Pith. sign in

REVIEW 5 major objections 5 minor 1 cited by

Federated Foundation Model for GI Endoscopy Images

T0 review · 5 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Federated pretraining of a masked autoencoder on unlabeled endoscopy images distributed across hospital clients yields representations that transfer to classification, detection, and segmentation nearly as well as centralized training.

desk verdict A useful but not groundbreaking empirical study of federated MAE pretraining for GI endoscopy; the main claim is confounded by data volume, and the paper needs variance estimates and cleaner baselines before it can be fully trusted. read the letter →

arxiv 2505.24108 v2 pith:7QOCTPFE submitted 2025-05-30 cs.CV cs.LG

Alina Devkota , Annahita Amireskandari , Joel Palko , Shyam Thakkar , Donald Adjeroh , Xiajun Jiang , Binod Bhattarai , Prashnna K. Gyawali This is my paper · ORCID
classification cs.CVcs.LG
keywords federatedlearningfoundationmodelmaskedautoencodergastrointestinalendoscopymedicalimageanalysisself-supervisedVisionTransformertransfer
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 a federated framework for pretraining a foundation model for gastrointestinal endoscopy images, where unlabeled images stay at local hospital nodes and only model updates are shared. Using a masked autoencoder objective, the authors pretrain a Vision Transformer backbone across five clients and a server, then fine-tune on classification (GastroVision) and on detection and segmentation (EDD2020). Their FedFoundAvg model improves classification accuracy by 8.9 points over a single-site lower bound (78.0% vs 69.1%), approaching the centralized upper bound (81.3%), with similar gains in detection and segmentation. The claim is that federated, label-free pretraining can build a general-purpose medical vision model without aggregating sensitive data.

What carries the argument

The central object is the federated masked autoencoder (MAE): a Vision Transformer encoder trained to reconstruct masked image patches with an L1 loss, where pretraining is distributed across N client nodes plus a server node and aggregated by a federated algorithm (FedAvg, FedAvgM, FedAdam, or FedAdagrad). This combines the task-agnostic representation learning of MAE with the privacy-preserving update exchange of federated learning; the encoder's frozen or fine-tuned features are then evaluated through task-specific heads (an MLP for classification, ViTDet with Mask R-CNN losses for detection and segmentation).

What would settle it

Train the same MAE model on the full 125,000 images for the same number of optimizer steps and epochs as the federated run (the current upper bound uses fewer compute hours); if this centralized model reaches or exceeds FedFoundAvg's downstream accuracy, the gap the paper attributes to federation is really a data-volume or compute effect.

Watch

Extended reading notes

Core claim

The central claim is that federated self-supervised pretraining with masked autoencoding is a viable route to general-purpose GI endoscopy representations. Across both homogeneous and heterogeneous client splits, the federated models (especially FedFoundAvg, which uses FedAvg aggregation) consistently outperform a lower bound pretrained on a single site's data and approach an upper bound trained on all pooled data. Concretely, FedFoundAvg reaches 78.0% classification accuracy (lower bound 69.1%, upper bound 81.3%), detection mAP of 18.649 (lower 14.008, upper 20.860), and segmentation Dice of 0.6301 (lower 0.5202, upper 0.6488) in the homogeneous setting; in the heterogeneous setting it reaches 77.8% accuracy, 20.353 mAP, and 0.6909 Dice, the latter exceeding the upper bound. The paper interprets these results as showing that federated foundation models can approach the performance of centralized training while preserving data privacy.

Load-bearing premise

The lower-bound baseline is pretrained on only the server's 10,000 images, whereas FedFound uses roughly 125,000 images, so the improvement the paper credits to federation could instead come from simply seeing more data.

Editorial extensions

If this is right

  • If the claim holds, hospitals can collaboratively build a shared endoscopy foundation model without exporting patient data, sidestepping a major barrier to medical foundation models.
  • FedAvg, the least sophisticated aggregation tested, delivers the best overall transfer, suggesting that simple averaging is sufficient for federated MAE pretraining on medical images.
  • Frozen features from FedFoundAvg support lightweight classifiers (MLP, SVM, XGBoost) with accuracy near the upper bound, so institutions with limited labels or compute can still benefit.
  • The heterogeneous-split results, where clients hold non-overlapping imaging domains, indicate the approach degrades gracefully under realistic domain shift, supporting multi-institutional deployment.
  • The early-epoch convergence behavior of FedFoundAvg suggests that federated pretraining can reach usable representations faster than single-site training.

