Pith. sign in

REVIEW 4 major objections 5 minor 26 references

Self-Supervised Multiview Xray Matching

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Automatically generated many-to-many correspondences between synthetic X-ray views, derived from unannotated CT volumes, can serve as self-supervised supervision to train a transformer that matches real X-ray views and improves multi-view…

desk verdict Novel DRR-derived many-to-many correspondence supervision with impressive real-X-ray matching, but the fracture-detection gain is not yet attributable to correspondence learning because the pretraining control is missing. read the letter →

arxiv 2507.00287 v1 pith:WU67QSIX submitted 2025-06-30 cs.CV cs.AI

Mohamad Dabboussi , Malo Huard , Yann Gousseau , Pietro Gori This is my paper · ORCID
classification cs.CVcs.AI
keywords multi-viewX-rayself-supervisedlearningcorrespondenceestimationdigitallyreconstructedradiographsfracturedetectiontransformermany-to-manymatchingpretraining
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 establish that a model can learn meaningful correspondences between different X-ray views without any manual annotation, by generating synthetic training pairs from CT volumes. The authors show that pretraining a transformer on these automatically generated correspondences transfers to real X-ray data, improving multi-view fracture detection accuracy. If true, this would remove a major bottleneck in multi-view X-ray analysis, where annotated correspondence data is scarce and existing matching methods assume one-to-one rather than many-to-many relationships.

What carries the argument

The central object is the voxel-wise outer product correspondence matrix: for each non-air voxel, the flattened projection matrices $P^v_1$ and $P^v_2$ are multiplied as $C^v_{1,2} = p^v_1 (p^v_2)^\top$, then the final matrix is the element-wise maximum over all voxels, computed at a coarsened patch resolution (downsampling factor $k=16$). This matrix serves both as training supervision for a transformer that predicts correspondences via a normalized dot-product between patch embeddings, and as an additive attention bias $A' = A + \alpha C$ in a downstream multi-view classification transformer.

What would settle it

Collect a densely annotated real X-ray correspondence test set (e.g., 50 studies with full landmark correspondences) and evaluate the pretrained transformer; if average precision on dense matching is far below the 83.8 AP reported on the three-positive-per-image test set, the transferability claim fails.

Watch

Extended reading notes

Core claim

The authors claim that synthetic X-ray views with dense many-to-many correspondence matrices can be generated automatically from unannotated CT volumes by projecting non-air voxels onto two DRR views and taking the element-wise maximum of the outer products of the flattened projection matrices. A transformer trained on these correspondences reaches 83.8 average precision on a real X-ray correspondence test set, and using the predicted correspondence matrix as an attention bias improves MURA fracture classification accuracy from 75.8 percent (early fusion baseline) to 80.6 percent, with a similar gain on a private dataset from 71.1 to 76.2 percent.

Load-bearing premise

The automatically generated correspondence matrices—computed by projecting every non-air CT voxel onto two DRR views and taking the element-wise maximum of outer products at coarse patch resolution—are an accurate and transferable notion of correspondence for real X-ray views.

Editorial extensions

If this is right

  • Correspondence pretraining on synthetic DRRs transfers to real X-ray data without any manual correspondence annotation, improving MURA accuracy from 75.8 percent to 80.1 percent even without attention guidance.
  • Incorporating the predicted correspondence matrix as an attention bias further raises accuracy to 80.6 percent on MURA and 76.2 percent on the private dataset, suggesting that explicit cross-view spatial relationships help multi-view classification.
  • The many-to-many correspondence formulation outperforms one-to-one matching baselines such as LoFTR and SuperGlue on both synthetic and real X-ray correspondence tests, consistent with the cumulative nature of X-ray image formation.
  • A compact 24M-parameter transformer pretrained and fine-tuned on this task surpasses a 1.1B-parameter DINOv2 model with a fine-tuned head on the real correspondence test (AP 83.8 vs 42.2).
  • The same self-supervised pipeline can produce training data across diverse anatomical regions from any unannotated CT volume, enabling scalable multi-view X-ray pretraining beyond chest radiographs.

