Pith. sign in

REVIEW 3 major objections 5 minor 35 references

PATHS: A Hierarchical Transformer for Efficient Whole Slide Image Analysis

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read PATHS, a top-down hierarchical transformer that reads only a few hundred patches per slide, reports superior survival-prediction performance on four of five TCGA datasets and over ten times faster inference than full-slide methods.

desk verdict A genuinely useful efficiency + interpretability contribution whose accuracy headline is over-reached by an encoder-confounded baseline comparison. read the letter →

arxiv 2411.18225 v1 pith:JICMTUYC submitted 2024-11-27 cs.CV cs.AI

Zak Buzzard , Konstantin Hemker , Nikola Simidjievski , Mateja Jamnik This is my paper · ORCID
classification cs.CVcs.AI
keywords wholeslideimagecomputationalpathologyhierarchicaltransformersurvivalpredictionpatchselectionmultipleinstancelearninggigapixelattention
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

Whole-slide images are gigapixel; standard methods treat each slide as a bag of tens of thousands of patches, which is expensive and noise-prone. This paper proposes PATHS, a top-down hierarchical transformer that mimics a pathologist: it starts at low magnification, learns an importance score for each patch, keeps only the top K (here K=20, so at most 80 per level), zooms into those regions, and repeats across five magnification levels. The paper reports that PATHS outperforms existing methods on survival prediction on four of five TCGA datasets (mean c-index 0.665 vs 0.618 for HIPT) while processing only a few hundred patches per slide and running over ten times faster at inference than full-slide methods. The authors' claim is that reading a small, carefully chosen subset of patches at multiple magnifications is both more accurate and far cheaper than reading the whole slide.

What carries the argument

The central object is the PATHS hierarchy: n processors, one per magnification level mi forming a geometric sequence with multiplier M=2 from 0.625x to 10x. Each processor computes contextualized patch features by adding an RNN output over the embeddings of ancestor patches at lower magnifications (hierarchical context), predicts a scalar importance α via an MLP with sigmoid, scales features by α, and feeds them to a transformer decoder with 2D positional encoding for global aggregation. A top-K filter keeps the 20 most important patches per level, and MAGNIFY expands those locations at the next magnification, giving a hard upper bound of $M^{2}$K = 80 patches per level. The slide-level features from all levels are summed as slide-level context, making the model a residual hierarchy. This machinery is what enables quadratic self-attention on a small patch set and provides the interpretable importance scores.

What would settle it

Recompute the five baselines with the same UNI encoder and 10x magnification on the same folds, or run PATHS with the SSL-ViT encoder used for the [6] baselines. If the mean c-index gap over HIPT (reported as 0.047) collapses to within noise, then the top-down hierarchical selection is not the source of the reported improvement.

Watch

Extended reading notes

Core claim

The paper's central claim is that a top-down, cross-magnification hierarchy with learned patch importance can outperform full-slide bag-of-patches and bottom-up hierarchical baselines on slide-level survival prediction, despite processing only a small fraction of the slide. The authors argue that most patches in a whole-slide image are uninformative noise, and that mimicking a pathologist's 'scan, zoom, re-scan' behavior both improves the signal-to-noise ratio and makes quadratic self-attention feasible. They support this with mean c-index 0.665 across IDC, CRC, CCRCC, PRCC and LUAD, best on four of five datasets, and an inference speed-up exceeding a factor of ten at 10x magnification. They also show the learned importance scores localize tumor regions in zero-shot CAMELYON17 visualizations.

Load-bearing premise

The claimed performance advantage rests on the comparability of PATHS scores (computed with the UNI encoder at magnifications up to 10x) to pre-calculated baseline risk scores taken from Chen et al. [6], which used a different patch encoder at 20x; if this encoder mismatch explains the gap, the architecture itself has not been shown superior.