Reading between the lines

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

  • Because the lower bound is pretrained on one-tenth of the data that FedFound sees, the reported gains may overstate the value of federation per se; a matched-data-volume control is needed to isolate the collaboration effect.
  • The result that FedFoundAvg beats the upper bound on heterogeneous segmentation Dice hints that federated averaging may act as an implicit regularizer or benefit from diverse data ordering, a hypothesis testable by comparing ensembles of local models against their average.
  • A practical extension is to apply the same protocol to non-medical image benchmarks with deliberately matched compute and data volume to quantify how much federation costs or saves relative to centralized training.
  • Another testable extension is to inject differential privacy or secure aggregation into the update exchange and measure the resulting accuracy drop, giving deployers a privacy-utility tradeoff curve.
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 a federated learning framework for pretraining masked-autoencoder foundation models on unlabeled GI endoscopy images distributed across five clients and a server. It compares four FL aggregation algorithms (FedAvg, FedAvgM, FedAdam, FedAdagrad) under homogeneous and heterogeneous data splits, then fine-tunes the pretrained encoders for classification on GastroVision and detection/segmentation on EDD2020. The central claimed result is that FedFoundAvg improves classification accuracy from 69.1% (lower bound) to 78.0%, approaching the 81.3% centralized upper bound, with analogous gains in detection and segmentation.

Significance. The paper addresses a relevant and timely problem: privacy-preserving pretraining of medical foundation models. Its strengths are the systematic comparison of four FL algorithms, two distribution settings, three downstream tasks, and a frozen-feature evaluation, all grounded in existing public datasets. If the federated-collaboration benefit were cleanly isolated, this would be a useful benchmark. At present, however, the comparison to the lower bound conflates federation with data volume and domain mix, and the absence of error bars and the downstream-based epoch selection weaken the quantitative claims. The contribution is therefore promising but not yet established.

major comments (5)
  1. [Section IV-A] The homogeneous split description states that 18,938 images are assigned to each of five clients and 10,000 images to the server, but these numbers sum to 104,690, not the 125,846 images implied by 87,970 + 37,876. This leaves 21,156 images unaccounted for and makes the pretraining data pool and the lower-bound data volume ambiguous. This matters because the main comparison in Section V.A is between FedFound models and a lower bound trained on only the server's 10,000 images; if the accounting is off, the data-volume confound is even harder to evaluate.
  2. [Section V.A versus Section IV.A] The central claim that federated collaboration improves over the lower bound is confounded with data volume and domain mix. The lower bound is pretrained on the server's 10,000 HyperKvasir images, whereas FedFoundAvg is pretrained on the distributed pool of about 125,846 images from both HyperKvasir and Kvasir-Capsule. The paper's own upper bound (centralized full-data training) reaches 81.3% accuracy, exceeding FedFoundAvg's 78.0%, so full data access is sufficient to explain the improvement. To support the claim that the federated aggregation mechanism itself is beneficial, the authors need a control that varies only the federation mechanism while matching total data and computational budget, such as a single-site model trained on the full pool of 125,846 images.
  3. [Section IV-B] The paper states that pretraining was run for 1,000 epochs, 'a value selected based on performance on downstream tasks.' This is a form of evaluation leakage: choosing pretraining epochs using the downstream test data can bias the comparison, especially if the epoch count was tuned to maximize FedFound performance. The authors should either select the epoch count using a validation split that is disjoint from the test set, or report downstream performance across a range of pretraining epochs for all methods so that the selection rule is transparent and fair to the baselines.
  4. [Tables B.1 and B.2, Section V] All reported metrics are single runs with no error bars, confidence intervals, or significance tests. Several important comparisons are close: for homogeneous segmentation, FedFoundAvgM achieves DSC 0.6959 versus FedFoundAdam 0.6617 and upper bound 0.6488; for homogeneous detection, FedFoundAvg achieves mAP 18.649 versus upper bound 20.860. Without repeated seeds and variance estimates, the claims that FedFound models 'consistently outperform' the lower bound and that one FL algorithm is better than another are not statistically supported. Please provide mean and standard deviation over at least three seeds for the main results, and use appropriate tests or confidence intervals when ranking algorithms.
  5. [Algorithm A.1 versus Eq. (4)] Equation (4) defines a sample-size-weighted federated objective with a 1/n sum of n_k terms, but Algorithm A.1 line 11 aggregates client updates as a uniform average 1/|S| times the sum of delta_{t,i}. If the experiments use Algorithm A.1, the implemented objective differs from Eq. (4); if they use weighted averaging, Algorithm A.1 is incorrectly specified. The distinction is non-trivial here because the server holds 10,000 images while clients hold approximately 18,938 each. Please state exactly which aggregation rule was used in the experiments and align the algorithm description and the mathematical objective.
minor comments (5)
  1. [Eq. (1)] The left-hand side should be theta* = arg min_theta ..., not L(theta*) = arg min_theta ...; as written, the equation defines the loss value circularly.
  2. [Section IV-A] There is a typo: 'HypeKvasir' should be 'HyperKvasir.'
  3. [Section V-B] The word 'acehives' should be 'achieves.'
  4. [Eqs. (1) and (4)] The self-supervised target is written with a tilde in Eq. (1) and as 'fX' in Eq. (4); please use consistent notation.
  5. [Tables B.1 and B.2] The captions state that underlined values represent the second-highest results, but no underlining is visible in the manuscript text; please ensure the formatting is present in the final version.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the paper's central claim is an empirical benchmark comparison and does not reduce by construction to its inputs.