Reading between the lines

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

  • A natural extension would apply the same correspondence supervision to more than two views, since the transformer processes concatenated patch embeddings from several views and the outer-product construction generalizes directly to multi-view outer products.
  • Because the ground-truth correspondences are purely geometric (ray paths through all non-air voxels), they may not align with the visible landmarks radiologists use; the reported 83.8 AP on real data should be read cautiously until denser annotations test that alignment.
  • The coarse $k=16$ patch resolution suggests that finer-grained tasks such as detecting small fractures or tracking surgical implants might require a smaller downsampling factor or a hierarchical refinement stage.
  • The method could be tested as a pretraining step for other multi-view tasks such as bone-age estimation, 3D reconstruction from X-rays, or registration, where cross-view correspondences are equally central.
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

4 major / 5 minor

Summary. The paper proposes a fully automatic pipeline for training multi-view X-ray correspondence models without manual annotations. Given a CT volume, the method generates pairs of digitally reconstructed radiographs (DRRs) and a many-to-many patch-level correspondence matrix by projecting non-air voxels onto two views and taking the element-wise maximum of the resulting outer products. A ResNet-50 backbone plus a transformer then predicts this correspondence matrix from the two views, with comparison against SuperGlue, LoFTR, and CNN-only baselines. The learned correspondence information is used in two downstream ways: as a self-supervised pretraining objective for multi-view fracture classification, and as an additive attention bias in a transformer classifier. Experiments are reported on a simulated test set, a small real X-ray correspondence dataset, the MURA public dataset, and a private multi-view dataset. The main quantitative claims are that correspondence pretraining improves MURA accuracy from 75.8% to 80.1% (early fusion) and to 80.6% with attention guidance, with similar gains on the private dataset, and that the proposed correspondence model reaches 83.8 AP on the real X-ray correspondence test set.

Significance. If the claims hold, the paper would make a practical contribution: it turns unannotated CT volumes into a large, automatically labeled source of multi-view X-ray correspondences, and it demonstrates a use of those correspondences for fracture classification. The geometric construction is simple, scalable, and does not rely on manual keypoints. The experimental results are mostly encouraging: the transformer-based correspondence predictor outperforms strong matching baselines on simulated data and transfers to real X-rays, and the classification gains on both MURA and the private dataset are consistent and reported with error bars. The main weakness is that the paper's central attribution claim—that the classification gains are specifically caused by learning correspondences rather than by generic exposure to synthetic radiographs—is not isolated in the experiments. The missing control is a load-bearing gap that can be addressed with additional experiments, so the paper is not ready for acceptance as is.

