Pith. sign in

REVIEW 6 major objections 6 minor 33 references

Grad-CL: Source Free Domain Adaptation with Gradient Guided Feature Disalignment

T0 review · 6 major / 6 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read Gradient-guided feature maps refine pseudolabels and disalign optic cup and disc features, enabling source-free domain adaptation for fundus image segmentation.

desk verdict A sensible, incremental SFDA method for cup/disc segmentation, but the core mechanism is untested on target data and the ablation doesn't ablate. read the letter →

arxiv 2509.10134 v1 pith:T2WOQ76U submitted 2025-09-12 cs.CV

Rini Smita Thakur , Rajeev Ranjan Dwivedi , Vinod K Kurmi This is my paper
classification cs.CV
keywords source-freedomainadaptationfundusimagesegmentationopticdiscandcupgradient-guidedpseudolabelscontrastivelearningcosinesimilarityGrad-CAMglaucomadiagnosis
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

Grad-CL is a source-free domain adaptation method for segmenting the optic cup and disc in fundus images. It starts from a model trained on a labeled source dataset and adapts it to an unlabeled target dataset using two mechanisms: gradient-guided pseudolabel refinement, which uses class-specific gradient maps to sharpen noisy pseudolabels, and a cosine-similarity contrastive loss that pushes cup and disc features apart. The paper reports that this combination outperforms previous unsupervised and source-free adaptation methods on two public fundus datasets, with the largest gains on the harder RIM-ONE-r3 target. If the method holds, it offers a way to adapt medical segmentation models across clinical sites without sharing source data.

What carries the argument

Gradient-guided class-specific feature maps (Grad-CAM-style heatmaps e_cup_GC and e_disc_GC) are the load-bearing object. They are computed by global average pooling of gradients of each class logit with respect to the final convolutional feature maps, then combined with raw features to form prototypes (Eq. 6-8) and contrasted via cosine similarity (Eq. 11-12). The gradient maps thus carry the argument: they localize class-discriminative pixels, making pseudolabel refinement and the contrastive disalignment target both class-aware.

What would settle it

Run Grad-CL with the gradient maps replaced by random or zero masks on the same target datasets; if the performance gain over DPL disappears, the claim that gradient-guided features drive the improvement is confirmed. Alternatively, measure the overlap between the gradient heatmaps and the true cup/disc masks on a heavily shifted target dataset—if the overlap is low, the method's Dice should drop sharply.

Watch

Extended reading notes

Core claim

The central claim is that gradient-based saliency maps of the two classes, computed from the pretrained source model on target images, carry information that improves both stages of source-free adaptation. The maps refine pseudolabel reliability by weighting feature prototypes, and they serve as class-specific feature anchors for a contrastive loss that explicitly disaligns optic cup and disc representations. The paper demonstrates the mechanism through ablations comparing divergence versus distance metrics, and through Dice and ASD scores on Drishti-GS and RIM-ONE-r3.

Load-bearing premise

The entire pipeline assumes the gradient maps computed from the source model on target images reliably point at the optic cup and disc; if domain shift corrupts this localization, both the pseudolabel refinement and the contrastive objective are built on wrong targets.

Editorial extensions

If this is right

  • If correct, source-free domain adaptation for fundus segmentation can be improved without any access to source data, easing privacy and confidentiality constraints in clinical settings.
  • Gradient-guided prototype estimation reduces the influence of noisy pseudolabels compared to prototypes built from combined features, as shown by the improved Dice scores over DPL.
  • Explicitly disaligning optic cup and disc features reduces misclassification between the two classes and improves boundary delineation, reflected in lower average surface distance.
  • The method's gains are largest on the more challenging RIM-ONE-r3 target, suggesting it helps when the domain shift is severe.
  • The approach combines pseudolabeling and contrastive learning in a way that could be adapted to other two-class medical segmentation tasks.