full rationale

The paper's contribution is a federated masked-autoencoder pretraining pipeline for GI endoscopy images, and its central claims are supported by direct experimental measurements on held-out downstream tasks (classification accuracy, detection mAP, segmentation Dice) rather than by a derivation that reduces to its own assumptions. Equation 4 is a standard FedAvg-style weighted aggregation of local MAE reconstruction losses; it defines the training objective, and the reported downstream performance is an empirical outcome, not a quantity forced by that equation. The comparison against lower and upper bounds is a benchmark design, not a fitted-parameter construction. The only text resembling tuning-based circularity is the statement in Section IV-B that the model 'was trained for 1,000 epochs — a value selected based on performance on downstream tasks,' which is a normal hyperparameter selection procedure, not a fitted parameter renamed as a prediction. The paper also cites the authors' own prior multimodal FL review (reference [17]), but that citation is background context only and is not load-bearing; no uniqueness theorem or methodological premise is imported from it. The skeptical concern that the lower bound is confounded with data volume (server-only 10,000 images versus the federated pool of roughly 125,000 images) is a legitimate experimental-design limitation and correctness risk, but it is not circularity: the claim could in principle be false, and the paper's evidence is empirical rather than definitional. No step in the claimed derivation chain is equivalent to its inputs by construction, so the circularity score is 0.

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

The empirical claim depends mainly on the choice of pretraining objective, the FL aggregation strategy, dataset splits, and the assumed separation between pretraining and downstream data. No new physical or mathematical entities are introduced.