major comments (4)
  1. [Section 4.2, Table 3] The central claim of the abstract and Section 3.3—that learning correspondences among synthetic views can be leveraged as a pretraining strategy—is not isolated by the experimental design. The comparison between row “– – Early” and row “✓ – Early” changes two things at once: the network is exposed to 207,600 synthetic DRR pairs, and the training objective is the Section 3.1 correspondence matrix. Since all models are ImageNet-initialized, the observed gain (75.8% to 80.1% on MURA; 71.1% to 75.0% on the private set) could plausibly come from generic domain adaptation to radiograph statistics rather than from the correspondence-specific supervision. Please add a control trained on the same synthetic pairs with a different pretext objective (for example, view-angle regression, masked-view reconstruction, or the same transformer trained with a shuffled or constant correspondence target). Without such a control, the causal attribution in the abstract and in Section 3.3 is not established.
  2. [Section 3.4, Eq. (2)] The attention-guidance mechanism is underspecified, and this matters because the “✓ ✓” rows of Table 3 are the only evidence for the attention-guidance contribution. It is not stated whether C in Eq. (2) is the ground-truth correspondence matrix, the predicted matrix from the pretrained correspondence network, or a thresholded/binarized version; how C is reshaped to match the concatenated QK^T attention matrix over patches from both views; and how the learnable scalar alpha is initialized. If C were available only from the synthetic generation process and not at real-image inference time, the reported classification gains would not transfer as stated. Please specify the full forward path for attention guidance during both training and inference.
  3. [Section 4.1, Table 2] The real X-ray correspondence evaluation is very sparse. Each of the 347 samples has only three annotated positive correspondences, and the 30% test split amounts to roughly 104 samples. AP and recall estimates on such a small set of positives are high-variance, and no error bars, confidence intervals, or repeated-seed variability are reported. Additionally, the selection of the 100 negative correspondences per sample is not described. Please enlarge the set of positive annotations per image (for example, by deriving dense correspondences from the same CT-based projections), report per-region or per-sample statistics, and provide bootstrapped confidence intervals for the Table 2 numbers.
  4. [Section 3.1] The definition C_{1,2} = max_v p_v1 p_v2^T means that any two pixels lying on the same ray through at least one non-air voxel are marked as corresponding, regardless of whether that voxel is salient in the radiograph. Combined with the heavy downsampling factor k=16 and the unspecified non-air threshold, the generated supervision may be over-complete relative to clinically meaningful correspondences. The transferability argument depends on this signal being meaningful, so please report at least one sensitivity experiment on the downstream classification task where k and the threshold are varied, and show that the pretraining benefit is not an artifact of a particular coarse geometric labeling.
minor comments (5)
  1. [Table 1] The formatting of Table 1 is confusing: the “Attention Model” column contains dashes, module names, and “Standard Transformer” with inconsistent “Message Pass” entries, and the CNN-only baseline is not explicitly labeled as such. Please restructure the table so each row is a named method.
  2. [Section 4.1] Please specify how the 100 negative correspondences per real sample are chosen (random patches, same-image hard negatives, or other), since this directly affects the reported AP values in Table 2.
  3. [Section 3.2] The manuscript refers to “patch-level correspondences” but does not state the patch size or the spatial resolution of the predicted correspondence matrix. With a ResNet-50 backbone on 256x256 inputs, the feature grid is 8x8 under stride 32, whereas the k=16 downsampled CT volume suggests a different grid size; please clarify the relationship.
  4. [Section 3.4, Eq. (2)] To avoid confusion with the ground-truth matrix C in Section 3.1, please use a distinct symbol (for example, C_hat) for the correspondence matrix used as an attention bias, and state explicitly when it is predicted versus taken from the dataset.
  5. [General] There are several typographical issues, including “T able” in Section 4.2 and inconsistent spacing around multiplication signs; a careful proofread is needed.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the correspondence labels are geometrically derived from CT projections, and the downstream claims are evaluated on external held-out data.

full rationale

The paper's derivation chain is self-contained but not circular. The correspondence ground truth is defined geometrically (Sec. 3.1) as max_v outer(p^v_1, p^v_2) from CT projection matrices; the model is trained to predict these matrices from synthetic DRR views and tested on held-out simulated and real X-ray correspondences (Sec. 4.1, Table 2). The pretraining claim is supported by Table 3, where accuracy is measured on external MURA and private classification labels, not on the correspondence targets. No equation reduces a predicted quantity to a fitted parameter, and no load-bearing argument relies on self-citations: the references (Joseph, ViT, LoFTR, DINOv2, MURA) are external prior work. The main weakness is not circularity but experimental control: the '✓ – Early' row differs from '– – Early' by both the synthetic DRR exposure and the correspondence objective, so the improvement cannot be uniquely attributed to correspondence learning without a non-correspondence pretext control. This is a validity and ablation concern, not a definitional or self-citation circularity, so the circularity score is 0.

