Pith. sign in

REVIEW 4 major objections 7 minor 19 references

Head and Neck Tumor Segmentation of MRI from Pre- and Mid-radiotherapy with Pre-training, Data Augmentation and Dual Flow UNet

T0 review · 4 major / 7 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A CT-pretrained dual-flow CNN segments head-and-neck tumors on T2-weighted MRI with 82.38% pre-RT and 72.53% mid-RT aggregated Dice scores.

desk verdict Credible external challenge scores and an honest write-up, but the ablation is too weak to attribute the gains to any single component. read the letter →

arxiv 2412.14846 v2 pith:HJGCJYV2 submitted 2024-12-19 eess.IV cs.AIcs.CV

Litingyu Wang , Wenjun Liao , Shichuan Zhang , Guotai Wang This is my paper · ORCID
classification eess.IVcs.AIcs.CV
keywords HNTS-MRG2024headandneckcancertumorsegmentationMRIpre-trainingMixUpdualencoderDicesimilaritycoefficient
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

The paper aims to show that a fully supervised CNN pipeline—CT pre-training with intensity matching, MixUp augmentation, and a dual-encoder network that fuses registered pre-treatment images and labels into the mid-treatment stream—can segment head-and-neck primary tumors (GTVp) and metastatic lymph nodes (GTVn) from T2-weighted MRI with aggregated Dice scores of 82.38% for pre-RT and 72.53% for mid-RT on the HNTS-MRG2024 test set. This matters because pixel-level tumor and node contours are required for adaptive radiotherapy planning, and MRI provides better soft-tissue contrast than the CT images most existing datasets provide. The paper's own ablations show the improvements are concentrated in lymph node segmentation; primary tumor gains are marginal and pre-training can hurt in some training folds.

What carries the argument

The Dual Flow UNet (DFUNet) is the paper's central architectural object: two separate encoders, one taking the single-channel mid-RT image and the other taking the registered pre-RT image with its label mask, merged at each resolution stage by a CNN-based cross-attention block that applies both channel and spatial attention (CBAM-style) to the secondary pre-RT features before injecting them into the primary mid-RT stream. For Task-1, the load-bearing mechanism is the training recipe: pre-training a VNet-like encoder-decoder on CT volumes after histogram matching and nonlinear intensity transforms, then fine-tuning on MRI with MixUp-generated mixed patches, using cross-entropy loss for mixed samples and a deep-supervised combination of cross-entropy and Dice loss for raw samples.

What would settle it

A head-to-head test on the HNTS-MRG2024 training set: train the identical Task-1 pipeline with and without CT pre-training, holding every other component fixed, and compare aggregated DSC for GTVp and GTVn across all five folds. If the no-pre-training model matches or beats the pre-trained model on GTVp in a majority of folds, the claimed benefit of cross-modal pre-training would not hold.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that combining (1) pre-training on the SegRap2023 CT dataset after histogram matching and nonlinear intensity transformations, (2) MixUp augmentation of training patches, and (3) a Dual Flow UNet whose second encoder ingests the registered pre-RT image and its mask through CNN-based cross-attention blocks, yields a competitive segmentation pipeline for the HNTS-MRG2024 challenge, with 82.38% aggregated DSC on pre-RT and 72.53% on mid-RT test data. The ablation tables show that adding the pre-RT image and label to the mid-RT input is the single largest factor for Task-2, lifting average aggregated DSC from 37.60% to 61.32% for GTVp, while the DFUNet architecture itself only surpasses the basic network in two of five folds. For Task-1, the combination of pre-training and MixUp improves GTVn by 1.30% on average while leaving GTVp essentially unchanged.

Load-bearing premise

The load-bearing premise is that pre-training on the SegRap2023 CT dataset, after histogram matching and nonlinear intensity transforms, transfers features that help T2-weighted MRI segmentation; the paper's own fold-level results show this transfer is inconsistent, lowering GTVp scores in some folds.

