Leveraging Pathology Co-occurrence for Test-Time Adaptation in Chest X-Ray Diagnosis

Dongbin Kim; Jaewook Lee; Soyeon Park; Woojin Jeong; Yujin Choi

REVIEW 4 major objections 6 minor 21 references

Leveraging Pathology Co-occurrence for Test-Time Adaptation in Chest X-Ray Diagnosis

T0 review · 4 major / 6 minor · reviewed 2026-07-12 · grok-4.5

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

Pith's one-line read CoWA reweights test-time adaptation by how well each chest X-ray prediction matches target disease co-occurrence patterns, improving multi-label diagnosis under domain shift.

desk verdict Solid, usable multi-label TTA for chest X-ray that actually respects co-occurrence; modest gains, real circularity risk, still worth a referee. read the letter →

arxiv 2607.03715 v2 pith:MDGZXFRI submitted 2026-07-04 cs.CV

Woojin Jeong , Yujin Choi , Dongbin Kim , Soyeon Park , Jaewook Lee This is my paper · ORCID

classification cs.CV

keywords test-timeadaptationchestX-raymulti-labelclassificationdiseaseco-occurrencedomainshiftentropyminimizationsamplereweighting

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

Chest X-ray classifiers often fail when moved to a new hospital because scanners, protocols, and patient mixes differ. Test-time adaptation tries to fix this by updating the model on unlabeled images from the new site, but standard methods treat every prediction as equally trustworthy and ignore the fact that diseases tend to appear together in structured ways. CoWA estimates those co-occurrence patterns on the fly from the model’s own soft predictions, scores each sample by how well its predicted label set matches the emerging structure, and uses the scores to reweight entropy minimization. Samples that look clinically plausible drive the update; samples that violate expected combinations are downweighted. Across six public domain-shift pairs the method produces higher mean AUROC than established baselines and is the only one that never falls sharply below the unadapted model on any pathology.

What carries the argument

Co-occurrence Weighted Adaptation (CoWA): the running co-occurrence matrix M estimated from soft-then-binarized predictions, the sample weight w_i = exp(-∥m_i - M∥_F^{2}/τ), and the weighted entropy loss that multiplies each sample’s contribution by max(w_i, w_min).

What would settle it

On a held-out domain-shift pair, replace CoWA’s estimated matrix with a deliberately scrambled or source-only co-occurrence matrix and check whether the AUROC gains and the “no severe drop” property disappear.

Watch

Extended reading notes

Core claim

By accumulating a normalized co-occurrence matrix from binarized model predictions on unlabeled target images and converting each sample’s local co-occurrence pattern into a Frobenius-norm consistency weight, CoWA turns entropy minimization into a structure-aware update that improves multi-label chest X-ray performance under domain shift without source data or extra labels.

Load-bearing premise

The co-occurrence matrix built from the model’s own early, source-biased predictions on unlabeled target data is already a reliable enough map of true target disease relationships that downweighting mismatches helps rather than hurts adaptation.

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.

Desk Editor's Note

read the letter

The one thing worth knowing: CoWA is a clean, source-free fix for the fact that standard entropy TTA (TENT and friends) treats multi-label chest X-ray pathologies as independent. They accumulate a running co-occurrence matrix from binarized target predictions, score each sample by Frobenius distance of its outer product to that matrix, and reweight the entropy loss. That combination is not in the single-label TTA literature they cite, and the six-shift table shows it is the only method that never tanks below the unadapted baseline on low-prevalence classes.

What they do well is straightforward. The method is simple (BN affine only, public TorchXRayVision backbones, code promised), the analyses in Figs. 3–4 are honest (matrix drifts toward a supervised target proxy; higher weights track higher AUPRC), and the ablation on the norm choice is useful. Gains are consistent if modest; worst-case robustness is the real selling point for clinical deployment. Citations look appropriate; no inventing of prior art.

Soft spots, in proportion. The stress-test concern is real but not fatal: M is built from the same initially source-biased predictions that are being adapted, so early batches can lock in a distorted reliability signal. Accumulation plus w_min dampens this; Fig. 3 shows progressive alignment, but they never freeze or randomize the early matrix to isolate the effect. Hyperparameters (τ, binarization threshold, LR) are selected per setting with no multi-seed error bars. These are standard empirical-TTA caveats, not load-bearing contradictions. The circularity is mild because final metrics are against held-out ground truth.

