Enhancing breast cancer detection on screening mammogram using self-supervised learning and a hybrid deep model of Swin Transformer and Convolutional Neural Network

Anne L. Martel; Han Chen

REVIEW 3 major objections 5 minor 36 references

Enhancing breast cancer detection on screening mammogram using self-supervised learning and a hybrid deep model of Swin Transformer and Convolutional Neural Network

T0 review · 3 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read A Swin Transformer pretrained with self-supervised learning and paired with a CNN branch distinguishes malignant from benign screening mammograms, reaching AUC 0.864 on CMMD and 0.889 on INbreast.

desk verdict Adding EsViT pretraining to a hybrid Swin/CNN mammogram model gives a modest AUC bump, but the paper's statistical evidence for the bump is shaky; the architecture itself is sound and worth reviewing. read the letter →

arxiv 2504.19888 v1 pith:RD4B5L74 submitted 2025-04-28 cs.CV

Han Chen , Anne L. Martel This is my paper · ORCID

classification cs.CV

keywords breastcancerdetectionscreeningmammogramself-supervisedlearningEsViTSwinTransformerhybriddeepmodelsaliencymapweaklysupervisedlocalization

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to show that self-supervised pretraining on unlabeled mammograms can substitute for large quantities of labeled data when building a breast-cancer classifier. The authors pretrain a Swin Transformer with the EsViT method on full-resolution screening mammograms, then attach a CNN branch that reads six patches chosen by a saliency map, calling the combined network HybMNet. On the CMMD test set the model reaches AUC 0.864 (95% CI 0.852, 0.875), and on INbreast 0.889 (95% CI 0.875, 0.903), beating both ImageNet-initialized HybMNet and GMIC fine-tuned on a large mammogram corpus. The CMMD AUC improvement from SSL is reported as statistically significant (paired t-test p=0.048); the INbreast gain is not (p=0.32), which the paper attributes to domain shift.

What carries the argument

The load-bearing object is the saliency-map-guided ROI patch selector. A 1x1 convolution plus sigmoid turns the Swin-T feature map at 1024x1024 input resolution into a probability map; a greedy algorithm repeatedly picks the 6 patches with the largest summed saliency, zeroes the already-selected locations so the same region is not chosen twice, and crops those patches from the original mammogram for the CNN branch. This mechanism is what converts image-level labels into weakly supervised local evidence, and the paper's claimed benefit of the hybrid fusion over the Swin-T-only backbone (AUC 0.856 vs 0.812 in the ablation) depends on those patches being informative.

What would settle it

Compare the six selected ROI patches against the pixel-level lesion annotations in INbreast: if on malignant exams the patches rarely overlap the annotated mass or calcifications, the claimed role of the CNN branch is not supported. A complementary check is to swap the saliency selector for random or fixed-grid patches and see whether AUC stays at 0.889; if it does, the saliency map is not the mechanism.

Watch

Extended reading notes

Core claim

On its own terms, the paper's finding is that an SSL pretraining stage is enough to make a hybrid transformer-CNN classifier competitive on mammograms when labeled examples are scarce. The architecture is HybMNet: a Swin Transformer processes the whole 1024x1024 mammogram with 16x16 windows, a 1x1 convolution turns its feature map into a saliency map, six high-saliency patches are cropped at full resolution, a ResNet-18 extracts local features from them, and a fusion layer concatenates global and local features for the final prediction. With EsViT pretraining on unlabeled CMMD images, the model reaches AUC 0.864 on CMMD and 0.889 on INbreast; without SSL it reaches 0.856 and 0.877. The paper claims the CMMD gain is significant and treats the INbreast gain as small but still transferable, noting the pretraining and downstream datasets differ in intensity distribution.

Load-bearing premise

The method depends on the saliency map's pinpointing the diagnostically relevant tissue so that the six greedily chosen patches carry the fine detail the CNN branch needs; the paper supports this only with qualitative examples and no measured localization accuracy.