Reading between the lines

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

  • The gradient-guided disalignment idea is not restricted to fundus images; any two-class segmentation with a clear object-pair structure could benefit, though the paper does not test this.
  • Because the gradient maps come from the source model, the method inherits the source model's biases; if the source model localizes classes poorly on target images, errors could be amplified rather than corrected.
  • An empirical test with larger domain shifts or different backbones (e.g., transformer-based segmentation models) would clarify whether the mechanism is general or specific to DeepLabv3+/MobileNetv2.
  • A direct check that the gradient maps actually overlap the true cup and disc on target images would validate the core assumption; the paper does not report such localization accuracy.
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

6 major / 6 minor

Summary. The paper proposes Grad-CL, a source-free domain adaptation method for optic cup and disc segmentation in fundus images. It consists of two stages: (i) gradient-guided pseudolabel refinement, where Grad-CAM-like maps computed from the pretrained source model on target images are used to modulate penultimate-layer features, build prototypes, and construct a refinement mask for pseudolabels; and (ii) contrastive feature disalignment, where cosine similarity between gradient-informed cup and disc features is minimized. The method is evaluated on REFUGE→RIM-ONE-r3 and REFUGE→Drishti-GS, reporting improvements over several unsupervised and source-free baselines. The paper also includes an 'ablation' table comparing divergence versus distance metrics on gradient maps.

Significance. If the mechanism works as claimed, Grad-CL would be a useful contribution to source-free segmentation in medical imaging, particularly for tasks with small, structured anatomical regions. The paper releases code and evaluates on two public benchmarks against multiple UDA and SFDA baselines, which is a strength. However, the core contribution is currently underspecified: the central operation involving gradient-modified features is dimensionally ambiguous, the contrastive loss is not fully defined, the loss weight λ is missing, and the 'ablation' does not test the proposed components. The reported gains are also small in several comparisons and are presented without error bars. The most serious concern is that the entire method rests on the assumption that source-model Grad-CAM maps localize the cup and disc on target images, an assumption that is never validated on target data and could be violated by domain shift.

major comments (6)
  1. [Section 3.1, Eqs. (5)-(8); Section 3.2, Eq. (11)] The quantity e_GCv is never defined. Eq. (5) defines e_cup_GC and e_disc_GC as spatial heatmaps, but Eq. (6) multiplies this quantity by the penultimate-layer feature vector e_v, and Eq. (11) adds it to e_v. If e_GCv is a scalar heatmap value, the vector addition is dimensionally inconsistent; if it is a per-pixel feature vector, its construction is missing. Since every component depends on this operation, the definition and dimensionality must be stated explicitly and the equations must be made consistent.
  2. [Section 3.1, Eq. (9)] The refinement mask only distinguishes object vs. background, although the segmentation task has two foreground classes (cup and disc). The relationship between the binary object indicator \hat{y}=1 and the two logits y_cup, y_disc is never defined. As written, the prototypes z_ob and z_bg are not class-wise cup/disc prototypes, so it is unclear how the method refines class-specific pseudolabels. This disconnect also makes it hard to see how the contrastive disalignment in Section 3.2 interacts with the refinement stage.
  3. [Section 3.2, Eq. (12); Algorithm 1] L_sim is written as a single-pixel cosine similarity with no summation or averaging over pixels; the set of contributing pixels is unspecified. In addition, Algorithm 1 (line 21) writes L_total = L_seg + L_sim without the λ of Eq. (13), and λ is never given in Section 4. The method is therefore not reproducible as stated.
  4. [Section 4, 'Ablation Study' / Table 2] This experiment does not ablate either proposed module. It compares divergence metrics (KL, JS) with distance metrics (MMD, Euclidean) on gradient maps, while the proposed loss is a cosine similarity combined with a segmentation loss. No run is reported that removes the gradient-guided prototypes or the contrastive loss. The claim that the 'ablation study validates' the approach is not supported by the presented experiment.
  5. [Section 3.1, Eqs. (4)-(5); Section 4, Table 2] The entire mechanism assumes that source-model Grad-CAM maps localize the optic cup and disc on target images. The paper provides no target-domain validation of this assumption: no Grad-CAM visualizations on target images, no quantitative localization test against target ground-truth masks, and no comparison with source-domain map quality. Table 2 evaluates distributional distances between cup and disc gradient maps, not localization accuracy. If domain shift corrupts these maps, Eq. (9) will discard correct pseudolabels and Eq. (12) will separate wrong features, so the reported gains could arise from extra training signal rather than from gradient-guided refinement. This is a load-bearing point that needs direct evidence.
  6. [Section 4, Table 1] No error bars or multiple-run statistics are reported. Several improvements over the strongest baselines are small (e.g., cup Dice 80.51 vs 79.78 on RIM-ONE-r3; disc Dice 84.67 vs 83.56 on Drishti-GS). Without variance estimates, the central claim of outperforming state-of-the-art methods is not statistically supported for those comparisons.
minor comments (6)
  1. [Eq. (12)] The denominator uses max(||a||||b||, ε). If the norm product is smaller than ε, the loss is scaled by 1/ε rather than by the actual norm product, which can amplify the loss. Consider using ||a||||b|| + ε or an equivalent standard stabilizer.
  2. [Table 2] The header row is garbled: the repeated 'Cup Disc' entries and the misaligned metric labels make the table difficult to read. The table would benefit from a conventional grouped-header format.
  3. [Section 2.2] PLPB is described as a 'robust image-classification SFDA network', but the cited work is a fundus image segmentation method; the description should be corrected.
  4. [Section 4] The paper gives η=0.05, K=10, and λ is missing. A sensitivity analysis for η and λ would help reproducibility, especially since the refinement mask depends directly on η.
  5. [Algorithm 1] Line 6 uses GAP for the same operation as Eq. (4), but the GAP notation is not defined in the algorithm. Also, the algorithm omits the λ from Eq. (13); this should be aligned.
  6. [Figure 1] Figure 1 is placed after Eq. (8) but is not referenced in the surrounding text. Please cite it where the two-stage pipeline is first described.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: Grad-CL is an empirical self-training method; the only self-citations are background and not load-bearing.

