Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

Learning Anatomy-Grounded CT Vision-Language Representations with Organ-Hierarchical Report Knowledge

T0 review · 3 major / 4 minor · reviewed 2026-07-14 · grok-4.5

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

Pith's one-line read Organ-level structure from radiology reports, reused as supervision, makes CT vision-language models more accurate and anatomy-sensitive.

desk verdict Real CT-VLP gains from a coherent two-stage organ-hierarchy package; the extraction-quality gap is real but does not erase the empirical result. read the letter →

arxiv 2607.10953 v1 pith:PITVCMKO submitted 2026-07-12 cs.CV

Guoliang You , Hongming Li , Yuanwang Zhang , Yong Fan This is my paper
classification cs.CV
keywords Medicalvision-languagepretrainingOrgan-hierarchicalknowledgeReportextractionStructuredcontrastivelearningCT-reportalignmentZero-shotabnormalitydiagnosisAnatomy-groundedrepresentation
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

Radiology reports describe findings organ by organ—abnormality status, disease concepts, location, and severity—but most CT vision-language pretraining either matches whole scans to whole reports or matches local patches to text fragments. This paper argues that the organ hierarchy itself is usable supervision. OKA-CT extracts that hierarchy automatically from free-text reports, then reuses it in two stages: first to train the CT encoder with organ-conditioned visual targets, and second to guide report–CT contrastive learning so that cases sharing the same organ-level findings become weak positives rather than hard negatives. A light query branch still builds a whole-scan embedding. On two public chest CT datasets the method raises zero-shot abnormality diagnosis and retrieval scores over prior CT vision-language baselines and, under patch occlusion, responds more strongly to disease-associated anatomy. The practical claim is that anatomy-grounded report structure is a scalable signal for better CT representations without dense manual labels.

What carries the argument

Organ-hierarchical knowledge (slots a, c, ℓ, h per organ) extracted by report parsing plus LLM structuring, reused as Stage-1 organ-conditioned heads on mask-pooled features and as Stage-2 soft target matrices that turn shared organ-level findings into weak positives in contrastive loss.

What would settle it

If independent expert review of the extracted organ slots shows high error rates on the same organs, and replacing those slots with random or shuffled slots collapses the reported AUROC and retrieval gains, the central claim fails.

Watch

Extended reading notes

Core claim

OKA-CT shows that organ-hierarchical knowledge extracted from free-text radiology reports—abnormality, concept, location, and attribute slots—can be reused as weak supervision for both organ-conditioned visual pretraining and structured report–CT contrastive learning, producing CT representations that improve zero-shot diagnosis, retrieval, and sensitivity to disease-linked anatomical regions.

Load-bearing premise

The automatically extracted organ slots from free-text reports are accurate enough as weak labels that supervising visual features and defining soft semantic neighborhoods improves rather than corrupts the learned evidence space.

Editorial extensions

If this is right

  • CT vision-language models can improve zero-shot abnormality diagnosis by treating report organ hierarchy as reusable weak supervision rather than only whole-report text.
  • Non-paired CT–report pairs that share organ-level findings should be soft positives in contrastive learning, not uniform hard negatives.
  • Global CT embeddings become more sensitive to disease-associated anatomical regions when organ-conditioned supervision is injected before report alignment.
  • The same organ-slot pipeline can, in principle, extend beyond the four chest organs studied here once masks and report structuring cover a larger anatomy set.

Reading between the lines

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

  • If extraction quality is the main bottleneck, better report parsers or lighter human verification of slots could raise performance more cheaply than scaling the visual backbone alone.
  • The soft-target idea may transfer to other paired medical modalities where free-text findings are also organ- or system-organized (e.g., MRI reports).
  • Domain shift between CT-RATE and RAD-ChestCT remains large; calibrating organ-slot distributions across institutions may be as important as the architecture itself.
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 / 4 minor

Summary. OKA-CT is a two-stage CT-report vision-language pretraining framework that converts free-text radiology reports into organ-conditioned hierarchical knowledge (abnormality status, disease/finding concepts, location, and severity attributes) via RadGraph-XL parsing and Qwen3-8B structuring. Stage 1 trains a 3D ResNet-18 with global disease supervision plus organ-mask-pooled visual tokens supervised by the four hierarchy slots (Eqs. 1–5); Stage 2 freezes the visual encoder and performs global report-CT contrastive learning plus organ-level alignment with hierarchy-derived soft targets that treat non-paired cases sharing organ-level findings as weak positives (Eqs. 6–13, Fig. 4). A query-based global branch aggregates volumetric evidence. On CT-RATE the method reports zero-shot abnormality diagnosis AUROC 84.9 (vs. 77.8 for fVLM) and strong retrieval gains; on RAD-ChestCT AUROC is 72.2. Ablations (Tables III–IV), organ-group deltas, and patch-occlusion maps support the contribution of the hierarchy.

