Pith. sign in

REVIEW 3 major objections 7 minor 65 references

DGSSA: Domain generalization with structural and stylistic augmentation for retinal vessel segmentation

T0 review · 3 major / 7 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Training on procedurally grown vessel structures plus style mixing raises cross-dataset retinal vessel segmentation Dice to 77.98%.

desk verdict Interesting structural-augmentation idea for retinal vessel DG, but a protocol gap could invalidate the headline result and the SOTA edge is within noise. read the letter →

arxiv 2501.03466 v2 pith:EGEJUJUK submitted 2025-01-07 eess.IV cs.CV

Bo Liu , Yudong Zhang , Shuihua Wang , Siyue Li , Jin Hong This is my paper · ORCID
classification eess.IVcs.CV
keywords domaingeneralizationretinalvesselsegmentationstructuralaugmentationspacecolonizationalgorithmstylePixMixmedicalimagefundusimages
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

This paper argues that a retinal vessel segmentation model generalizes to datasets it has never seen better when its training data are made more diverse in both image style and vessel structure, not just style. To that end it introduces DGSSA, which procedurally generates artificial vascular trees with a space-colonization algorithm, renders them into pseudo-retinal images with an improved Pix2Pix generator, and then further diversifies all training images with PixMix-style photometric mixing and uncertainty perturbation. In leave-one-out tests across four fundus datasets, the paper reports an average Dice similarity coefficient of 77.98%, about 0.38 percentage points above the style-only method AADG. If the result holds, procedural structure generation offers a cheap way to make segmentation models robust to new scanners and patient populations without collecting target-domain data.

What carries the argument

The load-bearing mechanism is the pairing of a procedural structure generator with a conditional image synthesizer. The space-colonization algorithm grows branching vascular networks by iteratively placing nodes toward attractor points, then applies Murray's-law-style thickening and post-processing to produce realistic vessel masks; an improved Pix2Pix model with a multi-scale discriminator converts these masks, together with real paired masks, into pseudo-retinal images. Style augmentation via PixMix with uncertainty perturbation then mixes each image with augmented variants and externally sourced images under random photometric operations. Together they provide the segmentation network with both structural and stylistic diversity during training, so that the evaluation protocol is the only part of the pipeline that sees the target domain.

What would settle it

Look at the training files for each leave-one-out fold and see whether the pseudo-retinal images produced by the Pix2Pix model trained on the target dataset are present in the source pool; if they are, the 77.98% average Dice is not a measurement of generalization to an unseen domain.

Watch

Extended reading notes

Core claim

The central claim is that combining structural augmentation with stylistic augmentation improves domain generalization for retinal vessel segmentation more than either alone. The paper's full method, DGSSA, reports average DSC 77.98% across STARE (82.17), HRF (72.66), DRIVE (78.62), and CHASEDB1 (78.45), surpassing the AADG baseline (77.60) and other comparison methods. The authors attribute the gain to structure augmentation: space-colonization-generated vessel masks rendered by a multi-scale Pix2Pix model into pseudo-retinal images, which forces the segmentor to learn a wider range of vessel morphologies; style augmentation (PixMix plus uncertainty perturbation) then adds photometric variation on top of both original and synthetic images. Their ablation shows structure augmentation alone accounts for most of the thin-vessel improvement, while the full combination reaches the highest average.

Load-bearing premise

The reported generalization scores are valid only if no image or generator trained on the held-out target dataset contributes to the source training pool, and the paper never states that exclusion rule.

Editorial extensions

If this is right

  • If correct, adding procedurally generated vascular structures to the training set improves cross-dataset retinal vessel segmentation beyond style-only augmentation.
  • The method requires no change to the segmentation architecture and no extra computation at inference, so it can be attached to existing segmentors.
  • The reported thin-vessel analysis implies the benefit concentrates on fine vessel morphology, where structure augmentation gives a 13.1% relative Dice improvement over baseline.
  • The combination is architecture-agnostic: the paper reports gains with both DeepLabv3+/MobileNetV2 and U-Net/ResNet backbones.
  • Using generated structures plus style mixing on the source domains could reduce the need for collecting labeled data from every new clinical site.