Editorial extensions

If this is right

  • If the reported test scores hold, this pipeline establishes a practical baseline for automated GTVp/GTVn contouring in adaptive radiotherapy workflows using T2-weighted MRI alone.
  • The large Task-2 gain from adding registered pre-RT images and labels implies that longitudinal imaging is a high-value input for mid-treatment segmentation, worth exploiting beyond simple input concatenation.
  • The observation that GTVn benefits more than GTVp from pre-training and MixUp suggests class- and structure-specific augmentation or loss strategies may be needed for hard classes.
  • The DFUNet's inconsistent gains indicate that two-stream attention fusion is not uniformly better than a single-stream network, and further architecture refinement is needed before it is reliable.

Reading between the lines

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

  • A natural extension the authors leave implicit is a standalone validation of the cross-modal pre-training: if the same fine-tuning is repeated with a modality-invariant pretraining objective (e.g., style augmentation or contrastive alignment), the GTVp gains might become consistent where the current intensity-matching transfer fails.
  • The dual-encoder design could transfer to other settings where registered baseline and follow-up scans exist, such as treatment response assessment or longitudinal tumor tracking, not just mid-RT radiotherapy.
  • The reversed difficulty ordering between modalities (GTVp easier than GTVn on CT, harder on MRI) suggests the intensity and contrast characteristics of each structure are modality-specific; a testable consequence is that reweighting the loss per structure or per fold should improve the weaker class's Dice without hurting the stronger one.
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 / 7 minor

Summary. The paper describes the authors' contribution to the HNTS-MRG2024 challenge for segmenting head and neck gross tumor volumes (GTVp) and metastatic lymph nodes (GTVn) on pre-radiotherapy and mid-radiotherapy T2-weighted MRI. For Task-1 they compare a basic VNet-like segmentation network trained with full supervision, pre-training on the CT-based SegRap2023 dataset with histogram matching and nonlinear intensity transforms, MixUp augmentation, and combinations of these. For Task-2 they introduce a Dual Flow UNet with two encoders and cross-attention to incorporate registered pre-RT images and their labels into the mid-RT segmentation stream. The best model per fold is selected from five-fold cross-validation and ensembled, yielding final test aggregated Dice scores of 82.38% for Task-1 and 72.53% for Task-2. The paper concludes that the proposed pre-training, MixUp, and DFUNet components contributed to the results.

Significance. If the causal claims were established, the paper would provide a useful practical recipe for MRI head and neck tumor segmentation: CT pre-training with intensity alignment, MixUp augmentation, and a dual-encoder network for using registered pre-RT images. The final test scores are measured on the challenge's external held-out test set, which gives objective support for the headline numbers. The authors make their code publicly available, use public datasets for pre-training, and include an honest discussion that acknowledges limitations. However, the internal evidence for the causal contribution of the proposed components is weak: the five-fold ablations are single runs without variance estimates, pre-training is inconsistent and sometimes harmful, and DFUNet is on average worse than the simpler baseline. These issues are load-bearing because the abstract and introduction attribute the final test performance to the proposed components. The manuscript needs sharper statistical evidence and more careful framing before the contribution claims can be accepted.