full rationale

Grad-CL is an empirical source-free domain adaptation pipeline. It generates pseudolabels from a pretrained source model on target images, filters them by Monte-Carlo uncertainty, reweights features by Grad-CAM-style gradient maps (Eqs. 4-6), computes prototypes (Eq. 7), and adds a cosine-similarity contrastive term (Eq. 12). The reported claim is a measured Dice/ASD improvement on held-out target ground truth (Table 1), so the central result is not derived from the method's own definitions or fitted parameters. The self-training loop (source predictions refine a model that then makes new predictions) is inherent to SFDA and is not a case of a predicted quantity being the same as a fitted input by construction. The paper's reliance on source-model gradient maps to localize cup and disc on target data is an unverified empirical assumption, not a circular step, unless one could show Eq. 5's maps are defined via the target ground truth they are supposed to predict, which the paper does not do. The only self-citations are Refs. [23] and [28], both cited in related work as examples of uncertainty/energy-based pseudolabel refinement; they are not used to justify the core mechanism, so they are not load-bearing. Overall, no equation reduces to its own input by construction, and the method is evaluated against external benchmarks, so circularity is minimal.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

No new entities are introduced. The method depends on several hand-set thresholds and an unspecified loss weight; no formal verification or code artifacts are provided.

free parameters (4)
  • Pseudolabel threshold gamma = 0.75
    Borrowed from DPL/BEAL protocol; not validated on the target domain.
  • Uncertainty threshold eta = 0.05
    Set by hand; no sensitivity analysis given.
  • Number of MC dropout passes K = 10
    Choice; no justification.
  • Contrastive loss weight lambda = unspecified
    The value is never stated in Section 4 or elsewhere, yet it controls the balance of the central loss in Eq. (13).
assumptions (3)
  • domain assumption Grad-CAM coefficients from the source model provide useful class-specific saliency in the target domain.
    Eqs. (4)-(12) assume the gradient maps localize cup and disc correctly under domain shift.
  • ad hoc to paper Deep features from the penultimate layer can be pointwise multiplied with gradient maps to form 'modified features'.
    Eq. (6) is a heuristic combination without theoretical or empirical justification beyond the reported results.
  • domain assumption The benchmark protocol (REFUGE as source, Drishti-GS/RIM-ONE as targets) is representative for SFDA in fundus imaging.
    The protocol is standard in the cited literature [5,32], but the paper does not test on other source-target pairs.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Grad-CL: Source Free Domain Adaptation with Gradient Guided Feature Disalignment." pith.science (2026). https://pith.science/paper/T2WOQ76U