This is for people who ship multi-label medical models under site shift and already know TENT/EATA. It is not a new principle, but it is a reproducible engineering step that fills a clear gap. I would send it to peer review; a referee can demand the multi-seed numbers and the early-M ablation without killing the paper. Worth engaging if you work in this lane.

Referee Report

4 major / 6 minor

Summary. The paper proposes CoWA, a source-free test-time adaptation method for multi-label chest X-ray classification. It estimates a pathology co-occurrence matrix M from the model's own soft-then-binarized target predictions (Eqs. 1–2), scores each sample by Frobenius consistency of its outer-product pattern with M (Eq. 3), and performs weighted entropy minimization over BN affine parameters (Eq. 4). The claim is that this structure-aware reweighting yields more reliable adaptation than uniform entropy methods designed for single-label natural images. Evaluation covers six source–target pairs among MIMIC-CXR, CheXpert, VinDr-CXR, and NIH using DenseNet-121 models from TorchXRayVision, reporting AUROC on six shared pathologies against AdaBN, TENT, CoTTA, EATA, and RoTTA. Supporting analyses include progressive alignment of M to a supervised target proxy (Fig. 3), correlation of weights with AUPRC (Fig. 4), and a norm ablation (Table 2).

Significance. Domain shift across clinical sites is a genuine barrier to deploying chest X-ray models, and source-free TTA is a practically relevant setting. The paper correctly identifies that standard entropy-based TTA ignores multi-label co-occurrence structure, which is a real property of thoracic disease. If the gains hold under stricter protocols, CoWA is a simple, architecture-preserving add-on that improves worst-case robustness on low-prevalence classes—clinically more important than mean AUROC alone. Strengths include public code, multiple real domain-shift pairs, and analyses that go beyond leaderboard tables (matrix evolution, weight–reliability correlation, gradient-variance check). The contribution is incremental rather than foundational, but it is a clear and usable idea for multi-label medical TTA.

major comments (4)

[Section 3, Eqs. 1–4; Fig. 3] Section 3 (Co-occurrence Matrix Estimation) and Fig. 3: M is accumulated from the same source-pretrained predictions that are being adapted, so early batches inherit source co-occurrence bias (Fig. 1). The progressive-alignment evidence compares M only to a fully supervised target proxy after adaptation has started; it does not isolate whether the early, source-contaminated M correctly ranks sample reliability. A load-bearing control is missing: freeze M after the first few batches (or replace early M with a random/source matrix) and show that the weighted objective still helps rather than locks in distorted structure. Without this, the central reliability-signal claim remains only partially supported.
[Section 4.1] Section 4.1 Implementation Details: learning rate, τ, and the binarization threshold are selected from small grids, but the selection protocol is not stated. If any selection used target labels or target validation AUROC, the comparison is not pure TTA and overstates gains relative to baselines run with original hyperparameters. Please state explicitly whether hyperparameters were chosen without target labels (e.g., fixed defaults, source-only proxy, or first-batch unsupervised criteria) and report sensitivity across the full grid for all six shifts.
[Table 1] Table 1: results appear to be single-run point estimates with no standard deviation, no multiple seeds, and no statistical tests. Several CoWA margins over the second-best method are small (e.g., 0.3–0.8 AUROC points). Given BN-only adaptation and stochastic batch order, variance can be material. Report mean±std over at least 3 seeds (or batch-order shuffles) and mark whether gains remain significant; otherwise the “consistent improvements” claim is overstated.
[Section 4.2, Table 1] Section 4.2 / Table 1 class-wise discussion: the strongest clinical claim—that CoWA is the only method avoiding severe drops below the unadapted baseline on all shifts, especially low-prevalence classes such as Pneumothorax—is important but only narrated. Please add a compact worst-case or min-class ΔAUROC summary (or per-class drop table) so this claim is directly verifiable rather than inferred from selected cells.

minor comments (6)