Reading between the lines

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

  • A testable extension is to vary the number of source domains and the number of generated structures per domain; the marginal value of structural augmentation may shrink once real source diversity is large.
  • The same space-colonization-plus-synthesis recipe could be applied to other tubular structures such as coronary arteries or bronchi, where a branching prior matches the anatomy.
  • The unresolved exclusion rule could be checked by rerunning one fold with target-trained generated images deliberately included and excluded; the difference would bound the contamination effect.
  • Because the paper reports that structure augmentation helps thin vessels most, a clinical deployment might measure generalization gains on microvessel density rather than overall Dice.
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 / 7 minor

Summary. The paper proposes DGSSA, a domain-generalization framework for retinal vessel segmentation that combines structural augmentation (space-colonization-generated vascular structures rendered into pseudo-retinal images by an improved Pix2Pix model) with stylistic augmentation (PixMix with photometric augmentations and uncertainty perturbations). The segmentation network is DeepLabv3+ with a MobileNetV2 backbone. The method is evaluated under a leave-one-out protocol on STARE, HRF, DRIVE, and CHASEDB1, reporting an average DSC of 77.98% and claiming state-of-the-art performance over prior DG methods such as AADG, DoFE, and ELCFS. The paper also includes ablations, a thin/thick-vessel decomposition, hyperparameter sensitivity analysis, and statistical comparisons.

Significance. If the reported results are valid, the central idea is interesting: adding procedurally generated structural diversity to style augmentation is a cheap and potentially effective DG mechanism for tubular structures like retinal vessels. The paper has several strengths: the structural generation pipeline is described algorithmically, the ablation study isolates structural and stylistic contributions, the thin-vessel analysis directly addresses the clinically relevant challenge of capillary segmentation, and the authors attempt statistical comparisons and hyperparameter sensitivity checks. However, the headline SOTA claim rests on a small 0.38-percentage-point average DSC margin over AADG, the paper does not provide code or error bars for the main table, and the leave-one-out protocol as written leaves open a serious data-leakage ambiguity. The significance of the contribution is therefore conditional on resolving these reporting and protocol issues.

major comments (3)
  1. [Section 4.1 and Section 4.2] The leave-one-out protocol does not state whether the Pix2Pix generator trained on the target dataset and its generated pseudo-images are excluded from the source training pool when that dataset is the held-out target. Section 4.2 says the improved Pix2Pix is trained "independently on each dataset" to synthesize 100 images per dataset, and Section 3.1 says the generated images are combined with the original dataset for segmentation training. The current text therefore permits a target-trained generator to contribute pseudo-images with target-domain appearance to the source-only training pool, which would violate the DG setting and invalidate the comparison with AADG and DoFE. Please state the exclusion rule explicitly for each fold and confirm that only generators trained on the three source datasets are used in that fold's training set.
  2. [Section 4.3, Table 1] The prose claims in Section 4.3 are numerically inconsistent with Table 1. The text says AADG achieves a 3.29% average DSC improvement over baseline, but Table 1 gives AADG average 77.60 vs baseline 75.13 (a 2.47-point gap). It then says DGSSA improves another 3.78% over baseline, but Table 1 gives DGSSA average 77.98 (2.85 points over baseline), and the stated "0.60% increase on the HRF dataset" is also inconsistent with the table's HRF values (DGSSA 72.66 vs baseline 72.23, a 0.43-point gap). Please correct these deltas and rephrase the SOTA claim to reflect the actual 0.38-point average margin over AADG.
  3. [Section 4.3 and Section 5.4, Tables 1 and 5] The statistical support for the SOTA claim is under-specified. Table 5 reports paired t-test p-values, but the paper does not state whether the unit of analysis is each fold (n=4), per-image DSC, or per-pixel DSC; no standard deviations or confidence intervals accompany the central Table 1 results. Given the small average margin over AADG, the claim that DGSSA "surpasses existing methods" is not yet quantitatively supported. Please report per-fold results with standard deviations across repeated runs and clarify the t-test unit, sample size, and whether the test is paired across the same test images.
minor comments (7)
  1. [Section 3.2 and Section 3.3] There are two algorithms labeled "Algorithm 1" (the space colonization procedure and the PixMix procedure); renumber the second as Algorithm 2 and update the in-text reference in Section 3.3.
  2. [Abstract and throughout] The dataset name appears as "CHASEDB" in the abstract but "CHASEDB1" elsewhere; please standardize the spelling.
  3. [Section 4.1] The opening sentence says "We conducted a comprehensive evaluation of the impact of AADG" but the paper evaluates DGSSA; this appears to be a typo.
  4. [Section 3.4, Eqs. (17)-(18)] These equations restate the loss functions but do not include the unpaired-data terms defined in Eqs. (5)-(7) and (11), and the hyperparameter values (λ_L1=100, λ_adv=0.2, λ_1=0.3, λ_GP=10) are only given in the text; please align the equations with the full loss definitions.
  5. [Figure 5] The tested ranges for λ_1, λ_adv, λ_GP, and λ_L1 are not stated in the caption or text; please specify them so the sensitivity analysis is reproducible.
  6. [Table 1 caption] The caption refers to "domains a, b, c, and d" while the columns are labeled A, B, C, D, and AVG; please clarify the notation and define what AVG averages over, and note explicitly which rows are borrowed from AADG rather than re-evaluated.
  7. [General reproducibility] The paper does not mention code or data release; providing the implementation and pretrained generators would substantially strengthen reproducibility.