free parameters (5)
  • Pretraining epoch count = 1,000
    Section IV.B states this value was 'selected based on performance on downstream tasks,' making it a hand-tuned choice tied to the evaluation.
  • MAE masking ratio = 0.6
    Section IV.B; chosen by the authors and affects pretraining difficulty and downstream transfer.
  • Server dataset size in homogeneous split = 10,000 images (clients 18,938 each)
    Section IV-A; the lower-bound baseline is defined by this size, so the comparison is sensitive to it.
  • Heterogeneous client assignment = Two clients receive Kvasir-Capsule images; three clients and server receive HyperKvasir
    Section IV-A; this hand-constructed split defines the non-IID scenario and is not generated by a standardized protocol.
  • Fine-tuning hyperparameters = Not reported
    Learning rates, schedules, augmentation, and downstream train/validation splits are not given, and these choices can affect the ranking among FedFound variants.
assumptions (4)
  • domain assumption MAE reconstruction loss is a sufficient self-supervised objective for learning transferable GI endoscopy representations.
    The entire framework is built on this choice; Section III-A adopts MAE with L1 loss without comparing to other self-supervised objectives.
  • domain assumption Averaging locally trained MAE weights via FedAvg, FedAvgM, FedAdam, or FedAdagrad yields a global model whose fine-tuned performance approximates the centralized model.
    Section III-C assumes federated aggregation of self-supervised updates is meaningful; convergence is shown only empirically in Section V-C.
  • domain assumption HyperKvasir and Kvasir-Capsule are representative enough of GI endoscopy distributions for conclusions to generalize.
    The authors acknowledge in Section VI that the study is limited to two public datasets and may not capture real-world diversity.
  • domain assumption Pretraining and downstream datasets are disjoint at the patient level.
    Not discussed in the paper. Gastrovision and HyperKvasir both include data from Baerum Hospital, so patient overlap is possible and could inflate downstream performance.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Federated Foundation Model for GI Endoscopy Images." pith.science (2026). https://pith.science/paper/7QOCTPFE

@misc{pith2026250524108,
  author       = {Pith},
  title        = {Pith review of: Federated Foundation Model for GI Endoscopy Images},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7QOCTPFE}},
  note         = {Machine review of arXiv:2505.24108}
}
read the original abstract

Gastrointestinal (GI) endoscopy is essential in identifying GI tract abnormalities in order to detect diseases in their early stages and improve patient outcomes. Although deep learning has shown success in supporting GI diagnostics and decision-making, these models require curated datasets with labels that are expensive to acquire. Foundation models offer a promising solution by learning general-purpose representations, which can be finetuned for specific tasks, overcoming data scarcity. Developing foundation models for medical imaging holds significant potential, but the sensitive and protected nature of medical data presents unique challenges. Foundation model training typically requires extensive datasets, and while hospitals generate large volumes of data, privacy restrictions prevent direct data sharing, making foundation model training infeasible in most scenarios. In this work, we propose a FL framework for training foundation models for gastroendoscopy imaging, enabling data to remain within local hospital environments while contributing to a shared model. We explore several established FL algorithms, assessing their suitability for training foundation models without relying on task-specific labels, conducting experiments in both homogeneous and heterogeneous settings. We evaluate the trained foundation model on three critical downstream tasks--classification, detection, and segmentation--and demonstrate that it achieves improved performance across all tasks, highlighting the effectiveness of our approach in a federated, privacy-preserving setting.

Figures

Figures reproduced from arXiv: 2505.24108 by the authors.

Figure 1
Figure 1. Workflow of Federated Foundation Model. The federated pretraining setup consists of multiple clients and a central server. Each client trains its local model using the local dataset and computes model updates and sends them to the server to form a global model (θ) via aggregation. The pretrained FFM is fine-tuned on three downstream tasks: classification, object detection, and semantic segmentation. our methods on b… view at source ↗
Figure 2
Figure 2. Conceptual trade-off between privacy and performance in foundation model training. Our federated training objective aims to achieve high performance with strong privacy, bridging the gap between centralized (upper bound) and single-node (lower bound) training. To formalize this idea, we introduce the notion of upper bound and lower bound performance in the context of FFM training. The upper bound represents the idea… view at source ↗
Figure 3
Figure 3. Image distribution across homogeneous and heterogeneous splits for federated pretraining. The bar chart shows the number of images assigned to each node (Client 1 to Client 5 and Server) using different homogeneous and heterogeneous distribution strategies. of 110,079 endoscopy images, and Kvasir-Capsule consists of 47,238 images; from these, 87,970 and 37,876 images were randomly selected, respectively, for model p… view at source ↗
Figures from the paper (2 more)
Figure 5
Figure 5. Figure 5: Performance comparison of various methods on the homo￾geneous split. FedFound Models are evaluated against baseline, lower bound, and upper bound models across different downstream tasks. V. RESULTS In this section, we present the detailed performance of the FFMs acros…
Figure 8
Figure 8. Figure 8: Evaluation of downstream classification performance using frozen features from various foundation models for homogeneous (Left) and heterogeneous (Right) settings. this strategy in both homogeneous and heterogeneous settings [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Leaderless Collective Motion in Affine Formation Control over the Complex Plane

    cs.RO 2026-04 unverdicted novelty 6.0 of 10

    Modifying Laplacian weights yields leaderless affine collective motion of planar robot swarms, with closed-form eigenvectors/eigenvalues designed via complex-plane analysis.