major comments (4)
  1. [Section 4, Tables 1 and 2] The five-fold cross-validation results are reported without variance, confidence intervals, or significance tests, with only a single run per condition. The claimed improvements are small relative to fold-to-fold variation: for Task-1, Pre-train+MixUp improves average GTVp by only 0.47 percentage points over Base (76.03 vs 75.56), and for GTVn by 1.30 points (85.12 vs 83.82). Fold-level differences, such as Fold 5 where Pre-train+MixUp reduces GTVp from 80.39 to 78.82, are of comparable or larger magnitude. The Discussion's statement that the strategies led to 'significant improvements' is therefore not supported. Please add repeated experiments or a statistical analysis, or soften the claims to describe observed trends in this specific dataset.
  2. [Section 2.2 and Table 1] The claim that pre-training on SegRap2023 CT data transfers useful features to T2-weighted MRI is not consistently supported. With pre-training alone, the average GTVp score is lower than Base (75.23 vs 75.56), and Pre-train+MixUp decreases GTVp relative to Base in Fold 5 (78.82 vs 80.39). The Discussion itself concedes that 'pre-training is not universally effective'. Because the final Task-1 ensemble is a per-fold selection among Base, MixUp, and Pre-train+MixUp, the contribution of pre-training to the 82.38% test score is unclear. Please report results separately for GTVp and GTVn by fold, state explicitly in which folds pre-training was selected, and avoid presenting pre-training as a uniformly beneficial component.
  3. [Section 2.1, Table 2, and Section 5] DFUNet, presented as a novel contribution in the introduction, underperforms the simpler Base+pre-RT model on average for both GTVp (60.37 vs 61.32) and GTVn (86.58 vs 86.65). The Discussion explicitly states that 'the DFUNet underperformed compared to the basic segmentation model in cross-validation'. The text also notes that DFUNet outperforms Base+pre-RT in folds 1 and 4; if those are the selected folds, the final ensemble contains two DFUNet models, not one. The manuscript should state the selection criterion precisely and quantify the contribution of DFUNet to the final 72.53% score, or it should temper the abstract and conclusion so that DFUNet is presented as a promising but not yet validated architecture.
  4. [Section 4, 'selected the highest-performing model'] The criterion for selecting the best model in each fold is not defined. It is unclear whether selection is based on aggregated DSC over both structures, on GTVp only, on GTVn only, or on a subjective compromise. The bold markers in Tables 1 and 2 are not visible in the text, and the caption says only that bolding indicates the selected models. Because the final ensemble depends on this selection, the criterion must be stated explicitly. Also, selecting the best model per fold on the validation set introduces optimism in the reported cross-validation average of 80.65% and 74.68%; please acknowledge this or report the non-oracle average.
minor comments (7)
  1. [Abstract] The phrase 'as HiLab' in the abstract is unclear; please clarify whether it means 'at HiLab' or 'on the HiLab evaluation platform'.
  2. [Section 1] In the second paragraph, 'per-RT images' appears to be a typo for 'pre-RT images'.
  3. [Section 2.4] Equation (2) sums over d from 0 to l, but the number of deep-supervision resolutions l is not specified in the implementation details. Please state the value used.
  4. [Section 3.2] The preprocessing threshold of intensity greater than 60 for the body mask and the morphological operation parameters are given without rationale or sensitivity analysis; a brief justification would improve reproducibility.
  5. [Figures 5 and 6] The captions say that cases are classified as well-predicted or poor-predicted based on 'whether they meet the method's improvements', which is circular. Please define an objective quantitative criterion for case selection.
  6. [Section 2.2 and Figure 3] The nonlinear intensity transformations are visualized in Figure 3, but the reference to 'Bézier Curve [8]' is not explained in the text; please clarify how the Bézier curve is used and why it is cited.
  7. [Table 2] The caption states that 'Pre-train+MixUp' uses the basic segmentation network, but the text in Section 2 for Task-2 lists only DFUNet as an additional training method. Please align the method list with the table entries.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper reports an empirical challenge submission evaluated on an external test set, with no derivation that reduces to fitted parameters or self-citations.

full rationale

The paper is a benchmark/systems report rather than a theoretical derivation. Its claims are empirical: five-fold cross-validation on the HNTS-MRG2024 training set and a single submission to the external challenge test set (Section 4, Table 3). The final test scores (82.38% for Task-1 and 72.53% for Task-2) are measured against held-out challenge data, so they cannot be forced by construction from training-set fits. Pre-training uses the public SegRap2023 CT dataset with histogram matching and nonlinear intensity transformations; the method does not define the reported DSC metric in terms of the pre-training procedure. The only self-citation is reference [8] for a Bézier-curve augmentation used during pre-training, and this is a routine technique citation, not a load-bearing assertion. The paper's own Discussion explicitly concedes negative results ('the DFUNet underperformed compared to the basic segmentation model' and 'pre-training is not universally effective'), which is inconsistent with a circular argument intended to force favorable conclusions. No fitted parameter is renamed as a prediction, no uniqueness theorem is imported from the authors' prior work, and the architecture choices are not defined in terms of the reported DSC. Thus the derivation chain, such as it is, is self-contained with respect to circularity; any concerns about statistical significance or causal attribution are correctness and evidence issues, not circularity.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The central claim is an empirical benchmark result, so there are no invented entities or derivations. The free parameters are standard training hyperparameters and preprocessing choices, none of which are used to construct the benchmark itself. The key domain assumptions concern transferability of CT pre-training and representativeness of the challenge data.