[Eq. (2)] Eq. (2): the normalization M_jk = P_jk / sqrt(P_jj P_kk + ε) is a cosine-style co-occurrence; briefly justify why this form is preferred over raw joint probabilities or PMI, especially under class imbalance.
[Fig. 1] Fig. 1 is referenced for cross-domain co-occurrence variation but is not quantitatively described in the text; a short numeric summary (e.g., Frobenius distance between domain matrices) would help.
[Table 1] VinDr results omit Edema (marked “–”); state whether the pathology is absent from labels or excluded for prevalence reasons so the mean is comparable across methods.
[Section 2] Related Work could briefly note multi-label TTA or medical TTA efforts beyond natural-image methods (even if limited) to better position novelty.
[Table 1, Fig. 2] Typographical inconsistencies: “CoW A” vs “CoWA” in figure captions and tables; “T arget” spacing artifacts in Table 1 headers.
[Eq. (4)] w_min = 0.01 is fixed without sensitivity; a one-line ablation would strengthen the “early-stage safeguard” claim.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: empirical TTA method whose co-occurrence reweighting is a heuristic, not a derivation that reduces to its inputs by construction.

full rationale

The paper proposes CoWA as a practical source-free TTA procedure (estimate pairwise co-occurrence matrix M from binarized model predictions via Eqs. 1-2, convert Frobenius deviation into sample weights via Eq. 3, then minimize weighted entropy on BN affine parameters via Eq. 4). It makes no first-principles uniqueness claim, no fitted-parameter-as-prediction claim, and no load-bearing self-citation of a prior uniqueness theorem or ansatz. The co-occurrence matrix is constructed from the model's own soft-then-binarized outputs, which is self-referential in the usual TTA sense (identical in kind to TENT's use of its own entropy), yet the central empirical claim—AUROC gains over independent published baselines on held-out ground-truth labels across six domain-shift pairs—is externally falsifiable and not forced by construction. Hyper-parameters are selected by ordinary search; evaluation metrics and baselines are independent of the weighting definition. Consequently the derivation chain contains no circular step of the kinds enumerated by the analyzer.

Assumptions & free parameters 4 free parameters · 4 assumptions · 1 invented entities

The method rests on standard multi-label classification and entropy TTA machinery plus a small set of free hyperparameters and the domain assumption that model-derived co-occurrence is a usable reliability signal. No new physical entities are postulated; the co-occurrence matrix and consistency weight are algorithmic constructs defined by Eqs. 1-3.

free parameters (4)

temperature τ = grid-searched per domain pair
Controls softness of the exponential weight in Eq. 3; selected from {0.01, 0.05, 0.1, 0.5} per setting.
binarization threshold for soft-to-hard labels = grid-searched
Converts continuous predictions to binary vectors before accumulating S; chosen from {0.4, 0.5, 0.6}.
learning rate for BN affine parameters = grid-searched
Selected from [1e-4, 1e-2] for each experiment.
w_min = 0.01
Floor on sample weights to prevent collapse early in adaptation; fixed at 0.01.

assumptions (4)

domain assumption Pathology labels in chest X-ray exhibit stable, informative co-occurrence structure that can be recovered from model predictions on unlabeled target data.
Stated in Introduction and Section 3; underpins the entire reliability-signal design.
domain assumption Minimizing prediction entropy on BN affine parameters is a valid TTA objective (inherited from TENT).
Section 3 Weighted Entropy Minimization and Related Work.
ad hoc to paper Frobenius distance between the outer-product of a sample's prediction vector and the global co-occurrence matrix M is a suitable scalar measure of structural consistency.
Eq. 3; justified post-hoc by the norm ablation in Table 2.
standard math Standard multi-label binary cross-entropy / AUROC evaluation and DenseNet-121 feature extractors are appropriate for the task.
Experimental Setup; common in the cited chest X-ray literature.

invented entities (1)

Co-occurrence Weighted Adaptation (CoWA) reliability weight w_i
purpose: Scalar per-sample multiplier that downweights predictions whose implied co-occurrence pattern deviates from the running target matrix M.
Defined by Eqs. 1-3; algorithmic construct rather than a physical entity. Independent evidence is limited to the correlation shown in Fig. 4 and the ablation in Table 2.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Leveraging Pathology Co-occurrence for Test-Time Adaptation in Chest X-Ray Diagnosis." pith.science (2026). https://pith.science/paper/MDGZXFRI