Editorial extensions

If this is right

On CMMD, SSL-pretrained HybMNet reaches AUC 0.864 (95% CI 0.852, 0.875), accuracy 0.797, and F1 0.790, beating ImageNet-initialized HybMNet and both GMIC variants.
On INbreast, the same pipeline reaches AUC 0.889 (95% CI 0.875, 0.903), accuracy 0.830, and F1 0.719, also above the GMIC baselines.
The ablation study reports that the full model with Swin-T, CNN, fusion, and L1 saliency regularization reaches AUC 0.856 versus 0.812 for the Swin-T branch alone, so the local branch and fusion are doing measurable work.
The paired t-test for the CMMD AUC gain from SSL is p=0.048; the same comparison on INbreast is p=0.32, which the paper attributes to domain shift between the CMMD pretraining set and INbreast images.

Reading between the lines

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

The paper does not quantify how often the six saliency-selected patches overlap the ground-truth lesion boundaries available in INbreast; measuring that overlap would directly test the localization assumption.
If the saliency selection is the key, then replacing the six greedy patches with six randomly chosen breast-tissue patches should lower accuracy; if it does not, the CNN branch is acting mainly as an ensemble member rather than a focused local reader.
A natural extension is to apply the same two-stage recipe to digital breast tomosynthesis, where labeled exams are even scarcer and the high-resolution local patches would come from different slices.
The paper's own numbers imply the cleanest test of transfer is linear probing the EsViT-pretrained backbone on INbreast without the CNN branch; a backbone that shows no feature gain there would locate the benefit in the downstream hybrid rather than in SSL itself.

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.

Desk Editor's Note

read the letter

The genuinely new bit here is applying EsViT self-supervised pretraining to whole high-resolution mammograms (512x512, with 1024x1024 downstream) and then slotting the pretrained Swin-T into the existing HybMNet hybrid. That combination is not in the prior IWBI paper, and the idea is reasonable: use unlabeled mammograms to adapt a large transformer before fine-tuning on small labeled sets. The evaluation is above average for this subfield: public datasets, patient-level splits, confidence intervals, an ablation, and a comparison against GMIC. The ablation shows the full model clearly beats the Swin-T-only backbone (AUC 0.856 vs 0.812), so the hybrid architecture itself seems to do real work.

The soft spots are mostly about the headline claim that SSL helps. The CMMD gain is 0.856 to 0.864, and the only statistical support is a paired t-test on five overlapping 80% resamples of the test set, with patient clustering ignored. That p=0.048 does not carry the weight the paper puts on it; a clustered or bootstrap-based test could easily flip it. The CIs overlap, which is not itself decisive but reinforces the fragility. On INbreast the AUC gain is 0.877 to 0.889 but the p-value is 0.32, yet the paper calls this evidence of robustness and cross-dataset transfer. That is overreading. The 'first to apply SSL to high-resolution mammograms' claim is also unsupported, since no comparison to prior SSL mammography work at lower resolutions or with other SSL methods is made. The saliency-map localization concern is real but secondary; even imperfect patch selection could add signal, and the architecture's gain over the backbone does not depend on the SSL comparison. Code is not public, which is a minor negative.

In short: the engineering is plausible, the ablation is informative, but the central quantitative claim about SSL is not convincingly established as written. This is a fixable problem — redo the significance testing properly, temper the language, add a real comparison to existing SSL approaches — not a broken method. The paper deserves a serious referee because the question matters and the core setup is honest and reproducible in principle. I would not cite it myself until the statistical claim is reworked, but I'd bring it to reading group as an example of how easy it is to overstate small AUC gains in medical imaging.

Referee Report

3 major / 5 minor

Summary. The paper proposes HybMNet, a hybrid deep model combining a Swin Transformer backbone with a ResNet-18 CNN branch for breast cancer classification on screening mammograms. The Swin Transformer is first pretrained with the EsViT self-supervised method on unlabeled CMMD mammograms, then fine-tuned together with the CNN branch in an end-to-end manner. A 1x1 convolutional layer on the Swin-T feature map generates a saliency map used to select six ROI patches, which are cropped from the high-resolution input and processed by the CNN branch; global and local features are fused for the final prediction. The method is evaluated on two public datasets, CMMD and INbreast, reporting AUC of 0.864 (95% CI: 0.852, 0.875) on CMMD and 0.889 (95% CI: 0.875, 0.903) on INbreast. The central quantitative claim is that self-supervised pretraining improves AUC over ImageNet-only initialization, supported by a paired t-test p-value of 0.048 on CMMD, while the corresponding INbreast comparison is not statistically significant (p=0.32).