Reference graph

Works this paper leans on

42 extracted references · 27 canonical work pages · cited by 1 Pith paper

  1. [1]

    Gastrovision: A multi-class endoscopy image dataset for computer aided gastrointestinal disease detection,

    D. Jha, V . Sharma, N. Dasu, N. K. Tomar, S. Hicks, M. K. Bhuyan, P. K. Das, M. A. Riegler, P. Halvorsen, U. Bagci et al., “Gastrovision: A multi-class endoscopy image dataset for computer aided gastrointestinal disease detection,” in Workshop on Machine Learning for Multimodal Healthcare Data. Springer, 2023, pp. 125–140

    work page 2023

  2. [2]

    Artificial intelligence in gastrointestinal endoscopy,

    R. Pannala, K. Krishnan, J. Melson, M. A. Parsi, A. R. Schulman, S. Sullivan, G. Trikudanathan, A. J. Trindade, R. R. Watson, J. T. Maple et al. , “Artificial intelligence in gastrointestinal endoscopy,” VideoGIE, vol. 5, no. 12, pp. 598–613, 2020

    work page 2020

  3. [3]

    Effect of a deep learning–based automatic upper gi endoscopic reporting system: a randomized crossover study (with video),

    L. Zhang, Z. Lu, L. Yao, Z. Dong, W. Zhou, C. He, R. Luo, M. Zhang, J. Wang, Y . Liet al., “Effect of a deep learning–based automatic upper gi endoscopic reporting system: a randomized crossover study (with video),” Gastrointestinal Endoscopy, vol. 98, no. 2, pp. 181–190, 2023

    work page 2023

  4. [4]

    Deep convolutional neural networks for computer-aided detection: Cnn architectures, dataset characteristics and transfer learning,

    H.-C. Shin, H. R. Roth, M. Gao, L. Lu, Z. Xu, I. Nogues, J. Yao, D. Mollura, and R. M. Summers, “Deep convolutional neural networks for computer-aided detection: Cnn architectures, dataset characteristics and transfer learning,” IEEE transactions on medical imaging , vol. 35, no. 5, pp. 1285–1298, 2016

    work page 2016

  5. [5]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2016, pp. 770–778

    2016

  6. [6]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    A. Dosovitskiy, “An image is worth 16x16 words: Transformers for image recognition at scale,” arXiv preprint arXiv:2010.11929 , 2020

    arXiv 2010

  7. [7]

    Automated polyp detection in colon capsule endoscopy,

    A. V . Mamonov, I. N. Figueiredo, P. N. Figueiredo, and Y .-H. R. Tsai, “Automated polyp detection in colon capsule endoscopy,” IEEE transactions on medical imaging , vol. 33, no. 7, pp. 1488–1502, 2014

    work page 2014

  8. [8]

    Deep learning for detection and segmentation of artefact and disease instances in gastrointestinal endoscopy,

    S. Ali, M. Dmitrieva, N. Ghatwary, S. Bano, G. Polat, A. Temizel, A. Krenzer, A. Hekalo, Y . B. Guo, B. Matuszewskiet al., “Deep learning for detection and segmentation of artefact and disease instances in gastrointestinal endoscopy,” Medical image analysis, vol. 70, p. 102002, 2021

    work page 2021

Show all 42 references
  1. [9]

    Novel deep learning–based computer-aided diag- nosis system for predicting inflammatory activity in ulcerative colitis,