Significance. If the gains are truly driven by reusable organ-hierarchical report knowledge rather than generic multi-task or soft-label effects, the work offers a practical and scalable way to inject anatomy-grounded structure into CT VLP without dense expert organ annotations. The two-stage reuse of the same slots for visual supervision and soft-target contrastive neighborhoods is a clear design contribution relative to prior global (CT-CLIP, Merlin) and local/anatomy-aware (fVLM, CT-GLIP) methods. Empirical strengths include multi-metric diagnosis and retrieval results, cross-dataset transfer, stepwise ablations, Stage-1-only analysis, and qualitative occlusion evidence of anatomy-sensitive localization. These make the paper a solid candidate for the medical VLP literature once the weakest assumption—slot quality—is better supported.

major comments (3)
  1. [§III.B, Fig. 3, Eqs. (3)–(5), (9)–(12)] §III.B and Fig. 3: The central claim attributes gains to organ-hierarchical slots z^o_i = {a,c,ℓ,h} used in Stage-1 heads (Eqs. 3–5) and Stage-2 soft targets (Eqs. 9–12). The paper never reports extraction precision/recall, error modes (negation, laterality, multi-organ findings), or agreement against expert organ-level annotations. Without this, Tables I–III cannot isolate the hierarchy mechanism from noisy multi-task labels or soft-label smoothing. A modest validation set or controlled noise/shuffle ablation of slots is needed to support the mechanism claim.
  2. [§III.B, Eq. (4), Table IV] §III.B and Stage-1 objective (Eq. 4): Case-level disease labels are used as fallbacks when organ concepts are missing. This can allow organ heads to re-learn global labels rather than anatomy-grounded evidence, weakening the claim that Stage 1 injects organ-conditioned visual geometry. Please quantify fallback frequency and report Stage-1 performance with vs. without fallbacks, or restrict organ supervision to report-derived slots only.
  3. [Tables I–II, §IV.B] Tables I–II: Several baseline numbers are collected from prior comparisons and some metrics are missing (e.g., COLIPRI-C). No error bars, multiple seeds, or significance tests are reported for the 7.1-point AUROC gain over fVLM or the large report-image retrieval lift. Given free parameters λ1, λ2, α, β, γ, σ and temperatures, at least seed variance or a sensitivity sweep on the soft-target weights would strengthen confidence that the gains are stable.
minor comments (4)
  1. [Fig. 5, §IV.D] Fig. 5: Occlusion heatmaps are informative but only qualitative; a quantitative summary (e.g., fraction of score drop inside organ masks vs. outside) would better support the anatomy-sensitivity claim.
  2. [Abstract, §I, §V] Notation: organ set O is restricted to lung/heart/esophagus/aorta; the discussion notes this limitation, but the abstract and introduction should more clearly scope claims to chest CT with these four organs.
  3. [Eq. (11)] Eq. (11): The weighted contrastive loss uses raw target mass rather than row-normalized targets; a short justification or comparison to normalized soft targets would help readers.
  4. [Table II, Abstract] Typos/clarity: “V ocabFine” spacing in Table II; “Stage~1” tilde usage is inconsistent with the rest of the text; some figure captions are dense and could be tightened.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: organ hierarchy is weak training supervision; zero-shot AUROC/retrieval are held-out empirical outcomes, not forced by construction.

full rationale

OKA-CT’s chain is methodological, not a first-principles derivation that collapses into its inputs. Report slots z^o_i = {a,c,ℓ,h} are extracted once (RadGraph-XL + Qwen3-8B) and reused as weak multi-task targets in Stage 1 (Eqs. 3–5) and as batch soft-target matrices in Stage 2 (Eqs. 9–12, Fig. 4). Those slots define training neighborhoods; they do not algebraically equal the reported zero-shot AUROCs, MAP, or Recall. Global Stage-1 disease labels y_i and held-out diagnosis labels share a clinical label space—standard supervised pretraining, not a self-definitional loop—and Table IV explicitly separates supervised Stage-1 AUROC from post-alignment zero-shot metrics in Table I. Soft-target weights (α,β,γ,σ) are fixed hyperparameters, not fits that are then re-reported as predictions. Cross-dataset RAD-ChestCT results and patch-occlusion analyses further sit outside any in-batch construction. No load-bearing uniqueness theorem or ansatz is imported from overlapping-author citations. Unvalidated slot quality is a correctness/assumption risk, not circularity: the paper does not claim the slots are proven ground truth, nor does any equation make the evaluation metric identical to the training target by definition. Score 0; steps empty.