Significance. If the results hold, the paper would provide evidence that self-supervised pretraining on a relatively small unlabeled mammogram dataset can improve downstream classification performance compared with ImageNet-only initialization, and that a hybrid transformer-CNN architecture with weakly supervised ROI selection can be effective for screening mammography. The study has several strengths: patient-level train/test splits on both public datasets, confidence intervals, an ablation study of the model components, and comparison against a strong GMIC baseline pretrained on a large mammogram corpus. However, the statistical evidence for the headline SSL gain is fragile, resting on a borderline p-value obtained from a poorly described resampling procedure that ignores patient-level clustering; the cross-dataset claim is based on a non-significant result; and the ROI localization mechanism is validated only qualitatively. The code is not released, limiting reproducibility. These issues make the paper's quantitative foundation weaker than its presentation suggests.

major comments (3)

[Section 3.1 and Table 3] The manuscript states that CMMD confidence intervals are obtained by randomly sampling 80% of the test set five times, and then reports a paired t-test p-value of 0.048 for the AUC difference between HybMNet and HybMNetssl. The paper does not state the unit of this random sampling or the number of pairs used in the t-test. If the five 80% resamples are the paired observations, the test has only four degrees of freedom, the resamples overlap heavily, and the independence assumption is violated. If the resampling is at image level, the 526 test images from 186 patients are not independent because images from the same patient are correlated. The overlapping confidence intervals for the two models (0.846-0.866 vs 0.852-0.875) are consistent with a fragile difference. The authors should replace this analysis with a valid patient-level clustered test, such as a patient-level bootstrap or permutation test, and clearly report the test procedure, the number of resamples, and the resulting p-value and effect size.
[Section 3.2, Table 4, and Discussion] The INbreast AUC comparison between HybMNet and HybMNetssl yields p=0.32, which is not statistically significant. Nevertheless, the text states that 'the marginal increase in AUC achieved by HybMNetssl still demonstrates that SSL pretraining in our approach can transfer across datasets,' and the Discussion asserts that HybMNetssl 'is expected to have better generalization and robustness.' A non-significant result cannot demonstrate transferability or robustness. The authors should either provide a properly powered cross-dataset analysis with an effect-size estimate and confidence interval, or temper these claims to state that no significant transfer benefit was observed on INbreast.
[Section 2.3.2, Section 3.4, and Table 5] The ROI-selection mechanism is central to the claimed benefit of the CNN branch, but the only validation is qualitative (Fig. 5). The ablation in Table 5 shows that adding the CNN branch improves AUC from 0.812 to 0.856 on CMMD, which is encouraging, but no quantitative localization metric is reported. Because INbreast provides pixel-level annotations, the authors could quantify how often the selected patches overlap with annotated masses or calcifications. Without such a measure, the generalizability of the saliency-map-based selection, and hence of the hybrid architecture, remains untested.

minor comments (5)

[Fig. 1 caption and Section 2.2.1] The figure caption contains a typo: 'Swi-T' should be 'Swin-T'.
[Section 2.3.1] The phrase 'without sacrificing spatial resolution' is misleading because Swin-T uses hierarchical patch merging that progressively reduces spatial resolution; the sentence should be rephrased to describe the trade-off more accurately.
[Section 3.2] The claim that Fig. 2 shows 'a notably more stable F1-score' is not supported by the overlapping confidence intervals; either report interval widths or a formal comparison of variability.
[References] Several references contain stray characters in the year field, such as 'arXiv preprint arXiv:2106.09785 0 (2021)', 'arXiv preprint arXiv:1711.05775 0 (2017)', and 'arXiv preprint arXiv:1711.05101 1 (2017)'; these appear to be citation artifacts and should be cleaned.
[Section 7 (Code and Data Availability)] The statement that code is available 'upon reasonable request' without a repository limits reproducibility; for a methods-focused journal, releasing code or providing a detailed implementation appendix would strengthen the paper.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the empirical evaluation is self-contained, with held-out test sets and external baselines.

