Detecting glaucoma early is crucial to preventing loss of eyesight. However, medical datasets often suffer from class imbalances, making detection more difficult for deep-learning algorithms. We use a generative-based framework to enhance glaucoma diagnosis, specifically addressing class imbalance through synthetic data generation. In addition, we collected the largest national dataset for glaucoma detection to support our study. The imbalance between normal and glaucomatous cases leads to performance degradation of classifier models. By combining our proposed framework leveraging diffusion models with a pretraining approach, we created a more robust classifier training process. This training process results in a better-performing classifier. The proposed approach shows promising results in improving the harmonic mean (sensitivity and specificity) and AUC for the roc for the glaucoma classifier. We report an improvement in the harmonic mean metric from 89.09% to 92.59% on the test set of our national dataset. We examine our method against other methods to overcome imbalance through extensive experiments. We report similar improvements on the AIROGS dataset. This study highlights that diffusion-based generation can be of great importance in tackling class imbalances in medical datasets to improve diagnostic performance.

Figures

Figures reproduced from arXiv: 2412.03629 by the authors.

Figure 1
Figure 1. Analysis of the embeddings on different training tech [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Samples of the collected dataset. The two images on the left are from grade 0 (non-glaucomatous) samples. Both middle images [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Framework for generating data using LDM. We show how to sample a generated dataset using LDM. Our method involves a [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Training procedure used for obtaining the final classifier. [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Samples from different generation methods. The left images show nonglaucoma samples from the dataset. Then we show the [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Results of augmentation using generated samples (Train [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 8
Figure 8. Figure 8: The top images are reconstruction images. The bottom [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 7
Figure 7. Figure 7: The top images are from conditional LDM and the bot [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 9
Figure 9. Figure 9: More Generated Samples [PITH_FULL_IMAGE:figures/full_fig_p012_9.png]
Figure 10
Figure 10. Figure 10: Confusion Matrices showing the right eyes separately from the left eyes. [PITH_FULL_IMAGE:figures/full_fig_p012_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 28 canonical work pages

  1. [1]

    Demystifying mmd gans

    Mikołaj Bi ´nkowski, Danica J Sutherland, Michael Arbel, and Arthur Gretton. Demystifying mmd gans. arXiv preprint arXiv:1801.01401, 2018. 7

    arXiv 2018

  2. [2]

    Multi-marginal wasserstein gan.Ad- vances in Neural Information Processing Systems, 32, 2019

    Jiezhang Cao, Langyuan Mo, Yifan Zhang, Kui Jia, Chunhua Shen, and Mingkui Tan. Multi-marginal wasserstein gan.Ad- vances in Neural Information Processing Systems, 32, 2019. 3, 6

    work page 2019

  3. [3]

    Smote: synthetic minority over- sampling technique

    Nitesh V Chawla, Kevin W Bowyer, Lawrence O Hall, and W Philip Kegelmeyer. Smote: synthetic minority over- sampling technique. Journal of artificial intelligence re- search, 16:321–357, 2002. 3

    work page 2002

  4. [4]

    Stargan v2: Diverse image synthesis for multiple domains

    Yunjey Choi, Youngjung Uh, Jaejun Yoo, and Jung-Woo Ha. Stargan v2: Diverse image synthesis for multiple domains. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 8188–8197, 2020. 3, 6

    work page 2020

  5. [5]

    Airogs: Artificial intelligence for robust glaucoma screening challenge

    Coen De Vente, Koenraad A Vermeer, Nicolas Jaccard, He Wang, Hongyi Sun, Firas Khader, Daniel Truhn, Temirgali Aimyshev, Yerkebulan Zhanibekuly, Tien-Dung Le, et al. Airogs: Artificial intelligence for robust glaucoma screening challenge. IEEE Transactions on Medical Imaging, 2023. 2, 4

    work page 2023

  6. [6]

    Explain- able framework for glaucoma diagnosis by image process- ing and convolutional neural network synergy: analysis with doctor evaluation

    Omer Deperlioglu, Utku Kose, Deepak Gupta, Ashish Khanna, Fabio Giampaolo, and Giancarlo Fortino. Explain- able framework for glaucoma diagnosis by image process- ing and convolutional neural network synergy: analysis with doctor evaluation. Future Generation Computer Systems , 129:152–169, 2022. 2

    work page 2022

  7. [7]