Circularity Check

0 steps flagged · score 2.0 of 10

No demonstrated circularity: the 77.98% DSC claim rests on held-out leave-one-out evaluation against external benchmarks, with only minor non-load-bearing self-citations in the UDA background.

full rationale

Reported DSCs are held-out leave-one-out evaluations on four external public datasets; the segmentation network is trained on source domains and the target test images are not used to fit the segmentation loss. Structural augmentation is produced by an independent space-colonization procedure (Algorithm 1) with stated parameters, not fit to the target DSC. Style augmentation is a standard PixMix-style pipeline with uncertainty perturbation, and the loss equations define a conventional GAN plus BCE segmentation training objective. I find no step where a reported prediction is algebraically or statistically identical to its inputs by construction. The only author self-citations are Refs [13]-[15], used for UDA background and not load-bearing for the central DG claim. I flag one procedural ambiguity: Sections 3.1 and 4.2 say the improved Pix2Pix model is trained "independently on each dataset" and that 100 structures are generated "for each of the four datasets," while Section 4.1 states that "only the training images from the multi-source domains were input into the network." The paper does not explicitly state the leave-one-out exclusion rule for the generator. If a target-trained generator contributed pseudo-images to the source pool, the "unseen target" premise would be violated and the DG comparison would be compromised. This is a real validity risk, but it is an unstated procedural condition, not a demonstrated circular reduction, so it does not by itself raise the circularity score.

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

Everything the central claim rests on is either public datasets, standard augmentation tools, or hand-set parameters. There are no invented entities. The free parameters are the procedural geometry constants (attraction radius 5, segment length 20, kill radius 5, Murray exponent 3), the four GAN loss weights (100, 0.2, 0.3, 10), the unreported PixMix hyperparameters (K, delta, p), and the thin-vessel radius threshold of 1.2 pixels. The key domain assumptions are that space-colonization trees rendered by Pix2Pix are realistic enough to transfer, that the four datasets are valid distinct domains for the leave-one-out protocol, and that generated masks can serve as supervision. The most fragile assumption is procedural: the paper does not state whether a target dataset's generator is excluded when that dataset is the target.

free parameters (5)
  • Space colonization geometry parameters (attraction radius d, segment length l, kill radius k) = 5, 20, 5 (arbitrary units)
    Hand-chosen in Section 3.2 and Algorithm 1; they determine branch density and vessel morphology of the generated structures and therefore the structural augmentation curriculum.
  • Murray's law exponent n = 3
    Set "typically to 3" in Section 3.2, Equation 2, to compute child branch radii from parent radius; controls simulated vessel thickness taper.
  • Pix2Pix loss weights (lambda_L1, lambda_adv, lambda_1, lambda_GP) = 100, 0.2, 0.3, 10
    Set in Section 3.4 with a sensitivity analysis (Figure 5) showing broad stability for the first three and a decline when lambda_GP grows.
  • PixMix hyperparameters (max mixing rounds K, mixing ratio delta, perturbation probability p) = not reported
    Defined in Section 3.3 but no concrete values are given; the strength of the style augmentation depends on them.
  • Thin/thick vessel radius threshold = 1.2 pixels
    Chosen in Section 5.5 to partition vessels for the thin-vessel DSC analysis; affects the reported 13.1% thin-vessel relative improvement.
assumptions (5)
  • domain assumption Space colonization trees are realistic enough vascular mimics that segmentation trained on their GAN renderings transfers to real unseen fundus images
    Core premise of the structural augmentation (Section 3.2, Figure 2); if the pseudo-images are unrealistic, the structural gain over baseline vanishes.
  • domain assumption The four public datasets (STARE, HRF, DRIVE, CHASEDB1) form distinct domains whose leave-one-out protocol measures generalization to unseen acquisitions
    Standard assumption in the cited DG literature [19,21], adopted in Section 4.1 with only four target evaluations.
  • domain assumption Pseudo-retinal images synthesized by the per-dataset Pix2Pix are paired with their generated masks and can serve as supervision for the segmentation model
    Implied in Section 3.1 and Figure 1; the segmentation loss (Equation 19) requires ground-truth masks for the generated images.
  • ad hoc to paper Gradient penalty and multi-scale discrimination stabilize the improved Pix2Pix without changing the distribution of generated anatomy
    The multi-scale discriminator and the paired/unpaired loss split (Equations 4-11) are new components introduced here with no independent validation beyond the reported DSC gains.
  • domain assumption Sample-level paired t-test across images is a valid statistical basis for the significance claims
    Used in Section 5.4; images within a dataset share acquisition conditions and are not independent samples, so the test's validity is questionable.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DGSSA: Domain generalization with structural and stylistic augmentation for retinal vessel segmentation." pith.science (2026). https://pith.science/paper/EGEJUJUK