@misc{pith2026260703715,
  author       = {Pith},
  title        = {Pith review of: Leveraging Pathology Co-occurrence for Test-Time Adaptation in Chest X-Ray Diagnosis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MDGZXFRI}},
  note         = {Machine review of arXiv:2607.03715}
}

read the original abstract

Medical imaging models often degrade when deployed at new clinical sites due to differences in imaging equipment, protocols, and patient populations. Test-time adaptation (TTA) addresses this by updating a pretrained model using only unlabeled target data, without access to source data. However, existing TTA methods were designed for single-label classification on natural image benchmarks, minimizing entropy uniformly across all samples without considering label dependencies. This overlooks a key property of multi-label medical imaging: pathologies do not occur independently but exhibit structured co-occurrence patterns. In this work, we propose Co-occurrence Weighted Adaptation (CoWA), which leverages disease co-occurrence patterns as a reliability signal for adaptation. CoWA estimates label co-occurrence structure from model predictions and downweights samples that deviate from expected patterns, enabling adaptation to rely more on consistent predictions while reducing the impact of noisy ones. We evaluate CoWA on chest X-ray benchmarks under domain shifts and demonstrate consistent improvements over established baselines.

Figures

Figures reproduced from arXiv: 2607.03715 by the authors.

**Figure 1.** Co-occurrence patterns vary across chest X-ray domains. In this work, we propose CoWA (Co-occurrence Weighted Adaptation) to bridge this gap. CoWA estimates the pathology co-occurrence matrix from incoming test predictions, capturing the evolving co-occurrence patterns of the target domain without accessing source data. This matrix serves as a reliability signal that measures how well each sample’s predicted label … view at source ↗

**Figure 2.** Overview of CoWA. CoWA estimates a pathology co-occurrence matrix from target predictions, computes sample-wise consistency scores, and uses them to weight samples for test-time domain adaptation. co-occurrence matrix: S = Xn i=1 y˜iy˜ ⊤ i , (1) where n is the number of processed samples and Sjk counts how often pathologies j and k are jointly predicted as positive. We then compute the empirical joint probabilities … view at source ↗

**Figure 3.** Co-occurrence matrix evolution (CheXpert → MIMIC). CoWA (top) gradually aligns toward the target-trained reference (right), while the baseline (bottom) progressively loses clear dependency structure over batches. Do CoWA Weights Reflect Sample Reliability? To examine whether CoWA weights reflect sample reliability, we analyze prediction quality as a function of weight under the CheXpert → MIMIC domain shift. Target… view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Sample weight distribution and prediction reliability. [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 1 linked inside Pith

[1]

World journal of radiology6(6), 230 (2014)

Cardinale, L., Priola, A.M., Moretti, F., Volpicelli, G.: Effectiveness of chest ra- diography, lung ultrasound and thoracic computed tomography in the diagnosis of congestive heart failure. World journal of radiology6(6), 230 (2014)

2014
[2]

In: Proceedings of The 5th International Conference on Medical Imaging with Deep Learning

Cohen, J.P., Viviano, J.D., Bertin, P., Morrison, P., Torabian, P., Guarrera, M., Lungren, M.P., Chaudhari, A., Brooks, R., Hashir, M., Bertrand, H.: Torchxrayvi- sion: A library of chest x-ray datasets and models. In: Proceedings of The 5th International Conference on Medical Imaging with Deep Learning. Proceedings of Machine Learning Research, vol. 172,...