full rationale

This paper reports an empirical machine-learning evaluation rather than a mathematical derivation. The central claims are measured AUC values for HybMNetssl on CMMD and INbreast, and a comparison against HybMNet initialized with ImageNet-1K weights. These metrics are computed on test sets that are disjoint from the training and pretraining sets at the patient level, as described in Section 2.1 and Tables 1-2. The loss function in Eq. (1) defines the training objective, but it does not define the evaluation metrics or the reported AUC values. No equation in the paper reduces a predicted quantity to a fitted parameter or to the claim being tested. The greedy saliency-map patch selection and fusion are architectural choices that are ablated in Table 5, and the comparison includes an external baseline, GMIC. The only self-citation, reference [25] to the authors' prior HybMNet work, is used as the baseline architecture and is fully described and re-evaluated in the present paper, so it is not load-bearing circularity. The statistical concern raised in the skeptic analysis, that the paired t-test p=0.048 for the SSL improvement may be unreliable because the confidence intervals were computed from overlapping 80% resamples of the test set, is a correctness and statistical-validity concern, not a circularity concern. The p-value is not constructed from the claim it supports; it is an empirical comparison. Therefore, no circular step can be identified under the required standard of quoting an equation or parameter that reduces the result to its inputs.

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

The central claim rests primarily on domain assumptions about label validity, resolution preservation, and saliency-based localization. The only tuned scalar entering the loss is alpha; the number of ROI patches and the Swin-T window size are additional hand choices. No new physical entities are introduced.

free parameters (3)

L1 regularization weight alpha = 3.26e-6
Introduced in Eq. (1) to sparsify the saliency map; value set by hand with no reported search or sensitivity analysis. Central to the saliency-based ROI selection.
Number of ROI patches = 6
Section 2.3.2: the greedy algorithm selects 6 patches per image; this determines how much local information the CNN branch sees and is chosen without an ablation over patch count.
Swin-T window size during pretraining = 16
Section 2.2.2: window size increased from 7 to 16 following ref 30; affects the model's receptive field and is a hand choice.

assumptions (4)

domain assumption BI-RADS 1-2 are negative, BI-RADS 4-6 are positive, BI-RADS 3 excluded, on INbreast (ref 29).
Section 2.1: this mapping converts INbreast to benign/malignant labels; if the mapping is inaccurate, the reported INbreast AUC does not measure true benign/malignant discrimination.
ad hoc to paper The saliency map from a 1x1 conv on the Swin-T feature map localizes the informative regions, so six greedily selected patches contain the diagnostic detail.
Section 2.3.2: the entire CNN branch and fusion depend on this. The paper shows only qualitative examples, no quantitative localization validation.
domain assumption Downsampling mammograms to 1024x1024 (and 512x512 for SSL) preserves the fine-grained information (micro-calcifications, masses) needed for classification.
Section 3.1: native INbreast resolution is 3328x4084 or 2560x3328; all inputs are resized. If clinically decisive details are lost in downsampling, the reported results may not transfer to full-resolution screening.
domain assumption EsViT's non-contrastive SSL objective, with 30 epochs of pretraining on 3,832 CMMD images, learns representations that transfer to downstream classification.
Section 3.1: the method relies on this transfer. The CMMD gain is small and the INbreast gain is not statistically significant, so the strength of this assumption is only partially supported.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Enhancing breast cancer detection on screening mammogram using self-supervised learning and a hybrid deep model of Swin Transformer and Convolutional Neural Network." pith.science (2026). https://pith.science/paper/RD4B5L74