@misc{pith2026250103466,
  author       = {Pith},
  title        = {Pith review of: DGSSA: Domain generalization with structural and stylistic augmentation for retinal vessel segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EGEJUJUK}},
  note         = {Machine review of arXiv:2501.03466}
}
read the original abstract

Retinal vascular morphology is crucial for diagnosing diseases such as diabetes, glaucoma, and hypertension, making accurate segmentation of retinal vessels essential for early intervention. Traditional segmentation methods assume that training and testing data share similar distributions, which can lead to poor performance on unseen domains due to domain shifts caused by variations in imaging devices and patient demographics. This paper presents a novel approach, DGSSA, for retinal vessel image segmentation that enhances model generalization by combining structural and style augmentation strategies. We utilize a space colonization algorithm to generate diverse vascular-like structures that closely mimic actual retinal vessels, which are then used to generate pseudo-retinal images with an improved Pix2Pix model, allowing the segmentation model to learn a broader range of structure distributions. Additionally, we utilize PixMix to implement random photometric augmentations and introduce uncertainty perturbations, thereby enriching stylistic diversity and significantly enhancing the model's adaptability to varying imaging conditions. Our framework has been rigorously evaluated on four challenging datasets-DRIVE, CHASEDB, HRF, and STARE-demonstrating state-of-the-art performance that surpasses existing methods. This validates the effectiveness of our proposed approach, highlighting its potential for clinical application in automated retinal vessel analysis.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

65 extracted references · 59 canonical work pages

  1. [1]

    Retinal imaging and image analysis,

    M. D. Abràmoff, M. i. Garvin, and M. Sonka, "Retinal imaging and image analysis," IEEE reviews in biomedical engineering, vol. 3, pp. 169-208, 2010

    work page 2010

  2. [2]

    Large-scale retrieval for medical image analytics: A comprehensive review,

    Z. Li, X. Zhang, H. Müller, and S. Zhang, "Large-scale retrieval for medical image analytics: A comprehensive review," Medical image analysis, vol. 43, pp. 66-84, 2018

    work page 2018

  3. [3]

    Refuge challenge: A unified framework for evaluating automated methods for glaucoma assessment from fundus photographs,

    J. I. Orlando et al., "Refuge challenge: A unified framework for evaluating automated methods for glaucoma assessment from fundus photographs," Medical image analysis, vol. 59, p. 101570, 2020

    work page 2020

  4. [4]

    Idrid: Diabetic retinopathy–segmentation and grading challenge,

    P. Porwal et al., "Idrid: Diabetic retinopathy–segmentation and grading challenge," Medical image analysis, vol. 59, p. 101561, 2020. 20

    work page 2020

  5. [5]

    Domain adaptation for medical image analysis: a survey,

    H. Guan and M. Liu, "Domain adaptation for medical image analysis: a survey," IEEE Transactions on Biomedical Engineering, vol. 69, no. 3, pp. 1173- 1185, 2021

    work page 2021

  6. [6]

    Domain-adversarial training of neural networks,

    Y . Ganin et al., "Domain-adversarial training of neural networks," Journal of machine learning research, vol. 17, no. 59, pp. 1-35, 2016

    work page 2016

  7. [7]

    Unsupervised domain adaptation in brain lesion segmentation with adversarial networks,

    i. iamnitsas et al., "Unsupervised domain adaptation in brain lesion segmentation with adversarial networks," in Information Processing in Medical Imaging: 25th International Conference, IPMI 2017, Boone, NC, USA, June 25-30, 2017, Proceedings 25, 2017: Springer, pp. 597-609

    work page 2017

  8. [8]

    Patch -based output space adversarial learning for joint optic disc and cup segmentation,

    S. Wang, L. Y u, X. Yang, C.-W. Fu, and P.-A. Heng, "Patch -based output space adversarial learning for joint optic disc and cup segmentation," IEEE transactions on medical imaging, vol. 38, no. 11, pp. 2485-2495, 2019

    work page 2019

Show all 65 references
  1. [9]