    Y . Fan, R. Mu, H. Xu, C. Xie, Y . Zhang, L. Liu, L. Wang, H. Shi, Y . Hu, J. Ren et al., “Novel deep learning–based computer-aided diag- nosis system for predicting inflammatory activity in ulcerative colitis,” Gastrointestinal Endoscopy, vol. 97, no. 2, pp. 335–346, 2023

    2023

  2. [10]

    Foundation model for endoscopy video analysis via large-scale self-supervised pre-train,

    Z. Wang, C. Liu, S. Zhang, and Q. Dou, “Foundation model for endoscopy video analysis via large-scale self-supervised pre-train,” in International Conference on Medical Image Computing and Computer- Assisted Intervention. Springer, 2023, pp. 101–111

    2023

  3. [11]

    A foundation model for generalizable disease detection from retinal images,

    Y . Zhou, M. A. Chia, S. K. Wagner, M. S. Ayhan, D. J. Williamson, R. R. Struyven, T. Liu, M. Xu, M. G. Lozano, P. Woodward-Court et al., “A foundation model for generalizable disease detection from retinal images,” Nature, vol. 622, no. 7981, pp. 156–163, 2023

    2023

  4. [12]

    Masked au- toencoders are scalable vision learners,

    K. He, X. Chen, S. Xie, Y . Li, P. Doll ´ar, and R. Girshick, “Masked au- toencoders are scalable vision learners,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2022, pp. 16 000–16 009

    2022

  5. [13]

    A simple framework for contrastive learning of visual representations,

    T. Chen, S. Kornblith, M. Norouzi, and G. Hinton, “A simple framework for contrastive learning of visual representations,” in International conference on machine learning . PMLR, 2020, pp. 1597–1607

    2020

  6. [14]

    Foundation models defining a new era in vision: a survey and outlook,

    M. Awais, M. Naseer, S. Khan, R. M. Anwer, H. Cholakkal, M. Shah, M.-H. Yang, and F. S. Khan, “Foundation models defining a new era in vision: a survey and outlook,” IEEE Transactions on Pattern Analysis and Machine Intelligence , 2025

    2025

  7. [15]

    Federated learning and differential privacy for medical image analysis,

    M. Adnan, S. Kalra, J. C. Cresswell, G. W. Taylor, and H. R. Tizhoosh, “Federated learning and differential privacy for medical image analysis,” Scientific reports, vol. 12, no. 1, p. 1953, 2022

    1953

  8. [16]

    Communication-efficient learning of deep networks from decentralized data,

    B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y Arcas, “Communication-efficient learning of deep networks from decentralized data,” in Artificial intelligence and statistics . PMLR, 2017, pp. 1273– 1282

    2017

  9. [17]

    Multimodal federated learning in health- care: a review,

    J. Thrasher, A. Devkota, P. Siwakotai, R. Chivukula, P. Poudel, C. Hu, B. Bhattarai, and P. Gyawali, “Multimodal federated learning in health- care: a review,” arXiv preprint arXiv:2310.09650 , 2023

    2023 arXiv

  10. [18]

    Deep learning-based prediction model for diagnosing gastrointestinal diseases using endoscopy images,

    A. Sharma, R. Kumar, and P. Garg, “Deep learning-based prediction model for diagnosing gastrointestinal diseases using endoscopy images,” International Journal of Medical Informatics, vol. 177, p. 105142, 2023

    2023

  11. [19]

    Development of a real-time endoscopic image diagnosis support system using deep learning technology in colonoscopy,

    M. Yamada, Y . Saito, H. Imaoka, M. Saiko, S. Yamada, H. Kondo, H. Takamaru, T. Sakamoto, J. Sese, A. Kuchiba et al. , “Development of a real-time endoscopic image diagnosis support system using deep learning technology in colonoscopy,” Scientific reports, vol. 9, no. 1, p. 14...