Editorial extensions

If this is right

  • On the five TCGA survival tasks, reading a few hundred carefully chosen patches per slide suffices to match or beat methods that read tens of thousands of patches.
  • The fixed bound of 80 patches per magnification level makes quadratic self-attention practical for slide-level aggregation.
  • The learned importance scores localize tumor regions without patch-level labels, as shown in zero-shot CAMELYON17 examples.
  • Inference latency for a new slide is cut by more than a factor of ten at 10x, because only selected patches are encoded.
  • Because the method is task-agnostic, it transfers to other weakly supervised gigapixel image tasks.

Reading between the lines

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

  • The reported gain over HIPT may stem largely from switching the patch encoder to UNI: the paper's own Table 5 shows the encoder moves mean c-index by 0.112, more than the 0.047 gain over HIPT, so an encoder-controlled comparison is needed before attributing the improvement to the hierarchy.
  • Random patch selection with cross-magnification context scores nearly as high as learned selection (mean 0.656 vs 0.665), suggesting the importance module's main added value may be interpretability rather than accuracy.
  • Since UNI was trained only at 20x, a cross-magnification pre-trained encoder, as the paper suggests, would likely push PATHS further.
  • The same top-down selection principle could apply to other gigapixel domains, such as satellite imagery, where most pixels are irrelevant to the task.
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 / 5 minor

Summary. PATHS is a top-down, hierarchical transformer for whole-slide image analysis that recursively selects a small number of patches (at most K per magnification level) from a WSI pyramid, encodes them with a frozen pre-trained encoder (UNI), and aggregates them with per-level transformers. The paper evaluates the method on survival prediction across five TCGA datasets (IDC, CRC, CCRCC, PRCC, LUAD), reporting a mean c-index of 0.665, which it claims is superior to previous methods while processing far fewer patches, and also reports large inference-speed gains. Additional contributions are an interpretable importance-based patch selection, an ablation study of the architectural components, and zero-shot visualizations on CAMELYON17 showing that selected regions align with tumor annotations.

Significance. If the accuracy claim held under controlled comparison, PATHS would be a meaningful advance: it shows that a top-down hierarchical selector can match or beat full-slide bag-of-patches methods with a fraction of the compute, and its interpretable importance values are clinically attractive. The efficiency evidence is strong and well documented (Figures 3 and 5, Appendix C), and the zero-shot interpretability results on CAMELYON17 (Figures 4 and 7) are convincing and reproducible because the code is provided. The main weakness is that the central accuracy claim, as stated in the abstract and Section 5, is not supported by a controlled architecture comparison: five of six baselines use pre-calculated scores from a different encoder and magnification, and the paper's own Table 5 shows that the encoder choice alone explains more than the reported advantage.