    Domain generalization via invariant feature representation,

    i. Muandet, D. Balduzzi, and B. Schölkopf, "Domain generalization via invariant feature representation," in International conference on machine learning, 2013: PMLR, pp. 10-18

    2013

  2. [10]

    Domain generalization: A survey,

    i. Zhou, Z. Liu, Y . Qiao, T. Xiang, and C. C. Loy, "Domain generalization: A survey," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 45, no. 4, pp. 4396-4415, 2022

    2022

  3. [11]

    A comprehensive survey on transfer learning,

    F. Zhuang et al., "A comprehensive survey on transfer learning," Proceedings of the IEEE, vol. 109, no. 1, pp. 43-76, 2020

    2020

  4. [12]

    Unsupervised domain adaptation in semantic segmentation: a review,

    M. Toldo, A. Maracani, U. Michieli, and P. Zanuttigh, "Unsupervised domain adaptation in semantic segmentation: a review," Technologies, vol. 8, no. 2, p. 35, 2020

    2020

  5. [13]

    Unsupervised domain adaptation for cross-modality liver segmentation via joint adversarial learning and self-learning,

    J. Hong, S. C.-H. Y u, and W. Chen, "Unsupervised domain adaptation for cross-modality liver segmentation via joint adversarial learning and self-learning," Applied Soft Computing, vol. 121, p. 108729, 2022

    2022

  6. [14]

    Source-free unsupervised domain adaptation for cross-modality abdominal multi-organ segmentation,

    J. Hong, Y .-D. Zhang, and W. Chen, "Source-free unsupervised domain adaptation for cross-modality abdominal multi-organ segmentation," Knowledge- Based Systems, vol. 250, p. 109155, 2022

    2022

  7. [15]

    Source-free unsupervised adaptive segmentation for knee joint MRI,

    S. Li, S. Zhao, Y . Zhang, J. Hong, and W. Chen, "Source-free unsupervised adaptive segmentation for knee joint MRI," Biomedical Signal Processing and Control, vol. 92, p. 106028, 2024/06/01/ 2024, doi: https://doi.org/10.1016/j.bspc.2024.106028

    2024

  8. [16]

    Domain generalization via model-agnostic learning of semantic features,

    Q. Dou, D. Coelho de Castro, i. iamnitsas, and B. Glocker, "Domain generalization via model-agnostic learning of semantic features," Advances in neural information processing systems, vol. 32, 2019

    2019

  9. [17]

    Generalizing to unseen domains: A survey on domain generalization,

    J. Wang et al., "Generalizing to unseen domains: A survey on domain generalization," IEEE transactions on knowledge and data engineering, vol. 35, no. 8, pp. 8052-8072, 2022

    2022

  10. [18]

    Domain generalization for medical image analysis: A survey,

    J. S. Yoon, i. Oh, Y . Shin, M. A. Mazurowski, and H. -I. Suk, "Domain generalization for medical image analysis: A survey," arXiv preprint arXiv:2310.08598, 2023

    2023 arXiv

  11. [19]

    Dofe: Domain-oriented feature embedding for generalizable fundus image segmentation on unseen datasets,

    S. Wang, L. Yu, i. Li, X. Yang, C.-W. Fu, and P .-A. Heng, "Dofe: Domain-oriented feature embedding for generalizable fundus image segmentation on unseen datasets," IEEE Transactions on Medical Imaging, vol. 39, no. 12, pp. 4237-4248, 2020

    2020

  12. [20]

    Feddg: Federated domain generalization on medical image segmentation via episodic learning in continuous frequency space,

    Q. Liu, C. Chen, J. Qin, Q. Dou, and P. -A. Heng, "Feddg: Federated domain generalization on medical image segmentation via episodic learning in continuous frequency space," in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021, pp. 1013-1023

    2021

  13. [21]

    AADG: Automatic augmentation for domain generalization on retinal image segmentation,

    J. Lyu, Y . Zhang, Y . Huang, L. Lin, P. Cheng, and X. Tang, "AADG: Automatic augmentation for domain generalization on retinal image segmentation," IEEE Transactions on Medical Imaging, vol. 41, no. 12, pp. 3699-3711, 2022

    2022

  14. [22]

    Retinal arteriolar narrowing and risk of coronary heart disease in men and women: the Atherosclerosis Risk in Communities Study,

    T. Y . Wong et al., "Retinal arteriolar narrowing and risk of coronary heart disease in men and women: the Atherosclerosis Risk in Communities Study," Jama, vol. 287, no. 9, pp. 1153-1159, 2002

    2002

  15. [23]

    Impact of current and past blood pressure on retinal arteriolar diameter in an older population,