    2019

  12. [20]

    Attention enabled multiresunet for bio-medical image segmentation,

    C. Rahman, R. K. Bhuiyan, S. P. Shyam, R. Subnom, and A. B. Rashid, “Attention enabled multiresunet for bio-medical image segmentation,” in 2024 6th International Conference on Electrical Engineering and Information & Communication Technology (ICEEICT) . IEEE, 2024, pp. 622–627

    2024

  13. [21]

    Deep-learning system detects neoplasia in patients with barrett’s esophagus with higher accuracy than endoscopists in a multistep training and validation study with benchmarking,

    A. J. de Groof, M. R. Struyvenberg, J. van der Putten, F. van der Sommen, K. N. Fockens, W. L. Curvers, S. Zinger, R. E. Pouw, E. Coron, F. Baldaque-Silva et al. , “Deep-learning system detects neoplasia in patients with barrett’s esophagus with higher accuracy than endoscopis...

    2020

  14. [22]

    Foundation models in gastrointestinal endoscopic ai: Impact of architecture, pre-training approach and data efficiency,

    T. G. Boers, K. N. Fockens, J. A. van der Putten, T. J. Jaspers, C. H. Kusters, J. B. Jukema, M. R. Jong, M. R. Struyvenberg, J. de Groof, J. J. Bergman et al., “Foundation models in gastrointestinal endoscopic ai: Impact of architecture, pre-training approach and data efficie...

    2024

  15. [23]

    Endodino: A foundation model for gi endoscopy,

    P. Dermyer, A. Kalra, and M. Schwartz, “Endodino: A foundation model for gi endoscopy,” arXiv preprint arXiv:2501.05488 , 2025

    2025 arXiv

  16. [24]

    Endovit: pretraining vision transformers on a large collection of endoscopic images,

    D. Bati ´c, F. Holm, E. ¨Ozsoy, T. Czempiel, and N. Navab, “Endovit: pretraining vision transformers on a large collection of endoscopic images,” International Journal of Computer Assisted Radiology and Surgery, vol. 19, no. 6, pp. 1085–1091, 2024

    2024

  17. [25]

    Divergence-aware federated self- supervised learning,

    W. Zhuang, Y . Wen, and S. Zhang, “Divergence-aware federated self- supervised learning,” arXiv preprint arXiv:2204.04385 , 2022

    2022 arXiv

  18. [26]

    Federated self-supervised learning for heterogeneous clients,

    D. Makhija, N. Ho, and J. Ghosh, “Federated self-supervised learning for heterogeneous clients,” arXiv preprint arXiv:2205.12493 , 2022

    2022 arXiv

  19. [27]

    Label-efficient self-supervised federated learning for tackling data heterogeneity in medical imaging,

    R. Yan, L. Qu, Q. Wei, S.-C. Huang, L. Shen, D. L. Rubin, L. Xing, and Y . Zhou, “Label-efficient self-supervised federated learning for tackling data heterogeneity in medical imaging,” IEEE Transactions on Medical Imaging, vol. 42, no. 7, pp. 1932–1943, 2023

    1932

  20. [28]

    Fedfms: Exploring federated foundation models for medical image segmentation,

    Y . Liu, G. Luo, and Y . Zhu, “Fedfms: Exploring federated foundation models for medical image segmentation,” in International Conference on Medical Image Computing and Computer-Assisted Intervention . Springer, 2024, pp. 283–293

    2024

  21. [29]

    Federated learning for medical image classification: A comprehensive benchmark,

    Z. Zhou, G. Luo, M. Chen, Z. Weng, and Y . Zhu, “Federated learning for medical image classification: A comprehensive benchmark,” arXiv preprint arXiv:2504.05238, 2025

    2025 arXiv

  22. [30]

    Federated endovit: Pretraining vision transformers via federated learning on endoscopic image collec- tions,

    M. Kirchner, A. C. Jenke, S. Bodenstedt, F. R. Kolbinger, O. Saldanha, J. N. Kather, M. Wagner, and S. Speidel, “Federated endovit: Pretraining vision transformers via federated learning on endoscopic image collec- tions,” arXiv preprint arXiv:2504.16612 , 2025

    2025

  23. [31]

    Measuring the effects of non- identical data distribution for federated visual classification,

    T.-M. H. Hsu, H. Qi, and M. Brown, “Measuring the effects of non- identical data distribution for federated visual classification,” arXiv preprint arXiv:1909.06335, 2019

    1909 arXiv

  24. [32]