major comments (3)
  1. [Table 1 and Table 5; Section 4 (Baselines and Patch Embedding)] The headline accuracy claim is confounded by the feature encoder and magnification. Five of the six baselines (ABMIL, DeepAttnMISL, GCN-MIL, DS-MIL, HIPT) are evaluated using 'pre-calculated risk scores' from Chen et al. [6], which used the SSL-ViT encoder and 20x patches, whereas PATHS uses UNI at magnifications from 0.625x to 10x. The paper's own Table 5 shows that changing the encoder from SSL-ViT to UNI shifts PATHS mean c-index from 0.553 to 0.665 (a +0.112 improvement), which is larger than the reported PATHS-vs-HIPT gain of +0.047 and the PATHS-vs-ZoomMIL gain of +0.049. Consequently, the 'superior performance' claimed in the abstract and Section 5 is not a controlled architecture comparison. The authors must either rerun the baselines with the same UNI encoder (or otherwise equalize the feature extraction), or substantially qualify the accuracy claim to be about the full pipeline rather than the PATHS architecture itself. The same issue applies to the statement in Section 5 that PATHS 'significantly improves' accuracy over prior methods.
  2. [Table 1, ZoomMIL row; Section 4 (Baselines)] The only same-encoder comparison, against ZoomMIL (both use UNI), does not firmly establish an architecture-level advantage. PATHS has a higher mean c-index (0.665 vs 0.616), but it is worse on LUAD (0.545 vs 0.551) and essentially tied on IDC (0.636 vs 0.634), with per-fold standard deviations between 0.036 and 0.097. No significance tests or confidence intervals are reported, so the difference could easily arise from fold variance. Please add paired significance tests (or at least per-fold paired differences) and discuss the LUAD/IDC results explicitly.
  3. [Table 2 and Algorithm 1, Eq. (7)] The random-selection ablation is not a clean test of the learned selection mechanism. The ablation replaces the predicted α_i with U[0,1] at inference time only, which changes both which patches are selected and the scaling of patch features in Eq. (7) (line 4 of Algorithm 1). Because the aggregator was trained with MLP-produced α values, random α at test time introduces a distribution shift that is not equivalent to random patch selection during training. The observed small drop of 0.009 mean c-index therefore does not cleanly show that learned selection is unimportant. The authors should retrain with a random top-K selection inside the training loop to isolate the effect of the learned importance model.
minor comments (5)
  1. [Abstract and Section 5] The phrase 'superior performance' overstates the evidence given the encoder confound; recommend rewording to 'comparable or better under matched encoder' or similar until the controlled comparison is provided.
  2. [Section 4 (Baselines)] Please state whether the 'pre-calculated risk scores' from [6] use the same five folds and the same preprocessing (patch size, Otsu threshold, tissue segmentation) as the PATHS experiments; otherwise the claim that 'all models are evaluated on the same folds' is not verifiable.
  3. [Algorithm 1] The assignment arrows '← −' are nonstandard and potentially confusing; please use standard notation such as '←' throughout the pseudo-code.
  4. [Table 2] The column headers 'Neither', 'Hierarchical only', and 'Slide-level only' are ambiguous; please define in the caption exactly which context components are included in each variant (e.g., 'Neither' = no hierarchical context and no slide-level context).
  5. [Discussion, second paragraph] The phrase 'reading in fewer uninformative slides' should be 'reading in fewer uninformative patches'; the current wording is a typo that could confuse readers.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: PATHS's predictions are computed from held-out folds and external encoders, and the baseline comparison issue is a confound, not a circular derivation.

full rationale

The paper's derivation chain is self-contained against external benchmarks. Alpha importance is learned end-to-end through a gating mechanism with gradients, and the c-index is evaluated on held-out cross-validation folds rather than on training data. The CAMELYON17 heatmaps are zero-shot after training on TCGA-BRCA, so they are not fitted to the evaluation inputs. The UNI encoder is an external, independently trained feature extractor and, per Section 4, was pre-trained excluding TCGA datasets used in evaluation, so encoder choice is not a fitted parameter of the paper's own method. The ablation study in Table 2 honestly reports that random patch selection costs only a small mean c-index drop, and Table 5 shows encoder sensitivity, meaning the paper does not disguise fitted components as predictions. The main concern—that five of six Table 1 baselines use pre-calculated risk scores from Chen et al. with a different patch encoder and magnification than PATHS—is a comparison-validity confound, not a circular step: those numbers are external evidence, not derived from PATHS's outputs, and they do not make PATHS's own result equivalent to its inputs by construction. No self-citation chain is load-bearing, no uniqueness theorem is imported from the authors' prior work, and no ansatz is smuggled in via citation. Therefore, the correct finding is no significant circularity.

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

The central claims rest on hand-chosen design values and an external encoder rather than fitted physical parameters: K=20, the five-level 0.625x-to-10x pyramid, b=4 buckets, weight 0.6, training hyperparameters, and UNI. None are fitted to test folds, but the evaluation does not isolate their influence: Appendix D shows the encoder is the largest factor in reported accuracy (0.553 SSL-ViT to 0.665 UNI), while Table 2 shows the learned selection is the smallest factor (0.656 random to 0.665 learned).