Assumptions & free parameters 4 free parameters · 5 assumptions · 3 invented entities

The central claim rests on standard medical VLP machinery plus several paper-specific modeling choices: that free-text reports can be reduced to four organ slots, that automatic masks define the right visual support, that soft semantic neighborhoods should be weighted by hand-chosen coefficients, and that freezing the Stage-1 encoder preserves useful geometry for Stage-2 alignment. No new physical entity is postulated; the invented pieces are algorithmic constructs.

free parameters (4)
  • Stage-1 organ loss weight λ1
    Hand-set to 0.05; controls how strongly noisy organ slots reshape the visual encoder relative to global disease supervision.
  • Stage-2 organ alignment weight λ2
    Hand-set to 0.1; balances global report-CT contrast against organ-level structured contrast.
  • Soft-target field weights α,β,γ,σ
    Set to 0.8, 0.8, 0.2, 0.15 for abnormality, concept, location, and attribute overlap; these directly define which non-paired cases become weak positives.
  • Contrast temperatures τg and τo
    Temperature parameters in global and organ similarity logits; values not fully specified but required for the contrastive objectives.
assumptions (5)
  • domain assumption Radiology reports are naturally organized around organs with abnormality, concept, location, and severity-related attributes as reusable supervisory slots.
    Stated in Introduction and Method as the key property motivating OKA-CT; without it the hierarchy is an arbitrary schema.
  • domain assumption Automatic organ masks (TotalSegmentator-derived) sufficiently localize organ-conditioned visual evidence for supervision and alignment.
    Used in Eq. (2) and implementation details; mask errors would misassign organ supervision.
  • ad hoc to paper Non-paired cases sharing organ-level hierarchy fields should be treated as weak positives rather than uniform hard negatives in contrastive learning.
    Encoded in Eqs. (9)-(12) and Fig. 4; this is a design choice, not a theorem.
  • ad hoc to paper Freezing the Stage-1 visual encoder preserves anatomy-aware geometry while Stage-2 learns report alignment on top.
    Stated in §III.D; alternative joint fine-tuning is not shown to be inferior beyond the chosen protocol.
  • standard math Standard bidirectional InfoNCE-style contrastive learning and BCE/CE multi-label heads are valid objectives for CT-report representation learning.
    Used throughout Stage 1 and Stage 2 losses; conventional in VLP.
invented entities (3)
  • Organ-hierarchical knowledge slots z^o_i = {a,c,ℓ,h}
    purpose: Convert free-text reports into organ-conditioned weak labels for visual supervision and soft-target construction.
    Paper-defined schema produced by parser+LLM; not independently validated organ ground truth.
  • Hierarchy-derived organ soft target matrix T^o = I + αS_a + βS_c + γS_ℓ + σS_h
    purpose: Define anatomy-grounded semantic neighborhoods for organ-level contrastive learning.
    Constructed inside the paper from extracted slots and hand weights; no external measurement of neighborhood correctness.
  • Query-based global branch for whole-scan CT representation
    purpose: Aggregate disease-relevant volumetric evidence beyond average pooling or organ tokens alone.
    Architectural component; usefulness is only shown via end-task gains, not isolated external evidence.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning Anatomy-Grounded CT Vision-Language Representations with Organ-Hierarchical Report Knowledge." pith.science (2026). https://pith.science/paper/PITVCMKO

@misc{pith2026260710953,
  author       = {Pith},
  title        = {Pith review of: Learning Anatomy-Grounded CT Vision-Language Representations with Organ-Hierarchical Report Knowledge},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PITVCMKO}},
  note         = {Machine review of arXiv:2607.10953}
}
read the original abstract