    An image is worth 16x16 words: Trans- formers for image recognition at scale

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, et al. An image is worth 16x16 words: Trans- formers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020. 2, 6

    arXiv 2010

  8. [8]

    Dr10k: Transfer learning using weak labels for grading diabetic retinopathy on dr10k dataset

    Mohamed ElHabebe, Shereen ElKordi, Ahmed Gamal El- Din, Noha Adly, Marwan Torki, Ahmed Elmassry, and Is- lam SH Ahmed. Dr10k: Transfer learning using weak labels for grading diabetic retinopathy on dr10k dataset. In Pro- ceedings of the IEEE/CVF Winter Conference on Applica- tions of Computer Vision, pages 7748–7758, 2024. 2

    work page 2024

Show all 50 references
  1. [9]

    Prevalence of diabetic retinopathy in patients with diabetes in alexandria and north-west delta, egypt

    Ahmed Elmassry, Islam SH Ahmed, Noha Adly, and Mar- wan Torki. Prevalence of diabetic retinopathy in patients with diabetes in alexandria and north-west delta, egypt. In- ternational Ophthalmology, 43(8):2883–2895, 2023. 3

    2023

  2. [10]

    Generative adversarial nets

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial nets. Advances in neural information processing systems, 27, 2014. 1, 2

    2014

  3. [11]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016. 6

    2016

  4. [12]

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

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

    2017

  5. [13]

    Denoising dif- fusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising dif- fusion probabilistic models. Advances in neural information processing systems, 33:6840–6851, 2020. 1, 3

    2020

  6. [14]

    Classifier-free diffusion guidance

    Jonathan Ho and Tim Salimans. Classifier-free diffusion guidance. arXiv preprint arXiv:2207.12598, 2022. 3, 6

    2022 arXiv

  7. [15]

    What do patients with glaucoma see? visual symptoms re- ported by patients with glaucoma

    Cindy X Hu, Camila Zangalli, Michael Hsieh, Lalita Gupta, Alice L Williams, Jesse Richman, and George L Spaeth. What do patients with glaucoma see? visual symptoms re- ported by patients with glaucoma. The American journal of the medical sciences, 348(5):403–409, 2014. 1

    2014

  8. [16]

    Ethnic diversity of retinal images used to train artificial intelligence models improves diagnostic accuracy to detect diabetic retinopathy

    Cris Martin P Jacoba, Duy Doan, Dean Zhang, Ward Fick- weiler, Catherine Jamison, Ahmed Souka, Monsef Khar- boush, Frank Albert, Kaye Locaylocay, Moises Dumapig, et al. Ethnic diversity of retinal images used to train artificial intelligence models improves diagnostic accuracy...

    2023

  9. [17]

    Decou- pling representation and classifier for long-tailed recogni- tion

    Bingyi Kang, Saining Xie, Marcus Rohrbach, Zhicheng Yan, Albert Gordo, Jiashi Feng, and Yannis Kalantidis. Decou- pling representation and classifier for long-tailed recogni- tion. arXiv preprint arXiv:1910.09217, 2019. 1, 3

    1910 arXiv

  10. [18]

    Unpaired image-to-image translation via neu- ral schr\” odinger bridge

    Beomsu Kim, Gihyun Kwon, Kwanyoung Kim, and Jong Chul Ye. Unpaired image-to-image translation via neu- ral schr\” odinger bridge. arXiv preprint arXiv:2305.15086,

    arXiv

  11. [19]

    Learning from imbalanced data: open challenges and future directions

    Bartosz Krawczyk. Learning from imbalanced data: open challenges and future directions. Progress in Artificial Intel- ligence, 5(4):221–232, 2016. 1, 3

    2016

  12. [20]

    A deep learning approach to automatic detection of early glau- coma from visual fields

    S ¸erife Seda Kucur, G´abor Holl ´o, and Raphael Sznitman. A deep learning approach to automatic detection of early glau- coma from visual fields. PloS one, 13(11):e0206081, 2018. 2

    2018

  13. [21]

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

    Christian Ledig, Lucas Theis, Ferenc Husz´ar, 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. In Proceedings of the IE...