    Adaptive federated optimization,

    S. Reddi, Z. Charles, M. Zaheer, Z. Garrett, K. Rush, J. Kone ˇcn`y, S. Kumar, and H. B. McMahan, “Adaptive federated optimization,”arXiv preprint arXiv:2003.00295, 2020

    2003 arXiv

  25. [33]

    Exploring plain vision transformer backbones for object detection,

    Y . Li, H. Mao, R. Girshick, and K. He, “Exploring plain vision transformer backbones for object detection,” in European conference on computer vision . Springer, 2022, pp. 280–296

    2022

  26. [34]

    Mask r-cnn,

    K. He, G. Gkioxari, P. Doll ´ar, and R. Girshick, “Mask r-cnn,” in Proceedings of the IEEE international conference on computer vision , 2017, pp. 2961–2969

    2017

  27. [35]

    Hyperkvasir, a comprehensive multi-class image and video dataset for gastrointestinal endoscopy,

    H. Borgli, V . Thambawita, P. H. Smedsrud, S. Hicks, D. Jha, S. L. Eskeland, K. R. Randel, K. Pogorelov, M. Lux, D. T. D. Nguyen et al., “Hyperkvasir, a comprehensive multi-class image and video dataset for gastrointestinal endoscopy,” Scientific data, vol. 7, no. 1, p. 283, 2020

    2020

  28. [36]

    Kvasir-capsule, a video capsule endoscopy dataset,

    P. H. Smedsrud, V . Thambawita, S. A. Hicks, H. Gjestang, O. O. Nedrejord, E. Næss, H. Borgli, D. Jha, T. J. D. Berstad, S. L. Eskeland et al. , “Kvasir-capsule, a video capsule endoscopy dataset,” Scientific Data, vol. 8, no. 1, p. 142, 2021. DEVKOTAet al.: FEDERATED FOUNDATI...

    2021

  29. [37]

    Endoscopy disease detection and segmentation (edd2020),

    S. A. B. B. D. L. S. R. A. B. R. C. N. G. J. R. C. D. J. East, “Endoscopy disease detection and segmentation (edd2020),” 2020. [Online]. Available: https://dx.doi.org/10.21227/f8xg-wb80 10 Supplementary Material APPENDIX A. FedFound Algorithms Algorithm A.1 Federated Learning ...

    2020 doi

  30. [38]

    BOLD VALUES INDICATE THE HIGHEST VALUES FOR EACH METRIC , AND UNDERLINED VALUES REPRESENT THE SECOND -HIGHEST VALUES

    Homogeneous Distribution: TABLE B.1 PERFORMANCE METRICS ACROSS VARIOUS DOWNSTREAM TASKS WITH DIFFERENT PRETRAINING METHODS FOR HOMOGENEOUS SPLIT . BOLD VALUES INDICATE THE HIGHEST VALUES FOR EACH METRIC , AND UNDERLINED VALUES REPRESENT THE SECOND -HIGHEST VALUES . Methods Cla...

  31. [39]

    BOLD VALUES INDICATE THE HIGHEST VALUES FOR EACH METRIC , AND UNDERLINED VALUES REPRESENT THE SECOND -HIGHEST VALUES

    Heterogeneous Distribution: TABLE B.2 PERFORMANCE METRICS ACROSS VARIOUS DOWNSTREAM TASKS WITH DIFFERENT PRETRAINING METHODS FOR HETEROGENEOUS SPLIT . BOLD VALUES INDICATE THE HIGHEST VALUES FOR EACH METRIC , AND UNDERLINED VALUES REPRESENT THE SECOND -HIGHEST VALUES . Methods...

  32. [40]

    MLP Classifier: Fig. C.1. Downstream classification performance using frozen features from different foundation models under homogeneous (Left) and heterogeneous (Right) settings, using an MLP classifier

  33. [41]

    SVM Classifier: Fig. C.2. Downstream classification performance using frozen features from different foundation models under homogeneous (Left) and heterogeneous (Right) settings, using an SVM classifier

  34. [42]

    XGBoost Classifier: Fig. C.3. Downstream classification performance using frozen features from different foundation models under homogeneous (Left) and heterogeneous (Right) settings, using an XGBoost classifier

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.