Medical vision-language pretraining (VLP) from paired CT images and radiology reports enables scalable representation learning, but most existing methods align either whole scans with entire reports or local image regions with text fragments. These formulations underuse a key property of radiology reports: findings are organized around anatomical structures, with abnormalities described by organs, disease concepts, locations, and severity-related attributes. We propose OKA-CT, an organ-hierarchical knowledge-augmented framework for CT-report VLP. OKA-CT first converts free-text reports into organ-conditioned knowledge using radiology report parsing and LLM-assisted semantic structuring. The extracted hierarchy is used across two learning stages. Stage~1 injects anatomy-grounded evidence into the CT visual representation through fine-grained organ-conditioned supervision, while Stage~2 uses organ-specific report evidence to guide structured report-CT contrastive learning, where hierarchy-derived semantic soft targets treat non-paired cases with shared organ-level findings as weak semantic positives rather than uniform negatives. A lightweight query-based global branch further aggregates disease-relevant volumetric evidence for whole-scan representation. On CT-RATE and RAD-ChestCT datasets, OKA-CT achieves zero-shot abnormality diagnosis AUROCs of 84.9 and 72.2, outperforming prior CT VLP baselines. Retrieval and patch-occlusion analyses further show improved report-image alignment and stronger sensitivity to disease-associated anatomical regions.

Figures

Figures reproduced from arXiv: 2607.10953 by the authors.

Figure 1
Figure 1. From coarse report alignment to organ-hierarchical evidence learning. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of OKA-CT. OKA-CT extracts organ-hierarchical knowledge from reports using report parsing and LLM-assisted semantic structuring. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 4
Figure 4. Organ-hierarchical soft target. For each organ, report-derived ab [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figures from the paper (2 more)
Figure 5
Figure 5. Figure 5: Patch-occlusion visualization. Each organ block shows a representative case and compares the original CT slice, CT-CLIP, and OKA-CT. Warmer [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Organ-group deltas. Bars report the mean per-disease AUROC [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

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

  1. OrganLens: Organ-Specific Representation Learning for CT Foundation Models

    cs.CV 2026-07 conditional novelty 6.0 of 10

    OrganLens conditions a shared CT encoder on an organ identity and uses mask-supervised pooling to produce 11 organ-specific representations from the same volume, with organ-matched representations improving downstream...

Reference graph

Works this paper leans on

28 extracted references · 2 linked inside Pith · cited by 1 Pith paper

  1. [1]

    Developing generalist foundation models from a multimodal dataset for 3d computed tomography,

    I. E. Hamamci, S. Er, C. Wang, F. Almas, A. G. Simsek, S. N. Esirgun, I. Dogan, O. F. Durugol, B. Hou, S. Shitet al., “Developing generalist foundation models from a multimodal dataset for 3d computed tomography,”arXiv preprint arXiv:2403.17834, 2024

    arXiv 2024

  2. [2]

    Merlin: A vision language foundation model for 3d computed tomography,

    L. Blankemeier, J. P. Cohen, A. Kumar, D. Van Veen, S. J. S. Gardezi, M. Paschali, Z. Chen, J.-B. Delbrouck, E. Reis, C. Truytset al., “Merlin: A vision language foundation model for 3d computed tomography,” Research Square, pp. rs–3, 2024

    2024

  3. [3]

    Gloria: A multimodal global-local representation learning framework for label- efficient medical image recognition,

    S.-C. Huang, L. Shen, M. P. Lungren, and S. Yeung, “Gloria: A multimodal global-local representation learning framework for label- efficient medical image recognition,” inProceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 3942–3951

    2021

  4. [4]

    Im- proving medical vision-language contrastive pretraining with semantics- aware triage,

    B. Liu, D. Lu, D. Wei, X. Wu, Y . Wang, Y . Zhang, and Y . Zheng, “Im- proving medical vision-language contrastive pretraining with semantics- aware triage,”IEEE Transactions on Medical Imaging, vol. 42, no. 12, pp. 3579–3589, 2023

    2023

  5. [5]

    Ecamp: entity-centered context-aware medical vision language pre- training,

    R. Wang, Q. Yao, Z. Jiang, H. Lai, Z. He, X. Tao, and S. K. Zhou, “Ecamp: entity-centered context-aware medical vision language pre- training,”Medical Image Analysis, vol. 105, p. 103690, 2025

    2025

  6. [6]

    Large-scale and fine-grained vision-language pre-training for enhanced ct image understanding,

    Z. Shui, J. Zhang, W. Cao, S. Wang, R. Guo, L. Lu, L. Yang, X. Ye, T. Liang, Q. Zhanget al., “Large-scale and fine-grained vision-language pre-training for enhanced ct image understanding,” inThe Thirteenth International Conference on Learning Representations, 2025