free parameters (5)
  • K, patches retained per magnification level = 20
    Hard top-K threshold in FILTER bounding each level at M^2 K = 80 patches; all headline results use K=20 (K=10 and K=40 appear only in the speed plot).
  • Magnification pyramid (m1, mn, n, M) = 0.625x to 10x, n=5, M=2
    Geometric pyramid in Eq. (1); the paper fixes it without sensitivity analysis, and the ZoomMIL baseline uses a different pyramid (1.25x, 2.5x, 10x).
  • Survival quantisation buckets b = 4
    Number of hazard buckets for the censored NLL objective; fixed 'in all experiments' without sensitivity analysis.
  • Censored loss weight alpha = 0.6
    Weight in the censored negative log-likelihood; fixed without sensitivity analysis.
  • Training hyperparameters and architecture dims = lr 2e-5, batch 32, epochs 40, aggregator 128/4/2, LSTM 256, IMP 128
    Chosen once and shared across datasets; not fitted to test folds, but no per-dataset validation is reported.
assumptions (5)
  • domain assumption UNI features computed at 20x training are informative for patches at 0.625x to 10x
    All embeddings come from UNI at out-of-domain magnifications; Appendix D mitigates but cannot eliminate this domain gap.
  • domain assumption WSI pyramid levels align exactly under the geometric magnification sequence
    Section 3.2, Eq. (1) requires parent-child patch containment for the RNN context (Eq. 3); real WSI pyramids are only approximately aligned.
  • ad hoc to paper Gating with non-differentiable top-K trains useful importance values
    FILTER is non-differentiable (Appendix A); the paper relies on gradients through IMP gating, and Table 2 shows random alpha is nearly as accurate, so this assumption is only weakly supported.
  • standard math Otsu thresholding reliably separates tissue from background
    Used in MAGNIFY and in the latency protocol (Appendix C); standard practice.
  • domain assumption Quantised censored negative log-likelihood is a valid survival objective
    Adopted from Zadeh and Schmid [33]; standard in computational pathology.

how reviews work

0 comments
Cite this review

Pith. "Pith review of PATHS: A Hierarchical Transformer for Efficient Whole Slide Image Analysis." pith.science (2026). https://pith.science/paper/JICMTUYC

@misc{pith2026241118225,
  author       = {Pith},
  title        = {Pith review of: PATHS: A Hierarchical Transformer for Efficient Whole Slide Image Analysis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JICMTUYC}},
  note         = {Machine review of arXiv:2411.18225}
}
abstract

Computational analysis of whole slide images (WSIs) has seen significant research progress in recent years, with applications ranging across important diagnostic and prognostic tasks such as survival or cancer subtype prediction. Many state-of-the-art models process the entire slide - which may be as large as $150,000 \times 150,000$ pixels - as a bag of many patches, the size of which necessitates computationally cheap feature aggregation methods. However, a large proportion of these patches are uninformative, such as those containing only healthy or adipose tissue, adding significant noise and size to the bag. We propose Pathology Transformer with Hierarchical Selection (PATHS), a novel top-down method for hierarchical weakly supervised representation learning on slide-level tasks in computational pathology. PATHS is inspired by the cross-magnification manner in which a human pathologist examines a slide, recursively filtering patches at each magnification level to a small subset relevant to the diagnosis. Our method overcomes the complications of processing the entire slide, enabling quadratic self-attention and providing a simple interpretable measure of region importance. We apply PATHS to five datasets of The Cancer Genome Atlas (TCGA), and achieve superior performance on slide-level prediction tasks when compared to previous methods, despite processing only a small proportion of the slide.

Figures

Figures reproduced from arXiv: 2411.18225 by the authors.

Figure 1
Figure 1. Overview of our novel method, PATHS, which predicts a patient’s relative hazard level given a whole slide image using a [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Architecture of the contextualisation module, which ac [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Inference speed, including I/O, patch pre-processing [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Left: whole slide images from the CAMELYON17 [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Number of patches loaded per slide for ABMIL (blue) [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]
Figure 6
Figure 6. Figure 6: Left-to-right: whole slide image from TCGA-BRCA, predicted semantic segmentation, PATHS heatmap. We observe that PATHS [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]
Figure 7
Figure 7. Figure 7: Further region of interest examples from CAMELYON17, with human-annotated tumourous regions marked in blue on the left of [PITH_FULL_IMAGE:figures/full_fig_p014_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

35 extracted references · 31 canonical work pages

  1. [6]