2022
[3]

In: International Conference on Machine Learning

Ganin, Y., Lempitsky, V.: Unsupervised domain adaptation by backpropagation. In: International Conference on Machine Learning. pp. 1180–1189. PMLR (2015)

2015
[4]

Ghafoorian, M., Mehrtash, A., Kapur, T., Karssemeijer, N., Marchiori, E., Pesteie, M., Guttmann, C.R., De Leeuw, F.E., Tempany, C.M., Van Ginneken, B., et al.: Transfer learning for domain adaptation in mri: Application in brain lesion seg- mentation.In:Internationalconferenceonmedicalimagecomputingandcomputer- assisted intervention. pp. 516–524. Springer (2017)

2017
[5]

In: International Conference on Learning Represen- tations (2019)

Hendrycks,D.,Dietterich,T.:Benchmarkingneuralnetworkrobustnesstocommon corruptions and perturbations. In: International Conference on Learning Represen- tations (2019)

2019
[6]

Jeong et al

Irvin, J., Rajpurkar, P., Ko, M., Yu, Y., Ciurea-Ilcus, S., Chute, C., Marklund, H., Haghgoo, B., Ball, R., Shpanskaya, K., et al.: Chexpert: A large chest radiograph 10 W. Jeong et al. dataset with uncertainty labels and expert comparison. In: Proceedings of the AAAI conference on artificial intelligence. vol. 33, pp. 590–597 (2019)

2019
[7]

PhysioNet101, 215–220 (2019)

Johnson, A., Lungren, M., Peng, Y., Lu, Z., Mark, R., Berkowitz, S., Horng, S.: Mimic-cxr-jpg-chest radiographs with structured labels. PhysioNet101, 215–220 (2019)

2019
[8]

Pattern Recognition80, 109–117 (2018)

Li, Y., Wang, N., Shi, J., Hou, X., Liu, J.: Adaptive batch normalization for prac- tical domain adaptation. Pattern Recognition80, 109–117 (2018)

2018

Show all 21 references

[9]

American journal of roentgenology144(5), 879–894 (1985)

Milne, E., Pistolesi, M., Miniati, M., Giuntini, C.: The radiologic distinction of cardiogenic and noncardiogenic edema. American journal of roentgenology144(5), 879–894 (1985)

1985
[10]

In: International Conference on Medical Image Computing and Computer-Assisted Intervention

Musa, A., Ibrahim Adamu, M., Kakudi, H.A., Hernandez, M., Lawal, Y.: Analyzing cross-population domain shift in chest x-ray image classification and mitigating the gap with deep supervised domain adaptation. In: International Conference on Medical Image Computing and Computer-...

2024
[11]

Scientific Data9(1), 429 (2022)

Nguyen, H.Q., Lam, K., Le, L.T., Pham, H.H., Tran, D.Q., Nguyen, D.B., Le, D.D., Pham, C.M., Tong, H.T., Dinh, D.H., et al.: Vindr-cxr: An open dataset of chest x-rays with radiologist’s annotations. Scientific Data9(1), 429 (2022)

2022
[12]

Niu, S., Wu, J., Zhang, Y., Chen, Y., Zheng, S., Zhao, P., Tan, M.: Efficient test- timemodeladaptationwithoutforgetting.In:Proceedingsofthe39thInternational Conference on Machine Learning. pp. 16888–16905. PMLR (2022)

2022
[13]

In: Thoracic Image Analysis (TIA 2020), Lecture Notes in Computer Science

Pooch, E.H.P., Ballester, P.L., Barros, R.C.: Can we trust deep learning based di- agnosis? the impact of domain shift in chest radiograph classification. In: Thoracic Image Analysis (TIA 2020), Lecture Notes in Computer Science. vol. 12502, pp. 74–83. Springer (2020)

2020
[14]

PloS one10(3), e0118432 (2015)

Saito, T., Rehmsmeier, M.: The precision-recall plot is more informative than the roc plot when evaluating binary classifiers on imbalanced datasets. PloS one10(3), e0118432 (2015)

2015
[15]

Wang, D., Shelhamer, E., Liu, S., Olshausen, B., Darrell, T.: Tent: Fully test- timeadaptationbyentropyminimization.In:InternationalConferenceonLearning Representations (2021)

2021
[16]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

Wang, Q., Fink, O., Van Gool, L., Dai, D.: Continual test-time domain adaptation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 7201–7211 (2022)

2022
[17]

In: Proceedings of the IEEE conference on computer vision and pattern recognition

Wang, X., Peng, Y., Lu, L., Lu, Z., Bagheri, M., Summers, R.M.: Chestx-ray8: Hospital-scale chest x-ray database and benchmarks on weakly-supervised classi- fication and localization of common thorax diseases. In: Proceedings of the IEEE conference on computer vision and patte...
[18]

arXiv preprint arXiv:1710.10501 (2017)