  14. [22]

    Attention based glaucoma detection: A large-scale database and cnn model

    Liu Li, Mai Xu, Xiaofei Wang, Lai Jiang, and Hanruo Liu. Attention based glaucoma detection: A large-scale database and cnn model. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 10571– 10580, 2019. 2, 4

    2019

  15. [23]

    Racial differences in reti- nal vessel geometric characteristics: a multiethnic study in healthy asians

    Xiang Li, Wan Ling Wong, Carol Yim-lui Cheung, Ching-Yu Cheng, Mohammad Kamran Ikram, Jialiang Li, Kee Seng Chia, and Tien Yin Wong. Racial differences in reti- nal vessel geometric characteristics: a multiethnic study in healthy asians. Investigative ophthalmology & visual sc...

    2013

  16. [24]

    Efficacy of a deep learning system for detecting glaucomatous optic neuropathy based on color fun- dus photographs

    Zhixi Li, Yifan He, Stuart Keel, Wei Meng, Robert T Chang, and Mingguang He. Efficacy of a deep learning system for detecting glaucomatous optic neuropathy based on color fun- dus photographs. Ophthalmology, 125(8):1199–1206, 2018. 2

    2018

  17. [25]

    Focal loss for dense object detection

    Tsung-Yi Lin, Priya Goyal, Ross Girshick, Kaiming He, and Piotr Doll´ar. Focal loss for dense object detection. In Pro- ceedings of the IEEE international conference on computer vision, pages 2980–2988, 2017. 3

    2017

  18. [26]

    Swin trans- former: Hierarchical vision transformer using shifted win- dows

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin trans- former: Hierarchical vision transformer using shifted win- dows. arXiv preprint arXiv:2103.14030, 2021. 6

    2021 arXiv

  19. [27]

    Exploring the limits of weakly supervised pretraining

    Dhruv Mahajan, Ross Girshick, Vignesh Ramanathan, Kaiming He, Manohar Paluri, Yixuan Li, Ashwin Bharambe, and Laurens Van Der Maaten. Exploring the limits of weakly supervised pretraining. In Proceedings of the European con- ference on computer vision (ECCV) , pages 181–196, 2018. 3

    2018

  20. [28]

    Refuge challenge: A unified framework for evaluat- ing automated methods for glaucoma assessment from fun- dus photographs

    Jos ´e Ignacio Orlando, Huazhu Fu, Jo ˜ao Barbosa Breda, Karel Van Keer, Deepti R Bathula, Andr ´es Diaz-Pinto, Ruogu Fang, Pheng-Ann Heng, Jeyoung Kim, JoonHo Lee, et al. Refuge challenge: A unified framework for evaluat- ing automated methods for glaucoma assessment from fun...

    2020

  21. [29]

    The effectiveness of data aug- mentation in image classification using deep learning

    Luis Perez and Jason Wang. The effectiveness of data aug- mentation in image classification using deep learning. arXiv preprint arXiv:1712.04621, 2017. 1

    2017 arXiv

  22. [30]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022. 1, 5

    2022

  23. [31]

    U- net: Convolutional networks for biomedical image segmen- tation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U- net: Convolutional networks for biomedical image segmen- tation. In Medical Image Computing and Computer-Assisted Intervention–MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Pa...

    2015

  24. [32]

    Improved techniques for training gans

    Tim Salimans, Ian Goodfellow, Wojciech Zaremba, Vicki Cheung, Alec Radford, and Xi Chen. Improved techniques for training gans. Advances in neural information processing systems, 29, 2016. 7

    2016

  25. [33]

    Simplifying neural network training under class imbalance

    Ravid Shwartz-Ziv, Micah Goldblum, Yucen Li, C Bayan Bruss, and Andrew G Wilson. Simplifying neural network training under class imbalance. Advances in Neural Infor- mation Processing Systems, 36, 2024. 6

    2024

  26. [34]

    Gans for medical image synthesis: An empirical study.Jour- nal of Imaging, 9(3):69, 2023

    Youssef Skandarani, Pierre-Marc Jodoin, and Alain Lalande. Gans for medical image synthesis: An empirical study.Jour- nal of Imaging, 9(3):69, 2023. 1

    2023

  27. [35]

    Learning structured output representation using deep conditional gen- erative models

    Kihyuk Sohn, Honglak Lee, and Xinchen Yan. Learning structured output representation using deep conditional gen- erative models. Advances in neural information processing systems, 28, 2015. 11

    2015

  28. [36]

    Denoising diffusion implicit models

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denoising diffusion implicit models. arXiv preprint arXiv:2010.02502, 2020. 5

    2010 arXiv

  29. [37]