@misc{pith2026250910134,
  author       = {Pith},
  title        = {Pith review of: Grad-CL: Source Free Domain Adaptation with Gradient Guided Feature Disalignment},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/T2WOQ76U}},
  note         = {Machine review of arXiv:2509.10134}
}
read the original abstract

Accurate segmentation of the optic disc and cup is critical for the early diagnosis and management of ocular diseases such as glaucoma. However, segmentation models trained on one dataset often suffer significant performance degradation when applied to target data acquired under different imaging protocols or conditions. To address this challenge, we propose \textbf{Grad-CL}, a novel source-free domain adaptation framework that leverages a pre-trained source model and unlabeled target data to robustly adapt segmentation performance without requiring access to the original source data. Grad-CL combines a gradient-guided pseudolabel refinement module with a cosine similarity-based contrastive learning strategy. In the first stage, salient class-specific features are extracted via a gradient-based mechanism, enabling more accurate uncertainty quantification and robust prototype estimation for refining noisy pseudolabels. In the second stage, a contrastive loss based on cosine similarity is employed to explicitly enforce inter-class separability between the gradient-informed features of the optic cup and disc. Extensive experiments on challenging cross-domain fundus imaging datasets demonstrate that Grad-CL outperforms state-of-the-art unsupervised and source-free domain adaptation methods, achieving superior segmentation accuracy and improved boundary delineation. Project and code are available at https://visdomlab.github.io/GCL/.

Figures

Figures reproduced from arXiv: 2509.10134 by the authors.

Figure 1
Figure 1. Overview of the proposed Grad-CL technique. The provided target image is fed [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Overlaying the final segmentation on an image from the RIM-ONE-r3 dataset. [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. Final segmentation overlay on a Drishti-GS image. The blue outline marks the [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

33 extracted references · 2 linked inside Pith

  1. [1]

    Unsupervised robust domain adaptation without source data

    Peshal Agarwal, Danda Pani Paudel, Jan-Nico Zaech, and Luc Van Gool. Unsupervised robust domain adaptation without source data. InProceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 2009–2018, 2022

    2009

  2. [2]

    Source-relaxed domain adaptation for image segmentation

    Mathilde Bateson, Hoel Kervadec, Jose Dolz, Hervé Lombaert, and Ismail Ben Ayed. Source-relaxed domain adaptation for image segmentation. In Anne L. Martel, Purang Abolmaesumi, Danail Stoyanov, Diana Mateus, Maria A. Zuluaga, S. Kevin Zhou, Daniel Racoceanu, and Leo Joskowicz, editors,Medical Image Computing and Com- puter Assisted Intervention – MICCAI 2...

    2020

  3. [3]

    Weight uncertainty in neural network

    Charles Blundell, Julien Cornebise, Koray Kavukcuoglu, and Daan Wierstra. Weight uncertainty in neural network. InInternational conference on machine learning, pages 1613–1622. PMLR, 2015

    2015

  4. [4]

    Local con- trastive loss with pseudo-label based self-training for semi-supervised medical image segmentation.Medical Image Analysis, 87:102792, 2023

    Krishna Chaitanya, Ertunc Erdil, Neerav Karani, and Ender Konukoglu. Local con- trastive loss with pseudo-label based self-training for semi-supervised medical image segmentation.Medical Image Analysis, 87:102792, 2023

    2023

  5. [5]