    2025

  7. [7]

    Mg-3d: Multi-grained knowledge-enhanced 3d medical vision-language pre-training,

    X. Ni, L. Wu, J. Zhuang, Q. Wang, M. Wu, V . Vardhanabhuti, L. Zhang, H. Gao, and H. Chen, “Mg-3d: Multi-grained knowledge-enhanced 3d medical vision-language pre-training,”arXiv preprint arXiv:2412.05876, 2024

    arXiv 2024

  8. [8]

    Ct-glip: 3d grounded language-image pretraining with ct scans and radi- ology reports for full-body scenarios,

    J. Lin, Y . Xia, J. Zhang, K. Yan, K. Cao, L. Lu, J. Luo, and L. Zhang, “Ct-glip: 3d grounded language-image pretraining with ct scans and radi- ology reports for full-body scenarios,”arXiv preprint arXiv:2404.15272, 2024

    arXiv 2024

Show all 28 references
  1. [9]

    Radgraph-xl: A large-scale expert-annotated dataset for entity and re- lation extraction from radiology reports,

    J.-B. Delbrouck, P. Chambon, Z. Chen, M. Varma, A. Johnston, L. Blankemeier, D. Van Veen, T. Bui, S. Truong, and C. Langlotz, “Radgraph-xl: A large-scale expert-annotated dataset for entity and re- lation extraction from radiology reports,” inFindings of the Association for Co...

    2024

  2. [10]

    Learning transferable visual models from natural language supervision,

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clarket al., “Learning transferable visual models from natural language supervision,” inInternational Conference on Machine Learning. PMLR, 2021, pp. 8748–8763

    2021

  3. [11]

    Scaling up visual and vision-language representation learning with noisy text supervision,

    C. Jia, Y . Yang, Y . Xia, Y .-T. Chen, Z. Parekh, H. Pham, Q. Le, Y .-H. Sung, Z. Li, and T. Duerig, “Scaling up visual and vision-language representation learning with noisy text supervision,” inInternational conference on machine learning. PMLR, 2021, pp. 4904–4916

    2021

  4. [12]

    Align before fuse: Vision and language representation learning with momentum distillation,

    J. Li, R. Selvaraju, A. Gotmare, S. Joty, C. Xiong, and S. C. H. Hoi, “Align before fuse: Vision and language representation learning with momentum distillation,”Advances in neural information processing systems, vol. 34, pp. 9694–9705, 2021

    2021

  5. [13]

    Chestx-ray: Hospital-scale chest x-ray database and benchmarks on weakly-supervised classification and localization of common thorax diseases,

    X. Wang, Y . Peng, L. Lu, Z. Lu, M. Bagheri, and R. M. Summers, “Chestx-ray: Hospital-scale chest x-ray database and benchmarks on weakly-supervised classification and localization of common thorax diseases,” inProceedings of the IEEE conference on computer vision and pattern ...

    2017

  6. [14]

    Chexpert: A large chest radiograph dataset with uncertainty labels and expert comparison,

    J. Irvin, P. Rajpurkar, M. Ko, Y . Yu, S. Ciurea-Ilcus, C. Chute, H. Mark- lund, B. Haghgoo, R. Ball, K. Shpanskayaet al., “Chexpert: A large chest radiograph dataset with uncertainty labels and expert comparison,” inProceedings of the AAAI conference on artificial intelligenc...

    2019

  7. [15]

    Mimic-cxr, a de- identified publicly available database of chest radiographs with free-text reports,

    A. E. Johnson, T. J. Pollard, S. J. Berkowitz, N. R. Greenbaum, M. P. Lungren, C.-y. Deng, R. G. Mark, and S. Horng, “Mimic-cxr, a de- identified publicly available database of chest radiographs with free-text reports,”Scientific data, vol. 6, no. 1, p. 317, 2019

    2019

  8. [16]

    Preparing a collection of radiology examinations for distribution and retrieval,

    D. Demner-Fushman, M. D. Kohli, M. B. Rosenman, S. E. Shooshan, L. Rodriguez, S. Antani, G. R. Thoma, and C. J. McDonald, “Preparing a collection of radiology examinations for distribution and retrieval,” Journal of the American Medical Informatics Association, vol. 23, no. 2,...