    H. Leung, J. J. Wang, E. Rochtchina, T. Y . Wong, R. ilein, and P. Mitchell, "Impact of current and past blood pressure on retinal arteriolar diameter in an older population," Journal of hypertension, vol. 22, no. 8, pp. 1543-1549, 2004

    2004

  16. [24]

    Retinal vascular image analysis as a potential screening tool for cerebrovascular disease: a rationale based on homology between cerebral and retinal microvasculatures,

    N. Patton, T. Aslam, T. MacGillivray, A. Pattie, I. J. Deary, and B. Dhillon, "Retinal vascular image analysis as a potential screening tool for cerebrovascular disease: a rationale based on homology between cerebral and retinal microvasculatures," Journal of anatomy, vol. 206...

    2005

  17. [25]

    Segmentation of retinal blood vessels by combining the detection of centerlines and morpholo gical reconstruction,

    A. M. Mendonca and A. Campilho, "Segmentation of retinal blood vessels by combining the detection of centerlines and morpholo gical reconstruction," IEEE transactions on medical imaging, vol. 25, no. 9, pp. 1200-1213, 2006

    2006

  18. [26]

    Detection of blood vessels in retinal images using two-dimensional matched filters,

    S. Chaudhuri, S. Chatterjee, N. iatz, M. Nelson, and M. Goldbaum, "Detection of blood vessels in retinal images using two-dimensional matched filters," IEEE Transactions on medical imaging, vol. 8, no. 3, pp. 263-269, 1989

    1989

  19. [27]

    Segmentation of vessel -like patterns using mathematical morphology and curvature evaluation,

    F. Zana and J. -C. ilein, "Segmentation of vessel -like patterns using mathematical morphology and curvature evaluation," IEEE transactions on image processing, vol. 10, no. 7, pp. 1010-1019, 2001. 21

    2001

  20. [28]

    U-net: Convolutional networks for biomedical image segmentation,

    O. Ronneberger, P . Fischer, and T. Brox, "U-net: Convolutional networks for biomedical image segmentation," in Medical image computing and computer- assisted intervention–MICCAI 2015: 18th international conference, Munich, Germany, October 5 -9, 2015, proceedings, part III 18...

    2015

  21. [29]

    Ce-net: Context encoder network for 2d medical image segmentation,

    Z. Gu et al., "Ce-net: Context encoder network for 2d medical image segmentation," IEEE transactions on medical imaging, vol. 38, no. 10, pp. 2281 - 2292, 2019

    2019

  22. [30]

    Unet++: Redesigning skip connections to exploit multiscale features in image segmentation,

    Z. Zhou, M. M. R. Siddiquee, N. Tajbakhsh, and J. Liang, "Unet++: Redesigning skip connections to exploit multiscale features in image segmentation," IEEE transactions on medical imaging, vol. 39, no. 6, pp. 1856-1867, 2019

    2019

  23. [31]

    CS-Net: Channel and spatial attention network for curvilinear structure segmentation,

    L. Mou et al., "CS-Net: Channel and spatial attention network for curvilinear structure segmentation," in Medical Image Computing and Computer Assisted Intervention–MICCAI 2019: 22nd International Conference, Shenzhen, China, October 13–17, 2019, Proceedings, Part I 22, 2019: ...

    2019

  24. [32]

    Retinal vessel segmentation using deep learning: a review,

    C. Chen, J. H. Chuah, R. Ali, and Y . Wang, "Retinal vessel segmentation using deep learning: a review," IEEE Access, vol. 9, pp. 111985-112004, 2021

    2021

  25. [33]

    Virtual adversarial training: a regularization method for supervised and semi-supervised learning,

    T. Miyato, S.-i. Maeda, M. ioyama, and S. Ishii, "Virtual adversarial training: a regularization method for supervised and semi-supervised learning," IEEE transactions on pattern analysis and machine intelligence, vol. 41, no. 8, pp. 1979-1993, 2018

    1979

  26. [34]

    Augmix: A simple data processing method to improve robustness and uncertainty,

    D. Hendrycks, N. Mu, E. D. Cubuk, B. Zoph, J. Gilmer, and B. Lakshminarayanan, "Augmix: A simple data processing method to improve robustness and uncertainty," arXiv preprint arXiv:1912.02781, 2019

    1912 arXiv

  27. [35]

    Augmax: Adversarial composition of random augmentations fo r robust training,

    H. Wang, C. Xiao, J. iossaifi, Z. Yu, A. Anandkumar, and Z. Wang, "Augmax: Adversarial composition of random augmentations fo r robust training," Advances in neural information processing systems, vol. 34, pp. 237-250, 2021