free parameters (6)
  • Learning rate, momentum, weight decay = 0.01, 0.99, 3e-5
    Chosen following nnUNetv2 defaults and SGD practice (Section 3.4); these affect optimization and final performance but are not derived.
  • Patch size and batch size = 56x224x160 and 2 (4 with MixUp)
    Chosen to fit GPU memory constraints (Section 3.4).
  • Number of training epochs = 1000
    Set to ensure convergence (Section 3.4); not tuned per fold.
  • Intensity threshold for body mask = 60
    Hand-selected in preprocessing to define the region of interest (Section 3.2).
  • Histogram matching reference image = Randomly selected pre-RT MRI from HNTS-MRG2024
    The choice of reference image affects the CT intensity mapping during pre-training (Section 3.2), yet it is chosen randomly.
  • MixUp beta distribution parameter = Not specified
    The paper states lambda is sampled from the Beta distribution but does not provide the Beta parameters (Section 2.3).
assumptions (5)
  • domain assumption Pre-training on CT with histogram matching improves downstream MRI segmentation
    The paper relies on this to justify Task-1 gains, but acknowledges it can hurt performance (Section 2.2 and Discussion).
  • domain assumption The training distribution of HNTS-MRG2024 is representative of the challenge test distribution
    Needed for the reported final test scores to be meaningful; standard for benchmark papers but unproven.
  • standard math CNN optimization via SGD converges to a useful solution
    All deep learning results rely on stochastic optimization; not proven in this paper (Section 3.4).
  • domain assumption Aggregated Dice coefficient is the appropriate evaluation metric
    The paper adopts the challenge's metric without justification (Section 3.4).
  • standard math MixUp linear interpolation creates valid training samples
    The paper uses Eq. 1 to define mixed samples and assumes the mixed labels are meaningful (Section 2.3).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Head and Neck Tumor Segmentation of MRI from Pre- and Mid-radiotherapy with Pre-training, Data Augmentation and Dual Flow UNet." pith.science (2026). https://pith.science/paper/HJGCJYV2

@misc{pith2026241214846,
  author       = {Pith},
  title        = {Pith review of: Head and Neck Tumor Segmentation of MRI from Pre- and Mid-radiotherapy with Pre-training, Data Augmentation and Dual Flow UNet},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HJGCJYV2}},
  note         = {Machine review of arXiv:2412.14846}
}
read the original abstract

Head and neck tumors and metastatic lymph nodes are crucial for treatment planning and prognostic analysis. Accurate segmentation and quantitative analysis of these structures require pixel-level annotation, making automated segmentation techniques essential for the diagnosis and treatment of head and neck cancer. In this study, we investigated the effects of multiple strategies on the segmentation of pre-radiotherapy (pre-RT) and mid-radiotherapy (mid-RT) images. For the segmentation of pre-RT images, we utilized: 1) a fully supervised learning approach, and 2) the same approach enhanced with pre-trained weights and the MixUp data augmentation technique. For mid-RT images, we introduced a novel computational-friendly network architecture that features separate encoders for mid-RT images and registered pre-RT images with their labels. The mid-RT encoder branch integrates information from pre-RT images and labels progressively during the forward propagation. We selected the highest-performing model from each fold and used their predictions to create an ensemble average for inference. In the final test, our models achieved a segmentation performance of 82.38% for pre-RT and 72.53% for mid-RT on aggregated Dice Similarity Coefficient (DSC) as HiLab. Our code is available at https://github.com/WltyBY/HNTS-MRG2024_train_code.

Figures

Figures reproduced from arXiv: 2412.14846 by the authors.