@misc{pith2026250419888,
  author       = {Pith},
  title        = {Pith review of: Enhancing breast cancer detection on screening mammogram using self-supervised learning and a hybrid deep model of Swin Transformer and Convolutional Neural Network},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RD4B5L74}},
  note         = {Machine review of arXiv:2504.19888}
}

read the original abstract

Purpose: The scarcity of high-quality curated labeled medical training data remains one of the major limitations in applying artificial intelligence (AI) systems to breast cancer diagnosis. Deep models for mammogram analysis and mass (or micro-calcification) detection require training with a large volume of labeled images, which are often expensive and time-consuming to collect. To reduce this challenge, we proposed a novel method that leverages self-supervised learning (SSL) and a deep hybrid model, named \textbf{HybMNet}, which combines local self-attention and fine-grained feature extraction to enhance breast cancer detection on screening mammograms. Approach: Our method employs a two-stage learning process: (1) SSL Pretraining: We utilize EsViT, a SSL technique, to pretrain a Swin Transformer (Swin-T) using a limited set of mammograms. The pretrained Swin-T then serves as the backbone for the downstream task. (2) Downstream Training: The proposed HybMNet combines the Swin-T backbone with a CNN-based network and a novel fusion strategy. The Swin-T employs local self-attention to identify informative patch regions from the high-resolution mammogram, while the CNN-based network extracts fine-grained local features from the selected patches. A fusion module then integrates global and local information from both networks to generate robust predictions. The HybMNet is trained end-to-end, with the loss function combining the outputs of the Swin-T and CNN modules to optimize feature extraction and classification performance. Results: The proposed method was evaluated for its ability to detect breast cancer by distinguishing between benign (normal) and malignant mammograms. Leveraging SSL pretraining and the HybMNet model, it achieved AUC of 0.864 (95% CI: 0.852, 0.875) on the CMMD dataset and 0.889 (95% CI: 0.875, 0.903) on the INbreast dataset, highlighting its effectiveness.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 33 canonical work pages

[1]

H. Sung, J. Ferlay, R. L. Siegel, et al. , ``Global cancer statistics 2020: Globocan estimates of incidence and mortality worldwide for 36 cancers in 185 countries,'' CA: a cancer journal for clinicians 71 (3), 209--249 (2021)

work page 2021
[2]

Pashayan, A

N. Pashayan, A. C. Antoniou, U. Ivanus, et al. , ``Personalized early detection and prevention of breast cancer: Envision consensus statement,'' Nature Reviews Clinical Oncology 17 (11), 687--705 (2020)

work page 2020
[3]

D. A. Zebari, D. A. Ibrahim, D. Q. Zeebaree, et al. , ``Systematic review of computing approaches for breast cancer detection based computer aided diagnosis using mammogram images,'' Applied Artificial Intelligence 35 (15), 2157--2203 (2021)

work page 2021
[4]

S. W. Duffy, L. Tab \'a r, A. M.-F. Yen, et al. , ``Beneficial effect of consecutive screening mammography examinations on mortality from breast cancer: a prospective study,'' Radiology 299 (3), 541--547 (2021)

work page 2021
[5]

Hovda, S

T. Hovda, S. R. Hoff, M. Larsen, et al. , ``True and missed interval cancer in organized mammographic screening: a retrospective review study of diagnostic and prior screening mammograms,'' Academic Radiology 29 , S180--S191 (2022)

work page 2022
[6]

S. M. McKinney, M. Sieniek, V. Godbole, et al. , ``International evaluation of an ai system for breast cancer screening,'' Nature 577 (7788), 89--94 (2020)

work page 2020
[7]

F. Yan, H. Huang, W. Pedrycz, et al. , ``Automated breast cancer detection in mammography using ensemble classifier and feature weighting algorithms,'' Expert Systems with Applications 227 , 120282 (2023)

work page 2023
[8]

Schaffter, D

T. Schaffter, D. S. Buist, C. I. Lee, et al. , ``Evaluation of combined artificial intelligence and radiologist assessment to interpret screening mammograms,'' JAMA network open 3 (3), e200265--e200265 (2020)

work page 2020

Show all 36 references

[9]

Lotter, A

W. Lotter, A. R. Diab, B. Haslam, et al. , ``Robust breast cancer detection in mammography and digital breast tomosynthesis using an annotation-efficient deep learning approach,'' Nature Medicine 27 (2), 244--249 (2021)

2021
[10]

Samek, G

W. Samek, G. Montavon, S. Lapuschkin, et al. , ``Explaining deep neural networks and beyond: A review of methods and applications,'' Proceedings of the IEEE 109 (3), 247--278 (2021)