    2016

  9. [17]

    Medclip: Contrastive learning from unpaired medical images and text,

    Z. Wang, Z. Wu, D. Agarwal, and J. Sun, “Medclip: Contrastive learning from unpaired medical images and text,” inProceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, 2022, pp. 3876–3887

    2022

  10. [18]

    Bootstrapping chest ct image understanding by distilling knowledge from x-ray expert models,

    W. Cao, J. Zhang, Y . Xia, T. C. W. Mok, Z. Li, X. Ye, L. Lu, J. Zheng, Y . Tang, and L. Zhang, “Bootstrapping chest ct image understanding by distilling knowledge from x-ray expert models,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024

    2024

  11. [19]

    Bridged semantic alignment for zero-shot 3d medical image diagnosis,

    H. Lai, Z. Jiang, Q. Yao, R. Wang, Z. He, X. Tao, W. Lv, W. Wei, and S. K. Zhou, “Bridged semantic alignment for zero-shot 3d medical image diagnosis,”IEEE Journal of Biomedical and Health Informatics, 2025

    2025

  12. [20]

    Joint learning of localized representations from medical images and reports,

    P. M ¨uller, G. Kaissis, C. Zou, and D. Rueckert, “Joint learning of localized representations from medical images and reports,” inEuropean conference on computer vision. Springer, 2022, pp. 685–701

    2022

  13. [21]

    Making the most of text semantics to improve biomedical vision– language processing,

    B. Boecking, N. Usuyama, S. Bannur, D. C. Castro, A. Schwaighofer, S. Hyland, M. Wetscherek, T. Naumann, A. Nori, J. Alvarez-Valleet al., “Making the most of text semantics to improve biomedical vision– language processing,” inEuropean conference on computer vision. Springer, ...

    2022

  14. [22]

    Bcnet: Bronchus classification via structure guided represen- tation learning,

    W. Huang, H. Gong, H. Zhang, Y . Wang, X. Wan, G. Li, H. Li, and H. Shen, “Bcnet: Bronchus classification via structure guided represen- tation learning,”IEEE Transactions on Medical Imaging, vol. 44, no. 1, pp. 489–498, 2024

    2024

  15. [23]

    Boundary as the bridge: Towards heterogeneous partially-labeled medical image segmentation and landmark detection,

    H. Gong, B. Wan, L. Kang, X. Wan, L. Zhang, and H. Li, “Boundary as the bridge: Towards heterogeneous partially-labeled medical image segmentation and landmark detection,”IEEE Transactions on Medical Imaging, 2025

    2025

  16. [24]

    Learning spatio-temporal features with 3d residual networks for action recognition,

    K. Hara, H. Kataoka, and Y . Satoh, “Learning spatio-temporal features with 3d residual networks for action recognition,” inProceedings of the IEEE international conference on computer vision workshops, 2017, pp. 3154–3160

    2017

  17. [25]

    Machine-learning-based multiple abnormality pre- diction with large-scale chest computed tomography volumes,

    R. L. Draelos, D. Dov, M. A. Mazurowski, J. Y . Lo, R. Henao, G. D. Rubin, and L. Carin, “Machine-learning-based multiple abnormality pre- diction with large-scale chest computed tomography volumes,”Medical image analysis, vol. 67, p. 101857, 2021

    2021

  18. [26]

    Compre- hensive language-image pre-training for 3d medical image understand- ing,

    T. Wald, I. E. Hamamci, Y . Gao, S. Bond-Taylor, H. Sharma, M. Ilse, C. Lo, O. Melnichenko, A. Schwaighofer, N. C. Codellaet al., “Compre- hensive language-image pre-training for 3d medical image understand- ing,”arXiv preprint arXiv:2510.15042, 2025

    2025

  19. [27]

    Totalsegmentator: robust segmentation of 104 anatomic structures in ct images,

    J. Wasserthal, H.-C. Breit, M. T. Meyer, M. Pradella, D. Hinck, A. W. Sauter, T. Heye, D. T. Boll, J. Cyriac, S. Yanget al., “Totalsegmentator: robust segmentation of 104 anatomic structures in ct images,”Radiology: Artificial Intelligence, vol. 5, no. 5, p. e230024, 2023

    2023

  20. [28]

    Qwen3 technical report,

    A. Yang, A. Li, B. Yang, B. Zhang, B. Hui, B. Zheng, B. Yu, C. Gao, C. Huang, C. Lvet al., “Qwen3 technical report,”arXiv preprint arXiv:2505.09388, 2025

    2025 arXiv

Pith tools

Reviewed July 14, 2026 · model on record in the stance chip above.