    Chen, Chengkuan Chen, Yicong Li, Tiffany Y

    Richard J. Chen, Chengkuan Chen, Yicong Li, Tiffany Y . Chen, Andrew D. Trister, Rahul G. Krishnan, and Faisal Mahmood. Scaling vision transformers to gigapixel images via hierarchical self-supervised learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 16144–16155, 2022. 1, 3, 5, 6, 2

    work page 2022

  2. [1]

    Structured crowdsourcing enables convolu- tional segmentation of histology images

    Mohamed Amgad, Habiba Elfandy, Hagar Hussein, Lamees A Atteya, Mai A T Elsebaie, Lamia S Abo Elnasr, Rokia A Sakr, Hazem S E Salem, Ahmed F Ismail, Anas M Saad, and et al. Structured crowdsourcing enables convolu- tional segmentation of histology images. Bioinformatics, 35 (18):3461–3467, 2019. 2

    work page 2019

  3. [2]

    Multiple instance learning: A survey of problem characteristics and applications

    Marc-Andr ´e Carbonneau, Veronika Cheplygina, Eric Granger, and Ghyslain Gagnon. Multiple instance learning: A survey of problem characteristics and applications. Pattern Recognition, 77:329–353, 2018. 1, 2

    work page 2018

  4. [3]

    Emerg- ing properties in self-supervised vision transformers

    Mathilde Caron, Hugo Touvron, Ishan Misra, Herv ´e J´egou, Julien Mairal, Piotr Bojanowski, and Armand Joulin. Emerg- ing properties in self-supervised vision transformers. In Pro- ceedings of the International Conference on Computer Vi- sion (ICCV), 2021. 3, 5

    work page 2021

  5. [4]

    Hi- erarchical perceiver.arXiv preprint arXiv:2202.10890, 2022

    Joao Carreira, Skanda Koppula, Daniel Zoran, Adria Re- casens, Catalin Ionescu, Olivier Henaff, Evan Shelhamer, Relja Arandjelovic, Matt Botvinick, Oriol Vinyals, et al. Hi- erarchical perceiver.arXiv preprint arXiv:2202.10890, 2022. 1, 3

    arXiv 2022

  6. [5]

    Chen, Ming Y

    Richard J. Chen, Ming Y . Lu, Wei-Hung Weng, Tiffany Y . Chen, Drew F.K. Williamson, Trevor Manz, Maha Shady, and Faisal Mahmood. Multimodal co-attention transformer for survival prediction in gigapixel whole slide images. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 4015–4025, 2021. 2, 5

    work page 2021

  7. [7]

    Towards a general-purpose foundation model for com- putational pathology

    Richard J Chen, Tong Ding, Ming Y Lu, Drew FK Williamson, Guillaume Jaume, Bowen Chen, Andrew Zhang, Daniel Shao, Andrew H Song, Muhammad Shaban, et al. Towards a general-purpose foundation model for com- putational pathology. Nature Medicine, 2024. 6, 1, 2, 3

    work page 2024

  8. [8]

    Differentiable patch selection for image recognition

    Jean-Baptiste Cordonnier, Aravindh Mahendran, Alexey Dosovitskiy, Dirk Weissenborn, Jakob Uszkoreit, and Thomas Unterthiner. Differentiable patch selection for image recognition. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 2351– 2360, 2021. 3

    work page 2021

Show all 35 references
  1. [9]

    Li, and Li Fei-Fei

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, K. Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. 2009 IEEE Conference on Computer Vision and Pattern Recognition, pages 248–255, 2009. 3

    2009

  2. [10]

    Neofytos Dimitriou, Ognjen Arandjelovi ´c, and Peter D. Caie. Deep learning for whole slide image analysis: An overview. Frontiers in Medicine, 6, 2019. 1

    2019

  3. [11]

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

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition ...