    Score-based generative modeling through stochastic differential equa- tions

    Yang Song, Jascha Sohl-Dickstein, Diederik P Kingma, Ab- hishek Kumar, Stefano Ermon, and Ben Poole. Score-based generative modeling through stochastic differential equa- tions. arXiv preprint arXiv:2011.13456, 2020. 3

    2011 arXiv

  30. [38]

    Cost-sensitive boosting for classification of imbalanced data

    Yanmin Sun, Mohamed S Kamel, Andrew KC Wong, and Yang Wang. Cost-sensitive boosting for classification of imbalanced data. Pattern recognition, 40(12):3358–3378,

  31. [39]

    Efficientnet: Rethinking model scaling for convolutional neural networks

    Mingxing Tan and Quoc Le. Efficientnet: Rethinking model scaling for convolutional neural networks. In International conference on machine learning, pages 6105–6114. PMLR,

  32. [40]

    Strategies for improving early detection of glau- coma: the combined structure–function index

    Andrew J Tatham, Robert N Weinreb, and Felipe A Medeiros. Strategies for improving early detection of glau- coma: the combined structure–function index. Clinical oph- thalmology, pages 611–621, 2014. 1

    2014

  33. [41]

    Cost-sensitive learning methods for imbalanced data

    Nguyen Thai-Nghe, Zeno Gantner, and Lars Schmidt- Thieme. Cost-sensitive learning methods for imbalanced data. In The 2010 International joint conference on neural networks (IJCNN), pages 1–8. IEEE, 2010. 1

    2010

  34. [42]

    Global prevalence of glau- coma and projections of glaucoma burden through 2040: a systematic review and meta-analysis

    Yih-Chung Tham, Xiang Li, Tien Y Wong, Harry A Quigley, Tin Aung, and Ching-Yu Cheng. Global prevalence of glau- coma and projections of glaucoma burden through 2040: a systematic review and meta-analysis. Ophthalmology, 121(11):2081–2090, 2014. 1

    2014

  35. [43]

    Racial varia- tions in the prevalence of primary open-angle glaucoma: the baltimore eye survey

    James M Tielsch, Alfred Sommer, Joanne Katz, Richard M Royall, Harry A Quigley, and Jonathan Javitt. Racial varia- tions in the prevalence of primary open-angle glaucoma: the baltimore eye survey. Jama, 266(3):369–374, 1991. 2

    1991

  36. [44]

    Going deeper with im- age transformers

    Hugo Touvron, Matthieu Cord, Alexandre Sablayrolles, Gabriel Synnaeve, and Herv´e J´egou. Going deeper with im- age transformers. arXiv preprint arXiv:2103.17239 , 2021. 6

    2021 arXiv

  37. [45]

    Experimental perspectives on learning from imbal- anced data

    Jason Van Hulse, Taghi M Khoshgoftaar, and Amri Napoli- tano. Experimental perspectives on learning from imbal- anced data. In Proceedings of the 24th international con- ference on Machine learning, pages 935–942, 2007. 1, 3

    2007

  38. [46]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszko- reit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. Advances in neural information processing systems, 30, 2017. 2

    2017

  39. [47]

    Vision transformers based classification for glau- comatous eye condition

    Moustafa Wassel, Ahmed M Hamdi, Noha Adly, and Mar- wan Torki. Vision transformers based classification for glau- comatous eye condition. In 2022 26th International Con- ference on Pattern Recognition (ICPR) , pages 5082–5088. IEEE, 2022. 2, 4

    2022

  40. [48]

    The unreasonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shecht- man, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 586–595, 2018. 5

    2018

  41. [49]

    Weakly-supervised simultaneous evidence identification and segmentation for automated glaucoma di- agnosis

    Rongchang Zhao, Wangmin Liao, Beiji Zou, Zailiang Chen, and Shuo Li. Weakly-supervised simultaneous evidence identification and segmentation for automated glaucoma di- agnosis. In Proceedings of the AAAI conference on artificial intelligence, volume 33, pages 809–816, 2019. 2

    2019

  42. [50]

    Unpaired image-to-image translation using cycle- consistent adversarial networks

    Jun-Yan Zhu, Taesung Park, Phillip Isola, and Alexei A Efros. Unpaired image-to-image translation using cycle- consistent adversarial networks. In Proceedings of the IEEE international conference on computer vision , pages 2223– 2232, 2017. 2, 6 A. Conditional Vs Unconditional...

    2017

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.