Yao,L.,Poblenz,E.,Dagunts,D.,Covington,B.,Bernard,D.,Lyman,K.:Learning to diagnose from scratch by exploiting dependencies among labels. arXiv preprint arXiv:1710.10501 (2017)

2017 arXiv
[19]

In: ProceedingsoftheIEEE/CVFConferenceonComputerVisionandPatternRecog- nition

Yuan, L., Xie, B., Li, S.: Robust test-time adaptation in dynamic scenarios. In: ProceedingsoftheIEEE/CVFConferenceonComputerVisionandPatternRecog- nition. pp. 15922–15932 (2023)

2023
[20]

PLoS Medicine15(11), e1002683 (2018)

Zech, J.R., Badgeley, M.A., Liu, M., Costa, A.B., Titano, J.J., Oermann, E.K.: Variable generalization performance of a deep learning model to detect pneumonia in chest radiographs: A cross-sectional study. PLoS Medicine15(11), e1002683 (2018)

2018
[21]

In: International Conference on Learning Representations (2021)

Zhou, K., Yang, Y., Qiao, Y., Xiang, T.: Domain generalization with mixstyle. In: International Conference on Learning Representations (2021)

2021

Pith tools

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

[1] [1]

World journal of radiology6(6), 230 (2014)

Cardinale, L., Priola, A.M., Moretti, F., Volpicelli, G.: Effectiveness of chest ra- diography, lung ultrasound and thoracic computed tomography in the diagnosis of congestive heart failure. World journal of radiology6(6), 230 (2014)

2014

[2] [2]

In: Proceedings of The 5th International Conference on Medical Imaging with Deep Learning

Cohen, J.P., Viviano, J.D., Bertin, P., Morrison, P., Torabian, P., Guarrera, M., Lungren, M.P., Chaudhari, A., Brooks, R., Hashir, M., Bertrand, H.: Torchxrayvi- sion: A library of chest x-ray datasets and models. In: Proceedings of The 5th International Conference on Medical Imaging with Deep Learning. Proceedings of Machine Learning Research, vol. 172,...

2022

[3] [3]

In: International Conference on Machine Learning

Ganin, Y., Lempitsky, V.: Unsupervised domain adaptation by backpropagation. In: International Conference on Machine Learning. pp. 1180–1189. PMLR (2015)

2015

[4] [4]

Ghafoorian, M., Mehrtash, A., Kapur, T., Karssemeijer, N., Marchiori, E., Pesteie, M., Guttmann, C.R., De Leeuw, F.E., Tempany, C.M., Van Ginneken, B., et al.: Transfer learning for domain adaptation in mri: Application in brain lesion seg- mentation.In:Internationalconferenceonmedicalimagecomputingandcomputer- assisted intervention. pp. 516–524. Springer (2017)

2017

[5] [5]

In: International Conference on Learning Represen- tations (2019)

Hendrycks,D.,Dietterich,T.:Benchmarkingneuralnetworkrobustnesstocommon corruptions and perturbations. In: International Conference on Learning Represen- tations (2019)

2019

[6] [6]

Jeong et al

Irvin, J., Rajpurkar, P., Ko, M., Yu, Y., Ciurea-Ilcus, S., Chute, C., Marklund, H., Haghgoo, B., Ball, R., Shpanskaya, K., et al.: Chexpert: A large chest radiograph 10 W. Jeong et al. dataset with uncertainty labels and expert comparison. In: Proceedings of the AAAI conference on artificial intelligence. vol. 33, pp. 590–597 (2019)

2019

[7] [7]

PhysioNet101, 215–220 (2019)

Johnson, A., Lungren, M., Peng, Y., Lu, Z., Mark, R., Berkowitz, S., Horng, S.: Mimic-cxr-jpg-chest radiographs with structured labels. PhysioNet101, 215–220 (2019)

2019

[8] [8]

Pattern Recognition80, 109–117 (2018)

Li, Y., Wang, N., Shi, J., Hou, X., Liu, J.: Adaptive batch normalization for prac- tical domain adaptation. Pattern Recognition80, 109–117 (2018)

2018