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

The central claim depends on the reliability of the automatically generated DRR correspondence labels, the learnability of those labels from 2D views, and the transfer of the resulting representations to real X-rays. The free parameters are hand-chosen values that control the granularity of matching and the evaluation protocol. No new physical entities are introduced.

free parameters (3)
  • Downsampling factor k = 16
    Chosen in Section 3.1 to compute patch-level correspondences from the CT volume. It sets the coarseness of the correspondence matrix and is a hand-selected design parameter.
  • Non-air voxel threshold = not specified
    Section 3.1 applies a threshold to select non-air voxels before projection. The threshold value is not reported, making it an unspecified free parameter that affects the correspondence labels.
  • Negative correspondences per real sample = 100
    Section 4.1 states each real X-ray sample includes three positive correspondences and 100 negative correspondences. This ratio is chosen by hand and shapes the precision/recall metrics and fine-tuning signal.
assumptions (3)
  • domain assumption The Joseph ray-tracing method generates DRRs that approximate clinical X-ray projections closely enough for correspondence learning.
    The correspondence ground truth is built by projecting voxels onto two DRR views using the Joseph method (Section 3.1). If the simulated projections diverge from real X-ray physics or anatomy, the learned correspondences may not transfer.
  • domain assumption The element-wise maximum over per-voxel outer products defines the correct many-to-many correspondence matrix.
    Section 3.1 defines C = max_v p_v1 p_v2^T. This assumes that aggregating voxel-level projection pairs by maximum yields a faithful representation of pixel correspondences between views.
  • domain assumption The correspondence matrix can be predicted from two 2D images without explicit depth or geometry information.
    Section 3.2 uses a transformer to map image pairs directly to the matrix C. This assumes the mapping is learnable from appearance and that synthetic pretraining supports real X-ray inference after fine-tuning.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Self-Supervised Multiview Xray Matching." pith.science (2026). https://pith.science/paper/WU67QSIX

@misc{pith2026250700287,
  author       = {Pith},
  title        = {Pith review of: Self-Supervised Multiview Xray Matching},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WU67QSIX}},
  note         = {Machine review of arXiv:2507.00287}
}
read the original abstract

Accurate interpretation of multi-view radiographs is crucial for diagnosing fractures, muscular injuries, and other anomalies. While significant advances have been made in AI-based analysis of single images, current methods often struggle to establish robust correspondences between different X-ray views, an essential capability for precise clinical evaluations. In this work, we present a novel self-supervised pipeline that eliminates the need for manual annotation by automatically generating a many-to-many correspondence matrix between synthetic X-ray views. This is achieved using digitally reconstructed radiographs (DRR), which are automatically derived from unannotated CT volumes. Our approach incorporates a transformer-based training phase to accurately predict correspondences across two or more X-ray views. Furthermore, we demonstrate that learning correspondences among synthetic X-ray views can be leveraged as a pretraining strategy to enhance automatic multi-view fracture detection on real data. Extensive evaluations on both synthetic and real X-ray datasets show that incorporating correspondences improves performance in multi-view fracture classification.

Figures

Figures reproduced from arXiv: 2507.00287 by the authors.

Figure 1
Figure 1. End-to-end self-supervised pipeline for predicting X-ray correspondences and multi-view fracture classification. (Left) From a CT volume we generate multiple 2D DRRs (digitally reconstructed radiographs) with a correspondence matrix link￾ing matching points. (Middle) A backbone network extracts features from each view, processed by a Transformer to predict a Correspondence Matrix via attention mecha￾nisms. (Right) T… view at source ↗
Figure 2
Figure 2. Many-to-many matching between two X-ray views of the same subject. The red patch in the left view corresponds to the red bar in the right view, illustrating multiple matching pixels across views. views [2], (2) 2D/3D image registration [16, 22, 23], and (3) generating labeled datasets for segmentation using manually labeled 3D CT volumes [9, 10, 24]. Unlike prior work, we leverage DRR to automatically generate corre… view at source ↗
Figure 3
Figure 3. Visualization of generated hand correspondences annotations. 3.2 Correspondence Prediction We formulate correspondence prediction as a similarity assessment between fea￾tures extracted from patches across different views. A pretrained backbone net￾work, denoted as f, extracts feature maps from each image, which are then projected into a lower-dimensional embedding space. The spatial grid of these feature maps is fla… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Generating a correspondence matrix for a 3 × 3 × 3 volume with a highlighted 1 × 1 × 2 cube (yellow) using two orthogonal views. Each row shows the projection of a yellow voxel v onto both views. The resulting projection matrices, P v 1 and P v 2 , are flattened, and t…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