    2021

  4. [12]

    Multimodal dynamics: Dynamical fusion for trustworthy multimodal classification

    Zongbo Han, Fan Yang, Junzhou Huang, Changqing Zhang, and Jianhua Yao. Multimodal dynamics: Dynamical fusion for trustworthy multimodal classification. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 20707–20717, 2022. 5

    2022

  5. [13]

    Zhang, Shaoqing Ren, and Jian Sun

    Kaiming He, X. Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. 2016 IEEE Confer- ence on Computer Vision and Pattern Recognition (CVPR) , pages 770–778, 2015. 1, 3

    2016

  6. [14]

    Registration-enhanced mul- tiple instance learning for cervical cancer whole slide image classification

    Qiming He, Chengjiang Wang, Siqi Zeng, Zhendong Liang, Hufei Duan, Jingying Yang, Feiyang Pan, Yonghong He, Wenting Huang, and Tian Guan. Registration-enhanced mul- tiple instance learning for cervical cancer whole slide image classification. International Journal of Imaging S...

    2024

  7. [15]

    Attention-based deep multiple instance learning

    Maximilian Ilse, Jakub Tomczak, and Max Welling. Attention-based deep multiple instance learning. InProceed- ings of the 35th International Conference on Machine Learn- ing, pages 2127–2136. PMLR, 2018. 1, 2, 5, 6

    2018

  8. [16]

    Tomczak, and Max Welling

    Maximilian Ilse, Jakub M. Tomczak, and Max Welling. Deep multiple instance learning for digital histopathology. In Handbook of Medical Image Computing and Computer As- sisted Intervention, pages 521–546. Academic Press, 2020. 1, 2

    2020

  9. [17]

    Eliceiri

    Bin Li, Yin Li, and Kevin W. Eliceiri. Dual-stream mul- tiple instance learning network for whole slide image clas- sification with self-supervised contrastive learning. 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 14313–14323, 2020. 3, 5, 6, 8

    2021

  10. [18]

    Sisk, Huihui Ye, William D

    Jiayun Li, Wenyuan Li, Anthony E. Sisk, Huihui Ye, William D. Wallace, W. Speier, and Corey W. Arnold. A multi-resolution model for histopathology image classifica- tion and localization with multiple instance learning. Com- puters in biology and medicine , 131:104253, 2020. 3...

    2020

  11. [19]

    Graph cnn for survival analysis on whole slide pathological images

    Ruoyu Li, Jiawen Yao, Xinliang Zhu, Yeqing Li, and Jun- zhou Huang. Graph cnn for survival analysis on whole slide pathological images. In International Conference on Med- ical Image Computing and Computer-Assisted Intervention ,

  12. [20]

    1399 h&e-stained sentinel lymph node sections of breast cancer patients: the camelyon dataset

    Geert Litjens, Peter Bandi, Babak Ehteshami Bejnordi, Os- car Geessink, Maschenka Balkenhol, Peter Bult, Altuna Halilovic, Meyke Hermsen, Rob van de Loo, Rob V ogels, Quirine F Manson, Nikolas Stathonikos, Alexi Baidoshvili, Paul van Diest, Carla Wauters, Marcory van Dijk, and...

    2018

  13. [21]

    Swin transformer: Hierarchical vision transformer using shifted windows

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2021. 1, 2, 3

    2021

  14. [22]

    Lu, Drew F

    Ming Y . Lu, Drew F. K. Williamson, Tiffany Y . Chen, Richard J. Chen, Matteo Barbieri, and Faisal Mahmood. 9 Data-efficient and weakly supervised computational pathol- ogy on whole-slide images. Nature Biomedical Engineering, 5:555 – 570, 2020. 1, 5

    2020

  15. [23]

    A threshold selection method from gray- level histograms

    Nobuyuki Otsu. A threshold selection method from gray- level histograms. IEEE Transactions on Systems, Man, and Cybernetics, 9(1):62–66, 1979. 1

    1979

  16. [24]

    TIAToolbox as an end-to-end li- brary for advanced tissue image analytics

    Johnathan Pocock, Simon Graham, Quoc Dang Vu, Mostafa Jahanifar, Srijay Deshpande, Giorgos Hadjigeorghiou, Adam Shephard, Raja Muhammad Saad Bashir, Mohsin Bilal, Wenqi Lu, David Epstein, Fayyaz Minhas, Nasir M Rajpoot, and Shan E Ahmed Raza. TIAToolbox as an end-to-end li- br...