    Source-free do- main adaptive fundus image segmentation with denoised pseudo-labeling

    Cheng Chen, Quande Liu, Yueming Jin, Qi Dou, and Pheng-Ann Heng. Source-free do- main adaptive fundus image segmentation with denoised pseudo-labeling. InMedical Image Computing and Computer Assisted Intervention–MICCAI 2021: 24th Interna- tional Conference, Strasbourg, France, September 27–October 1, 2021, Proceedings, Part V 24, pages 225–235. Springer, 2021

    2021

  6. [6]

    Uc-sfda: Source-free domain adapta- tion via uncertainty prediction and evidence-based contrastive learning.Knowledge- Based Systems, 275:110728, 2023

    Dong Chen, Hongqing Zhu, and Suyi Yang. Uc-sfda: Source-free domain adapta- tion via uncertainty prediction and evidence-based contrastive learning.Knowledge- Based Systems, 275:110728, 2023. ISSN 0950-7051. doi: https://doi.org/10.1016/ j.knosys.2023.110728. URLhttps://www.sciencedirect.com/science/ article/pii/S0950705123004781

    arXiv 2023

  7. [7]

    Encoder-decoder with atrous separable convolution for semantic image seg- mentation

    Liang-Chieh Chen, Yukun Zhu, George Papandreou, Florian Schroff, and Hartwig Adam. Encoder-decoder with atrous separable convolution for semantic image seg- mentation. InProceedings of the European conference on computer vision (ECCV), pages 801–818, 2018

    2018

  8. [8]

    Unsupervised domain adaptive fundus image segmentation with category-level regu- larization

    Wei Feng, Lin Wang, Lie Ju, Xin Zhao, Xin Wang, Xiaoyu Shi, and Zongyuan Ge. Unsupervised domain adaptive fundus image segmentation with category-level regu- larization. InInternational Conference on Medical Image Computing and Computer- Assisted Intervention, pages 497–506. Springer, 2022

    2022

Show all 33 references
  1. [9]

    Rim-one: An open retinal image database for optic nerve evaluation

    Francisco Fumero, Silvia Alayón, José L Sanchez, Jose Sigut, and M Gonzalez- Hernandez. Rim-one: An open retinal image database for optic nerve evaluation. In2011 24th international symposium on computer-based medical systems (CBMS), pages 1–6. IEEE, 2011

    2011

  2. [10]

    Context-aware pseudo-label re- finement for source-free domain adaptive fundus image segmentation

    Zheang Huai, Xinpeng Ding, Yi Li, and Xiaomeng Li. Context-aware pseudo-label re- finement for source-free domain adaptive fundus image segmentation. InInternational Conference on Medical Image Computing and Computer-Assisted Intervention, pages 618–628. Springer, 2023. 12RINI...

    2023

  3. [11]

    Model adaptation: Historical contrastive learning for unsupervised domain adaptation without source data.Advances in Neural Information Processing Systems, 34:3635–3649, 2021

    Jiaxing Huang, Dayan Guan, Aoran Xiao, and Shijian Lu. Model adaptation: Historical contrastive learning for unsupervised domain adaptation without source data.Advances in Neural Information Processing Systems, 34:3635–3649, 2021

    2021

  4. [12]

    What uncertainties do we need in bayesian deep learning for computer vision?Advances in neural information processing systems, 30, 2017

    Alex Kendall and Yarin Gal. What uncertainties do we need in bayesian deep learning for computer vision?Advances in neural information processing systems, 30, 2017

    2017

  5. [13]

    Pseudo-label: The simple and efficient semi-supervised learn- ing method for deep neural networks

    Dong-Hyun Lee et al. Pseudo-label: The simple and efficient semi-supervised learn- ing method for deep neural networks. InWorkshop on challenges in representation learning, ICML, volume 3, page 896. Atlanta, 2013

    2013

  6. [14]

    Haijun Lei, Weixin Liu, Hai Xie, Benjian Zhao, Guanghui Yue, and Baiying Lei. Un- supervised domain adaptation based image synthesis and feature alignment for joint optic disc and cup segmentation.IEEE Journal of Biomedical and Health Informatics, 26(1):90–102, 2021