26 extracted references · 17 canonical work pages

  1. [1]

    In: European Conference on Computer Vision

    Chen, H., Luo, Z., Zhou, L., Tian, Y., Zhen, M., Fang, T., Mckinnon, D., Tsin, Y., Quan, L.: Aspanformer: Detector-free image matching with adaptive span trans- former. In: European Conference on Computer Vision. pp. 20–36. Springer (2022)

    work page 2022

  2. [2]

    In: 2022 44th annual international conference of the IEEE engineering in medicine & Biology society (EMBC)

    Corona-Figueroa,A.,Frawley,J.,Bond-Taylor,S.,Bethapudi,S.,Shum,H.P.,Will- cocks, C.G.: Mednerf: Medical neural radiance fields for reconstructing 3d-aware ct-projections from a single x-ray. In: 2022 44th annual international conference of the IEEE engineering in medicine & Biology society (EMBC). pp. 3843–3848. IEEE (2022)

    work page 2022

  3. [3]

    Radiotherapy and Oncology 153, 213–219 (2020)

    Dhont, J., Verellen, D., Mollaert, I., Vanreusel, V., Vandemeulebroucke, J.: Realdrr–rendering of realistic digitally reconstructed radiographs using locally trained image-to-image translation. Radiotherapy and Oncology 153, 213–219 (2020)

    work page 2020

  4. [4]

    arXiv preprint arXiv:2010.11929 (2020)

    Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., et al.: An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929 (2020)

    arXiv 2020

  5. [5]

    In: Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition

    Edstedt, J., Sun, Q., Bökman, G., Wadenbäck, M., Felsberg, M.: Roma: Robust dense feature matching. In: Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition. pp. 19790–19800 (2024)

    2024

  6. [6]

    He,K.,Zhang,X.,Ren,S.,Sun,J.:Deepresiduallearningforimagerecognition.In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 770–778 (2016)

    2016

  7. [7]

    Computer vision and image under- standing 110(3), 346–359 (2008)

    Herbert, B.: Surf: Speeded up robust features. Computer vision and image under- standing 110(3), 346–359 (2008)

    work page 2008

  8. [8]

    IEEE transactions on medical imaging1(3), 192–196 (1982)

    Joseph, P.M.: An improved algorithm for reprojecting rays through pixel images. IEEE transactions on medical imaging1(3), 192–196 (1982)

    work page 1982

Show all 26 references
  1. [9]

    In: Machine Learning for Medical Image Reconstruction: Third International Workshop, MLMIR 2020, Held in Conjunction with MICCAI 2020, Lima, Peru, October 8, 2020, Proceedings

    Kasten, Y., Doktofsky, D., Kovler, I.: End-to-end convolutional neural network for 3d reconstruction of knee bones from bi-planar x-ray images. In: Machine Learning for Medical Image Reconstruction: Third International Workshop, MLMIR 2020, Held in Conjunction with MICCAI 2020...

    2020

  2. [10]

    pp. 123–133. Springer (2020)

    2020

  3. [11]

    arXiv preprint arXiv:2403.08059 (2024)

    Killeen, B.D., Wang, L.J., Zhang, H., Armand, M., Taylor, R.H., Osgood, G., Unberath, M.: Fluorosam: A language-aligned foundation model for x-ray image segmentation. arXiv preprint arXiv:2403.08059 (2024)

    2024 arXiv

  4. [12]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)

    Lindenberger, P., Sarlin, P.E., Pollefeys, M.: Lightglue: Local feature matching at light speed. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV). pp. 17627–17638 (October 2023)

    2023

  5. [13]

    Interna- tional journal of computer vision60, 91–110 (2004)

    Lowe, D.G.: Distinctive image features from scale-invariant keypoints. Interna- tional journal of computer vision60, 91–110 (2004)

    2004

  6. [14]

    arXiv preprint arXiv:2304.07193 (2023)