Figure 1
Figure 1. Two architectures used for training. (a) An encoder-decoder network, named basic segmentation network, with MixUp diagram. (b) Two separate encoders with one decoder named Dual Flow UNet (DFUNet). For Task-2, we introduced an additional training method: • Fully supervised learning employing the Dual Flow UNet (DFUNet) archi￾tecture, illustrated in [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Details of the network’s configuration are elaborated in Section 3.3. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 2
Figure 2. CNN-based cross attention block to integrate secondary information fpre into primary information fmid [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figures from the paper (4 more)
Figure 3
Figure 3. Figure 3: Raw image (the first column) and corresponding augmented images using Bézier Curve [8]. 2.2 Model Pre-training We utilized the SegRap2023 Challenge Dataset [9] and the basic segmentation network depicted in Fig.1 (a) for fully supervised pre-training. However, since Se…
Figure 4
Figure 4. Figure 4: Histogram matching on the SegRap2023 dataset. (a) shows the source image and its grayscale histogram from the SegRap2023 dataset. (b) displays the reference image from the HNTS-MRG2024 dataset, and (c) presents the matched image. HNTS-MRG2024 Challenge Dataset Head and…
Figure 5
Figure 5. Figure 5: Examples of Task-1 illustrate the segmentation of GTVp (red) and GTVn (green) with mean DSC values (numbers in white). (a) represents a well-predicted case, while (b) shows a poorly predicted one. Classification of samples as well- or poor￾predicted here refers to whet…
Figure 6
Figure 6. Figure 6: Examples of Task-2 illustrate the segmentation of GTVp (red) and GTVn (green) with mean DSC values (numbers in white). (a) represents a well-predicted case, while (b) is the bad one. ing category, like GTVp, have been marginal and could even introduce adverse effects, …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

19 extracted references · 9 canonical work pages

  1. [1]

    In: 3D head and neck tumor segmentation in PET/CT challenge, pp

    Andrearczyk, V., Oreiller, V., Boughdad, S., Rest, C.C.L., Elhalawani, H., Jreige, M., Prior, J.O., Vallières, M., Visvikis, D., Hatt, M., et al.: Overview of the heck- tor challenge at miccai 2021: automatic head and neck tumor segmentation and outcome prediction in pet/ct images. In: 3D head and neck tumor segmentation in PET/CT challenge, pp. 1–37. Spr...

    2021

  2. [2]

    International journal of computer assisted radiology and surgery 16, 529–542 (2021)

    Badrigilan, S., Nabavi, S., Abin, A.A., Rostampour, N., Abedi, I., Shirvani, A., Ebrahimi Moghaddam, M.: Deep learning approaches for automated classification andsegmentationofheadandneckcancersandbraintumorsinmagneticresonance images: a meta-analysis study. International journal of computer assisted radiology and surgery 16, 529–542 (2021)

    work page 2021

  3. [3]

    International Journal of Radiation Oncology* Biology* Physics81(4), e439–e445 (2011)

    Brændengen, M., Hansson, K., Radu, C., Siegbahn, A., Jacobsson, H., Glimelius, B.: Delineation of gross tumor volume (gtv) for radiation treatment planning of locally advanced rectal cancer using information from mri or fdg-pet/ct: a prospec- tive study. International Journal of Radiation Oncology* Biology* Physics81(4), e439–e445 (2011)

    work page 2011

  4. [4]

    arXiv preprint arXiv:2102.04306 (2021)

    Chen, J., Lu, Y., Yu, Q., Luo, X., Adeli, E., Wang, Y., Lu, L., Yuille, A.L., Zhou, Y.:Transunet:Transformersmakestrongencodersformedicalimagesegmentation. arXiv preprint arXiv:2102.04306 (2021)

    arXiv 2021

  5. [5]

    Clinical radiology 73(1), 45–59 (2018)

    Dai, Y., King, A.: State of the art mri in head and neck cancer. Clinical radiology 73(1), 45–59 (2018)

    work page 2018

  6. [6]

    Nature methods 18(2), 203–211 (2021)

    Isensee, F., Jaeger, P.F., Kohl, S.A., Petersen, J., Maier-Hein, K.H.: nnu-net: a self-configuring method for deep learning-based biomedical image segmentation. Nature methods 18(2), 203–211 (2021)

    2021

  7. [7]

    Medical Image Analysis 72, 102135 (2021) 12 Wang et al

    Liu, J., Liu, H., Gong, S., Tang, Z., Xie, Y., Yin, H., Niyoyita, J.P.: Automated cardiac segmentation of cross-modal medical images using unsupervised multi- domain adaptation and spatial neural attention structure. Medical Image Analysis 72, 102135 (2021) 12 Wang et al

    work page 2021

  8. [8]