    2021

  28. [36]

    The many faces of robustness: A critical analysis of out-of-distribution generalization,

    D. Hendrycks et al., "The many faces of robustness: A critical analysis of out-of-distribution generalization," in Proceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 8340-8349

    2021

  29. [37]

    Autoaugment: Learning augmentation policies from data,

    E. D. Cubuk, B. Zoph, D. Mane, V . V asudevan, and Q. V . Le, "Autoaugment: Learning augmentation policies from data," arXiv preprint arXiv:1805.09501, 2018

    2018 arXiv

  30. [38]

    Improving robustness against common corruptions with frequency biased models,

    T. Saikia, C. Schmid, and T. Brox, "Improving robustness against common corruptions with frequency biased models," in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 10211-10220

    2021

  31. [39]

    Prime: A few primitives can boost robustness to common corruptions,

    A. Modas, R. Rade, G. Ortiz-Jiménez, S.-M. Moosavi-Dezfooli, and P. Frossard, "Prime: A few primitives can boost robustness to common corruptions," in European Conference on Computer Vision, 2022: Springer, pp. 623-640

    2022

  32. [40]

    Amplitude-phase recombination: Rethinking robustness of convolutional neural networks in frequency domain,

    G. Chen, P . Peng, L. Ma, J. Li, L. Du, and Y . Tian, "Amplitude-phase recombination: Rethinking robustness of convolutional neural networks in frequency domain," in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 458-467

    2021

  33. [41]

    Pixmix: Dreamlike pictures comprehensively improve safety measures,

    D. Hendrycks et al., "Pixmix: Dreamlike pictures comprehensively improve safety measures," in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 16783-16792

    2022

  34. [42]

    Generative adversarial networks,

    I. Goodfellow et al., "Generative adversarial networks," Communications of the ACM, vol. 63, no. 11, pp. 139-144, 2020

    2020

  35. [43]

    DR -GAN: conditional generative adversarial network for fine -grained lesion synthesis on diabetic retinopathy images,

    Y . Zhou, B. Wang, X. He, S. Cui, and L. Shao, "DR -GAN: conditional generative adversarial network for fine -grained lesion synthesis on diabetic retinopathy images," IEEE journal of biomedical and health informatics, vol. 26, no. 1, pp. 56-66, 2020

    2020

  36. [44]

    Unsupervised domain adaptation for cross -modality retinal vessel segmentation via disentangling representation style transfer and collaborative consistency learning,

    L. Peng, L. Lin, P . Cheng, Z. Huang, and X. Tang, "Unsupervised domain adaptation for cross -modality retinal vessel segmentation via disentangling representation style transfer and collaborative consistency learning," in 2022 IEEE 19th International Symposium on Biomedical I...

    2022

  37. [45]

    I-secret: Importance-guided fundus image enhancement via semi-supervised contrastive constraining,

    P. Cheng, L. Lin, Y . Huang, J. Lyu, and X. Tang, "I-secret: Importance-guided fundus image enhancement via semi-supervised contrastive constraining," in Medical Image Computing and Computer Assisted Intervention –MICCAI 2021: 24th International Conference, Strasbourg, France,...

    2021

  38. [46]

    Missing MRI pulse sequence synthesis using multi -modal generative adversarial network,

    A. Sharma and G. Hamarneh, "Missing MRI pulse sequence synthesis using multi -modal generative adversarial network," IEEE transactions on medical imaging, vol. 39, no. 4, pp. 1170-1183, 2019

    2019

  39. [47]

    Towards adversarial retinal image synthesis,

    P. Costa et al., "Towards adversarial retinal image synthesis," arXiv preprint arXiv:1701.08974, 2017

    2017 arXiv

  40. [48]

    Image-to-image translation with conditional adversarial networks,

    P. Isola, J.-Y . Zhu, T. Zhou, and A. A. Efros, "Image-to-image translation with conditional adversarial networks," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 1125-1134

    2017

  41. [49]

    End-to-end adversarial retinal image synthesis,

    P. Costa et al., "End-to-end adversarial retinal image synthesis," IEEE transactions on medical imaging, vol. 37, no. 3, pp. 781-791, 2017

    2017

  42. [50]

    Synthetic medical images from dual generative adversarial networks,

    J. T. Guibas, T. S. Virdi, and P. S. Li, "Synthetic medical images from dual generative adversarial networks," arXiv preprint arXiv:1709.01872, 2017. 22

    2017 arXiv

  43. [51]

    Synthesizing retinal and neuronal images with generative adversarial nets,

    H. Zhao, H. Li, S. Maurer-Stroh, and L. Cheng, "Synthesizing retinal and neuronal images with generative adversarial nets," Medical image analysis, vol. 49, pp. 14-26, 2018

    2018

  44. [52]

    SkrGAN: Sketching-rendering unconditional generative adversarial networks for medical image synthesis,

    T. Zhang et al., "SkrGAN: Sketching-rendering unconditional generative adversarial networks for medical image synthesis," in Medical Image Computing and Computer Assisted Intervention –MICCAI 2019: 22nd International Conference, Shenzhen, China, October 13 –17, 2019, Proceedin...