2021
[11]

H. Chen, Y. Jiang, H. Ko, et al. , ``A teacher--student framework with fourier transform augmentation for covid-19 infection segmentation in ct images,'' Biomedical Signal Processing and Control 79 , 104250 (2023)

2023
[12]

H. Chen, Y. Jiang, M. Loew, et al. , ``Unsupervised domain adaptation based covid-19 ct infection segmentation network,'' Applied Intelligence 52 (6), 6340--6353 (2022)

2022
[13]

H. Chen, Y. Jiang, and H. Ko, ``Pose-guided graph convolutional networks for skeleton-based action recognition,'' IEEE Access 10 , 111725--111731 (2022)

2022
[14]

Y. Shen, N. Wu, J. Phang, et al. , ``An interpretable classifier for high-resolution breast cancer screening images utilizing weakly supervised localization,'' Medical image analysis 68 , 101908 (2021)

2021
[15]

Ribli, A

D. Ribli, A. Horv \'a th, Z. Unger, et al. , ``Detecting and classifying lesions in mammograms with deep learning,'' Scientific reports 8 (1), 4165 (2018)

2018
[16]

Rangarajan, A

K. Rangarajan, A. Gupta, S. Dasgupta, et al. , ``Ultra-high resolution, multi-scale, context-aware approach for detection of small cancers on mammography,'' Scientific reports 12 (1), 11622 (2022)

2022
[17]

J. Guo, K. Han, H. Wu, et al. , ``Cmt: Convolutional neural networks meet vision transformers,'' in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 12175--12185 (2022)

2022
[18]

S. M. Pinto Pereira, V. A. McCormack, S. M. Moss, et al. , ``The spatial distribution of radiodense breast tissue: a longitudinal study,'' Breast Cancer Research 11 (3), 1--12 (2009)

2009
[19]

Wei, H.-P

J. Wei, H.-P. Chan, Y.-T. Wu, et al. , ``Association of computerized mammographic parenchymal pattern measure with breast cancer risk: a pilot case-control study,'' Radiology 260 (1), 42--49 (2011)

2011
[20]

Z. Liu, Y. Lin, Y. Cao, et al. , ``Swin transformer: Hierarchical vision transformer using shifted windows,'' in Proceedings of the IEEE/CVF international conference on computer vision , 10012--10022 (2021)

2021
[21]

Jing and Y

L. Jing and Y. Tian, ``Self-supervised visual feature learning with deep neural networks: A survey,'' IEEE transactions on pattern analysis and machine intelligence 43 (11), 4037--4058 (2020)

2020
[22]

Caron, H

M. Caron, H. Touvron, I. Misra, et al. , ``Emerging properties in self-supervised vision transformers,'' in Proceedings of the IEEE/CVF international conference on computer vision , 9650--9660 (2021)

2021
[23]

C. Li, J. Yang, P. Zhang, et al. , ``Efficient self-supervised vision transformers for representation learning,'' arXiv preprint arXiv:2106.09785 0 (2021)

2021 arXiv
[24]

Grill, F

J.-B. Grill, F. Strub, F. Altch \'e , et al. , ``Bootstrap your own latent-a new approach to self-supervised learning,'' Advances in neural information processing systems 33 , 21271--21284 (2020)

2020
[25]

Chen and A

H. Chen and A. L. Martel, ``Towards improved breast cancer detection on digital mammograms using local self-attention-based transformer,'' in 17th International Workshop on Breast Imaging (IWBI 2024) , 13174 , 455--461, SPIE (2024)

2024
[26]

H. Cai, J. Wang, T. Dan, et al. , ``An online mammography database with biopsy confirmed types,'' Scientific Data 10 (1), 123 (2023)

2023
[27]

Clark, B

K. Clark, B. Vendt, K. Smith, et al. , ``The cancer imaging archive (tcia): maintaining and operating a public information repository,'' Journal of digital imaging 26 , 1045--1057 (2013)

2013
[28]

I. C. Moreira, I. Amaral, I. Domingues, et al. , ``Inbreast: toward a full-field digital mammographic database,'' Academic radiology 19 (2), 236--248 (2012)