    2021

  7. [15]

    Maximum density divergence for domain adaptation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(11):3918–3930, 2021

    Jingjing Li, Erpeng Chen, Zhengming Ding, Lei Zhu, Ke Lu, and Heng Tao Shen. Maximum density divergence for domain adaptation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(11):3918–3930, 2021. doi: 10.1109/TPAMI. 2020.2991050

    2021

  8. [16]

    A comprehensive survey on source-free domain adaptation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 46(8):5743–5762, 2024

    Jingjing Li, Zhiqi Yu, Zhekai Du, Lei Zhu, and Heng Tao Shen. A comprehensive survey on source-free domain adaptation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 46(8):5743–5762, 2024. doi: 10.1109/TPAMI.2024.3370978

    2024

  9. [17]

    Robust source-free domain adaptation for fundus image segmentation

    Lingrui Li, Yanfeng Zhou, and Ge Yang. Robust source-free domain adaptation for fundus image segmentation. In2024 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pages 7825–7834, 2024. doi: 10.1109/W ACV57701.2024. 00766

    2024

  10. [18]

    Do we really need to access the source data? source hypothesis transfer for unsupervised domain adaptation

    Jian Liang, Dapeng Hu, and Jiashi Feng. Do we really need to access the source data? source hypothesis transfer for unsupervised domain adaptation. InInternational conference on machine learning, pages 6028–6039. PMLR, 2020

    2020

  11. [19]

    Deep unsupervised domain adaptation: A review of recent advances and perspectives.APSIPA Transactions on Signal and Information Process- ing, 11(1), 2022

    Xiaofeng Liu, Chaehwa Yoo, Fangxu Xing, Hyejin Oh, Georges El Fakhri, Je-Won Kang, Jonghye Woo, et al. Deep unsupervised domain adaptation: A review of recent advances and perspectives.APSIPA Transactions on Signal and Information Process- ing, 11(1), 2022

    2022

  12. [20]

    Open compound domain adaptation

    Ziwei Liu, Zhongqi Miao, Xingang Pan, Xiaohang Zhan, Dahua Lin, Stella X Yu, and Boqing Gong. Open compound domain adaptation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 12406–12415, 2020

    2020

  13. [21]

    Representation learning with con- trastive predictive coding.arXiv preprint arXiv:1807.03748, 2018

    Aaron van den Oord, Yazhe Li, and Oriol Vinyals. Representation learning with con- trastive predictive coding.arXiv preprint arXiv:1807.03748, 2018

    2018 arXiv

  14. [22]

    Bathula, Andrés Diaz-Pinto, Ruogu Fang, Pheng-Ann Heng, Jeyoung Kim, JoonHo Lee, Joonseok Lee, Xiaoxiao Li, Peng Liu, Shuai Lu, Balamurali Murugesan, Valery Naranjo, Sai Samarth R

    José Ignacio Orlando, Huazhu Fu, João Barbosa Breda, Karel van Keer, Deepti R. Bathula, Andrés Diaz-Pinto, Ruogu Fang, Pheng-Ann Heng, Jeyoung Kim, JoonHo Lee, Joonseok Lee, Xiaoxiao Li, Peng Liu, Shuai Lu, Balamurali Murugesan, Valery Naranjo, Sai Samarth R. Phaye, Sharath M....