    2022

  17. [25]

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

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U- net: Convolutional networks for biomedical image segmen- tation. In Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015 , pages 234–241, Cham, 2015. Springer International Publishing. 2

    2015

  18. [26]

    Transmil: Transformer based correlated multiple instance learning for whole slide image classification

    Zhuchen Shao, Hao Bian, Yang Chen, Yifeng Wang, Jian Zhang, Xiangyang Ji, et al. Transmil: Transformer based correlated multiple instance learning for whole slide image classification. Advances in Neural Information Processing Systems, 34:2136–2147, 2021. 1, 2, 5

    2021

  19. [27]

    Very deep convo- lutional networks for large-scale image recognition

    Karen Simonyan and Andrew Zisserman. Very deep convo- lutional networks for large-scale image recognition. CoRR, abs/1409.1556, 2014. 1

    2014 arXiv

  20. [28]

    Differentiable zooming for multiple instance learn- ing on whole-slide images

    Kevin Thandiackal, Boqi Chen, Pushpak Pati, Guillaume Jaume, Drew FK Williamson, Maria Gabrani, and Orcun Goksel. Differentiable zooming for multiple instance learn- ing on whole-slide images. In The European Conference on Computer Vision (ECCV), 2022. 3, 5, 6, 8, 2

    2022

  21. [29]

    Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N

    Ashish Vaswani, Noam M. Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. Attention is all you need. In Neural Infor- mation Processing Systems, 2017. 2, 5

    2017

  22. [30]

    Steiner, Zhaoyang Xu, Apaar Sad- hwani, Hongwu Wang, Isabelle Flament-Auvigne, Craig H

    Ellery Wulczyn, David F. Steiner, Zhaoyang Xu, Apaar Sad- hwani, Hongwu Wang, Isabelle Flament-Auvigne, Craig H. Mermel, Po-Hsuan Cameron Chen, Yun Liu, and Martin C. Stumpe. Deep learning-based survival prediction for multi- ple cancer types using histopathology images. PLOS ...

    2020

  23. [31]

    Nystr¨omformer: A nystr¨om-based algorithm for approximat- ing self-attention

    Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Moo Fung, Yin Li, and Vikas Singh. Nystr¨omformer: A nystr¨om-based algorithm for approximat- ing self-attention. Proceedings of the ... AAAI Conference on Artificial Intelligence. AAAI Conference on Artif...

    2021

  24. [32]

    Whole slide images based cancer survival prediction using attention guided deep multi- ple instance learning networks

    Jiawen Yao, Xinliang Zhu, Jitendra Jonnagaddala, Nicholas J Hawkins, and Junzhou Huang. Whole slide images based cancer survival prediction using attention guided deep multi- ple instance learning networks. Medical image analysis, 65: 101789, 2020. 3, 5, 6

    2020

  25. [33]

    Bias in Cross- Entropy-Based training of deep survival networks

    Shekoufeh Gorgi Zadeh and Matthias Schmid. Bias in Cross- Entropy-Based training of deep survival networks. IEEE Trans Pattern Anal Mach Intell, 43(9):3126–3137, 2021. 5

    2021

  26. [34]

    Arik, and Tomas Pfister

    Zizhao Zhang, Han Zhang, Long Zhao, Ting Chen, Ser- can O. Arik, and Tomas Pfister. Nested hierarchical trans- former: Towards accurate, data-efficient and interpretable vi- sual understanding. In AAAI Conference on Artificial Intelli- gence (AAAI), 2022. 1, 3

    2022

  27. [35]

    Predicting lymph node metastasis using histopathological images based on multiple instance learning with deep graph convolution

    Yu Zhao, Fan Yang, Yuqi Fang, Hailing Liu, Niyun Zhou, Jun Zhang, Jiarui Sun, Sen Yang, Bjoern H Menze, Xinjuan Fan, and Jianhua Yao. Predicting lymph node metastasis using histopathological images based on multiple instance learning with deep graph convolution. 2020 IEEE/CVF ...

    2020

Pith tools

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