    In: 2024 IEEE International Symposium on Biomedical Imaging (ISBI)

    Liu, X., Wu, J., Luo, X., Liao, W., Zhang, S., Zhang, S., Wang, G.: Rpl-sfda: Reliable pseudo label-guided source-free cross-modality adaptation for npc gtv segmentation. In: 2024 IEEE International Symposium on Biomedical Imaging (ISBI). pp. 1–5. IEEE (2024)

    work page 2024

Show all 19 references
  1. [9]

    arXiv preprint arXiv:2312.09576 (2023)

    Luo, X., Fu, J., Zhong, Y., Liu, S., Han, B., Astaraki, M., Bendazzoli, S., Toma- Dasu, I., Ye, Y., Chen, Z., et al.: Segrap2023: A benchmark of organs-at-risk and gross tumor volume segmentation for radiotherapy planning of nasopharyngeal carcinoma. arXiv preprint arXiv:2312....

    2023 arXiv

  2. [10]

    In: 2016 fourth international confer- ence on 3D vision (3DV)

    Milletari, F., Navab, N., Ahmadi, S.A.: V-net: Fully convolutional neural networks for volumetric medical image segmentation. In: 2016 fourth international confer- ence on 3D vision (3DV). pp. 565–571. Ieee (2016)

    2016

  3. [11]

    Current treatment options in oncology7, 23–34 (2006)

    Rumboldt, Z., Gordon, L., Bonsall, R., Ackermann, S.: Imaging in head and neck cancer. Current treatment options in oncology7, 23–34 (2006)

    2006

  4. [12]

    Concepts in Magnetic Reso- nance Part A2019(1), 4282754 (2019)

    Sager, O., Dincoglan, F., Demiral, S., Gamsiz, H., Uysal, B., Ozcan, F., Colak, O., Dirican, B., Beyzadeoglu, M.: Evaluation of the impact of magnetic resonance imaging (mri) on gross tumor volume (gtv) definition for radiation treatment plan- ning (rtp) of inoperable high gra...

    2019

  5. [13]

    In: 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

    Sun, J., Dai, Y., Zhang, X., Xu, J., Ai, R., Gu, W., Chen, X.: Efficient spatial- temporal information fusion for lidar-based 3d moving object segmentation. In: 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). pp. 11456–11463. IEEE (2022)

    2022

  6. [14]

    Advances in Neural Information Processing Systems (2017)

    Vaswani, A.: Attention is all you need. Advances in Neural Information Processing Systems (2017)

    2017

  7. [15]

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

    Wang, H., Ma, C., Zhang, J., Zhang, Y., Avery, J., Hull, L., Carneiro, G.: Learnable cross-modal knowledge distillation for multi-modal learning with missing modality. In: International Conference on Medical Image Computing and Computer-Assisted Intervention. pp. 216–226. Spri...

    2023

  8. [16]

    In: Proceedings of the European conference on computer vision (ECCV)

    Woo, S., Park, J., Lee, J.Y., Kweon, I.S.: Cbam: Convolutional block attention module. In: Proceedings of the European conference on computer vision (ECCV). pp. 3–19 (2018)

    2018

  9. [17]

    In: 2024 International Joint Conference on Neural Networks (IJCNN)

    Xia, Y., Feng, S., Zhao, J., Yuan, Z.: Robust cross-modal medical image transla- tion via diffusion model and knowledge distillation. In: 2024 International Joint Conference on Neural Networks (IJCNN). pp. 1–8. IEEE (2024)

    2024

  10. [18]

    arXiv preprint arXiv:1710.09412 (2017)

    Zhang, H.: mixup: Beyond empirical risk minimization. arXiv preprint arXiv:1710.09412 (2017)

    2017 arXiv

  11. [19]

    In: Proceed- ings of the IEEE/CVF conference on computer vision and pattern recognition

    Zhou, Z., Qi, L., Yang, X., Ni, D., Shi, Y.: Generalizable cross-modality medical image segmentation via style augmentation and dual normalization. In: Proceed- ings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 20856–20865 (2022)

    2022

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.