    2020

  15. [24]

    Grad-cam: Visual explanations from deep networks via gradient-based localization

    Ramprasaath R Selvaraju, Michael Cogswell, Abhishek Das, Ramakrishna Vedantam, Devi Parikh, and Dhruv Batra. Grad-cam: Visual explanations from deep networks via gradient-based localization. InProceedings of the IEEE international conference on computer vision, pages 618–626, 2017

    2017

  16. [25]

    A comprehensive retinal image dataset for the assessment of glaucoma from the optic nerve head analysis.JSM Biomedical Imaging Data Papers, 2(1):1004, 2015

    Jayanthi Sivaswamy, S Krishnadas, Arunava Chakravarty, G Joshi, A Syed Tabish, et al. A comprehensive retinal image dataset for the assessment of glaucoma from the optic nerve head analysis.JSM Biomedical Imaging Data Papers, 2(1):1004, 2015

    2015

  17. [26]

    Source-free do- main adaptive fundus image segmentation with class-balanced mean teacher

    Longxiang Tang, Kai Li, Chunming He, Yulun Zhang, and Xiu Li. Source-free do- main adaptive fundus image segmentation with class-balanced mean teacher. In Hayit Greenspan, Anant Madabhushi, Parvin Mousavi, Septimiu Salcudean, James Dun- can, Tanveer Syeda-Mahmood, and Russell ...

    2023

  18. [27]

    Nearest neighborhood-based deep clustering for source data-absent unsupervised domain adaptation.arXiv preprint arXiv:2107.12585, 2021

    Song Tang, Yan Yang, Zhiyuan Ma, Norman Hendrich, Fanyu Zeng, Shuzhi Sam Ge, Changshui Zhang, and Jianwei Zhang. Nearest neighborhood-based deep clustering for source data-absent unsupervised domain adaptation.arXiv preprint arXiv:2107.12585, 2021

    2021 arXiv

  19. [28]

    Uncertainty and energy based loss guided semi-supervised semantic segmentation

    Rini Smita Thakur and Vinod K Kurmi. Uncertainty and energy based loss guided semi-supervised semantic segmentation. In2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pages 8035–8045, 2025. doi: 10.1109/ W ACV61041.2025.00780

    2025

  20. [29]

    Vibashan, V

    S. Vibashan, V . Jeya Maria Jose Valanarasu, and Vishal M. Patel. Target and task spe- cific source-free domain adaptive image segmentation. InProceedings of the Medical Imaging with Deep Learning (MIDL), pages 1627–1639, June 2024

    2024

  21. [30]

    Advent: Adversarial entropy minimization for domain adaptation in semantic segmen- tation

    Tuan-Hung Vu, Himalaya Jain, Maxime Bucher, Matthieu Cord, and Patrick Pérez. Advent: Adversarial entropy minimization for domain adaptation in semantic segmen- tation. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 2517–2526, 2019

    2019

  22. [31]

    Tent: Fully test-time adaptation by entropy minimization

    Dequan Wang, Evan Shelhamer, Shaoteng Liu, Bruno Olshausen, and Trevor Darrell. Tent: Fully test-time adaptation by entropy minimization. InProceedings of the 9th International Conference on Learning Representations (ICLR), 2021. URLhttps: //openreview.net/forum?id=uXl3bZLkr3c...

    2021

  23. [32]

    Boundary and entropy-driven adversarial learning for fundus image segmentation

    Shujun Wang, Lequan Yu, Kang Li, Xin Yang, Chi-Wing Fu, and Pheng-Ann Heng. Boundary and entropy-driven adversarial learning for fundus image segmentation. In Medical Image Computing and Computer Assisted Intervention–MICCAI 2019: 22nd International Conference, Shenzhen, China...

    2019

  24. [33]

    Patch- based output space adversarial learning for joint optic disc and cup segmentation.IEEE transactions on medical imaging, 38(11):2485–2495, 2019

    Shujun Wang, Lequan Yu, Xin Yang, Chi-Wing Fu, and Pheng-Ann Heng. Patch- based output space adversarial learning for joint optic disc and cup segmentation.IEEE transactions on medical imaging, 38(11):2485–2495, 2019

    2019

  25. [34]

    Divide and contrast: Source-free domain adaptation via adaptive contrastive learning

    Ziyi Zhang, Weikai Chen, Hui Cheng, Zhen Li, Siyuan Li, Liang Lin, and Guanbin Li. Divide and contrast: Source-free domain adaptation via adaptive contrastive learning. Advances in Neural Information Processing Systems, 35:5137–5149, 2022

    2022

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.