    Oquab, M., Darcet, T., Moutakanni, T., Vo, H., Szafraniec, M., Khalidov, V., Fernandez, P., Haziza, D., Massa, F., El-Nouby, A., et al.: Dinov2: Learning robust visual features without supervision. arXiv preprint arXiv:2304.07193 (2023)

    2023 arXiv

  7. [15]

    arXiv preprint arXiv:1712.06957 (2017)

    Rajpurkar, P., Irvin, J., Bagul, A., Ding, D., Duan, T., Mehta, H., Yang, B., Zhu, K., Laird, D., Ball, R.L., et al.: Mura: Large dataset for abnormality detection in musculoskeletal radiographs. arXiv preprint arXiv:1712.06957 (2017)

    2017 arXiv

  8. [16]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (June 2020) 10 M

    Sarlin, P.E., DeTone, D., Malisiewicz, T., Rabinovich, A.: Superglue: Learning feature matching with graph neural networks. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (June 2020) 10 M. Dabboussi et al

    2020

  9. [17]

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

    Shrestha, P., Xie, C., Shishido, H., Yoshii, Y., Kitahara, I.: X-ray to ct rigid reg- istration using scene coordinate regression. In: International Conference on Medi- cal Image Computing and Computer-Assisted Intervention. pp. 781–790. Springer (2023)

    2023

  10. [18]

    Physics in Medicine & Biology30(8), 817 (1985)

    Siddon, R.L.: Prism representation: a 3d ray-tracing algorithm for radiotherapy applications. Physics in Medicine & Biology30(8), 817 (1985)

    1985

  11. [19]

    Neurocomputing568, 127063 (2024)

    Su, J., Ahmed, M., Lu, Y., Pan, S., Bo, W., Liu, Y.: Roformer: Enhanced trans- former with rotary position embedding. Neurocomputing568, 127063 (2024)

    2024

  12. [20]

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

    Sun, J., Shen, Z., Wang, Y., Bao, H., Zhou, X.: Loftr: Detector-free local feature matching with transformers. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). pp. 8922–8931 (June 2021)

    2021

  13. [21]

    In: Medical Image Computing and Computer Assisted Intervention–MICCAI 2018: 21st International Conference, Granada, Spain, September 16-20, 2018, Proceed- ings, Part IV 11

    Unberath, M., Zaech, J.N., Lee, S.C., Bier, B., Fotouhi, J., Armand, M., Navab, N.: Deepdrr–a catalyst for machine learning in fluoroscopy-guided procedures. In: Medical Image Computing and Computer Assisted Intervention–MICCAI 2018: 21st International Conference, Granada, Spa...

    2018

  14. [22]

    Advances in neural information pro- cessing systems 30 (2017)

    Vaswani,A.,Shazeer,N.,Parmar,N.,Uszkoreit,J.,Jones,L.,Gomez,A.N.,Kaiser, Ł., Polosukhin, I.: Attention is all you need. Advances in neural information pro- cessing systems 30 (2017)

    2017

  15. [23]

    Wu, H., Zhang, J., Fang, Y., Liu, Z., Wang, N., Cui, Z., Shen, D.: Multi-view vertebralocalizationandidentificationfromctimages.In:InternationalConference on Medical Image Computing and Computer-Assisted Intervention. pp. 136–145. Springer (2023)

    2023

  16. [24]

    Zhang, B., Faghihroohi, S., Azampour, M.F., Liu, S., Ghotbi, R., Schunkert, H., Navab, N.: A patient-specific self-supervised model for automatic x-ray/ct registra- tion.In:MedicalImageComputingandComputerAssistedIntervention–MICCAI

  17. [26]

    In: Interna- tional conference on medical image computing and computer-assisted intervention

    Zhang, Y., Miao, S., Mansi, T., Liao, R.: Task driven generative modeling for unsu- pervised domain adaptation: Application to x-ray image segmentation. In: Interna- tional conference on medical image computing and computer-assisted intervention. pp. 599–607. Springer (2018)

    2018

  18. [2023]

    pp. 515–524. Springer Nature Switzerland, Cham (2023)

    2023

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.