    2019

  45. [53]

    Yolocurvseg: You only label one noisy skeleton for vessel -style curvilinear structure segmentation,

    L. Lin et al., "Yolocurvseg: You only label one noisy skeleton for vessel -style curvilinear structure segmentation," Medical Image Analysis, vol. 90, p. 102937, 2023

    2023

  46. [54]

    Data augmentation using learned transformations for one -shot medical image segmentation,

    A. Zhao, G. Balakrishnan, F. Durand, J. V . Guttag, and A. V . Dalca, "Data augmentation using learned transformations for one -shot medical image segmentation," in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 8543-8553

    2019

  47. [55]

    Improving the generalizability of convolutional neural network-based segmentation on CMR images,

    C. Chen et al., "Improving the generalizability of convolutional neural network-based segmentation on CMR images," Frontiers in cardiovascular medicine, vol. 7, p. 105, 2020

    2020

  48. [56]

    Staining invariant features for improving generalization of d eep convolutional neural networks in computational pathology,

    S. Otálora, M. Atzori, V . Andrearczyk, A. ihan, and H. Müller, "Staining invariant features for improving generalization of d eep convolutional neural networks in computational pathology," Frontiers in bioengineering and biotechnology, vol. 7, p. 198, 2019

    2019

  49. [57]

    Generalizing deep learning for medical image segmentation to unseen domains via deep stacked transformation,

    L. Zhang et al., "Generalizing deep learning for medical image segmentation to unseen domains via deep stacked transformation," IEEE transactions on medical imaging, vol. 39, no. 7, pp. 2531-2540, 2020

    2020

  50. [58]

    Modeling trees with a space colonization algorithm,

    A. Runions, B. Lane, and P. Prusinkiewicz, "Modeling trees with a space colonization algorithm," Nph, vol. 7, no. 63-70, p. 6, 2007

    2007

  51. [59]

    Locating blood vessels in retinal images by piecewise threshold probing of a matched filter response,

    A. Hoover, V . iouznetsova, and M. Goldbaum, "Locating blood vessels in retinal images by piecewise threshold probing of a matched filter response," IEEE Transactions on Medical imaging, vol. 19, no. 3, pp. 203-210, 2000

    2000

  52. [60]

    Robust vessel segmentation in fundus images,

    A. Budai, R. Bock, A. Maier, J. Hornegger, and G. Michelson, "Robust vessel segmentation in fundus images," International journal of biomedical imaging, vol. 2013, no. 1, p. 154860, 2013

    2013

  53. [61]

    Ridge -based vessel segmentation in color images of the retina,

    J. Staal, M. D. Abràmoff, M. Niemeijer, M. A. Viergever, and B. Van Ginneken, "Ridge -based vessel segmentation in color images of the retina," IEEE transactions on medical imaging, vol. 23, no. 4, pp. 501-509, 2004

    2004

  54. [62]

    An ensemble classification-based approach applied to retinal blood vessel segmentation,

    M. M. Fraz et al., "An ensemble classification-based approach applied to retinal blood vessel segmentation," IEEE Transactions on Biomedical Engineering, vol. 59, no. 9, pp. 2538-2548, 2012

    2012

  55. [63]

    Manifold mixup: Better representations by interpolating hidden states,

    V . Verma et al., "Manifold mixup: Better representations by interpolating hidden states," in International conference on machine learning, 2019: PMLR, pp. 6438-6447

    2019

  56. [64]

    Cutmix: Regularization strategy to train strong classifiers with localizable features,

    S. Yun, D. Han, S. J. Oh, S. Chun, J. Choe, and Y . Yoo, "Cutmix: Regularization strategy to train strong classifiers with localizable features," in Proceedings of the IEEE/CVF international conference on computer vision, 2019, pp. 6023-6032

    2019

  57. [65]

    Statistical comparisons of classifiers over multiple data sets,

    J. Demšar, "Statistical comparisons of classifiers over multiple data sets," Journal of Machine learning research, vol. 7, no. Jan, pp. 1-30, 2006

    2006

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.