2012
[29]

Shen, ``End-to-end training for whole image breast cancer diagnosis using an all convolutional design,'' arXiv preprint arXiv:1711.05775 0 (2017)

L. Shen, ``End-to-end training for whole image breast cancer diagnosis using an all convolutional design,'' arXiv preprint arXiv:1711.05775 0 (2017)

2017 arXiv
[30]

P. Ren, C. Li, G. Wang, et al. , ``Beyond fixation: Dynamic window visual transformer,'' in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 11987--11997 (2022)

2022
[31]

D. Ma, M. R. Hosseinzadeh Taher, J. Pang, et al. , ``Benchmarking and boosting transformers for medical image classification,'' in MICCAI Workshop on Domain Adaptation and Representation Transfer , 12--22, Springer (2022)

2022
[32]

K. He, X. Zhang, S. Ren, et al. , ``Deep residual learning for image recognition,'' in Proceedings of the IEEE conference on computer vision and pattern recognition , 770--778 (2016)

2016
[33]

Loshchilov and F

I. Loshchilov and F. Hutter, ``Decoupled weight decay regularization,'' arXiv preprint arXiv:1711.05101 1 (2017)

2017 arXiv
[34]

Paszke, S

A. Paszke, S. Gross, F. Massa, et al. , ``Pytorch: An imperative style, high-performance deep learning library,'' Advances in neural information processing systems 32 (2019)

2019
[35]

V. Rani, M. Kumar, A. Gupta, et al. , ``Self-supervised learning for medical image analysis: a comprehensive review,'' Evolving Systems 15 (4), 1607--1633 (2024)

2024
[36]

write newline

" write newline "" before.all 'output.state := FUNCTION blank.sep after.quote 'output.state := FUNCTION fin.entry output.state after.quoted.block = 'skip 'add.period if write newline FUNCTION new.block output.state before.all = 'skip output.state after.quote = after.quoted.blo...

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.

[1] [1]

H. Sung, J. Ferlay, R. L. Siegel, et al. , ``Global cancer statistics 2020: Globocan estimates of incidence and mortality worldwide for 36 cancers in 185 countries,'' CA: a cancer journal for clinicians 71 (3), 209--249 (2021)

work page 2021

[2] [2]

Pashayan, A

N. Pashayan, A. C. Antoniou, U. Ivanus, et al. , ``Personalized early detection and prevention of breast cancer: Envision consensus statement,'' Nature Reviews Clinical Oncology 17 (11), 687--705 (2020)

work page 2020

[3] [3]

D. A. Zebari, D. A. Ibrahim, D. Q. Zeebaree, et al. , ``Systematic review of computing approaches for breast cancer detection based computer aided diagnosis using mammogram images,'' Applied Artificial Intelligence 35 (15), 2157--2203 (2021)

work page 2021

[4] [4]

S. W. Duffy, L. Tab \'a r, A. M.-F. Yen, et al. , ``Beneficial effect of consecutive screening mammography examinations on mortality from breast cancer: a prospective study,'' Radiology 299 (3), 541--547 (2021)

work page 2021

[5] [5]

Hovda, S

T. Hovda, S. R. Hoff, M. Larsen, et al. , ``True and missed interval cancer in organized mammographic screening: a retrospective review study of diagnostic and prior screening mammograms,'' Academic Radiology 29 , S180--S191 (2022)

work page 2022

[6] [6]

S. M. McKinney, M. Sieniek, V. Godbole, et al. , ``International evaluation of an ai system for breast cancer screening,'' Nature 577 (7788), 89--94 (2020)

work page 2020

[7] [7]

F. Yan, H. Huang, W. Pedrycz, et al. , ``Automated breast cancer detection in mammography using ensemble classifier and feature weighting algorithms,'' Expert Systems with Applications 227 , 120282 (2023)

work page 2023

[8] [8]

Schaffter, D

T. Schaffter, D. S. Buist, C. I. Lee, et al. , ``Evaluation of combined artificial intelligence and radiologist assessment to interpret screening mammograms,'' JAMA network open 3 (3), e200265--e200265 (2020)

work page 2020