Pith. sign in

REVIEW 4 major objections 8 minor 20 references

HResFormer: Hybrid Residual Transformer for Volumetric Medical Image Segmentation

T0 review · 4 major / 8 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 The paper claims that combining a 2D Transformer that reads each CT/MRI slice closely with a 3D Transformer that reads the whole volume, joined by a mutual local-global fusion module and residual prediction addition, sets a new state of…

desk verdict A promising hybrid 2D/3D transformer for volumetric segmentation, but adopted baselines and internal reporting errors leave the SOTA claim unverified. read the letter →

arxiv 2412.11458 v1 pith:AAXVJXPR submitted 2024-12-16 cs.CV

Sucheng Ren , Xiaomeng Li This is my paper · ORCID
classification cs.CV
keywords volumetricmedicalimagesegmentationhybridTransformer2D-3DfeaturefusionresiduallearningLocal-GlobalModuleself-attentioncomputedtomographymagneticresonanceimaging
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 the best way to segment 3D medical scans is to combine two Transformers rather than choose between them: a 2D Transformer that examines each slice closely and a 3D Transformer that understands how slices connect into anatomy. On three benchmarks (Synapse multi-organ CT, brain-tumor MRI, cardiac MRI), the proposed HResFormer reports the highest Dice scores, beating the previous Transformer-based method nnFormer by 2.5% on Synapse and the convolutional method nnU-Net by 2.1%. The argument is that radiologists themselves work this way—read the axial plane, then review coronal views to build a 3D picture—so a network that mirrors that workflow should segment better than one that only sees slices or only sees volumes. If true, it would shift design practice: hybrid architectures become a viable default for volumetric segmentation, and the specific fusion and residual mechanisms would be reusable components.

What carries the argument

The Hybrid Local-Global Fusion Module (HLGM) is the load-bearing component. It takes features from the 2D prediction volume and the raw 3D volume and performs mutual cross-feature attention twice: once in non-overlapping 3D windows (local mutual fusion) and once with spatially reduced key and value projections (global mutual fusion), then passes the sum through a Cross-Position Feed-Forward layer that uses depth-wise convolution. A second mechanism, residual learning, sets the final prediction as the 2D prediction plus the 3D prediction, letting the 3D model refine rather than replace the 2D segmentation. Together they let the 3D Transformer use the 2D prior without discarding its own volumetric context.

What would settle it

Take the exact Synapse 8-class split used in Table I, retrain nnFormer and nnU-Net with the same augmentation, optimizer, and 1000-epoch schedule the paper uses for HResFormer, and compare average DSC; if the gaps shrink to near zero or change sign, the claimed benefit of the hybrid architecture is not established.

Watch

Extended reading notes

Core claim

The central claim is that a hybrid residual Transformer, built from standard 2D and 3D Transformer backbones, outperforms both pure 2D and pure 3D methods on volumetric medical image segmentation. The 2D branch produces fine-grained per-slice predictions; the 3D branch consumes the original volume together with those predictions, fused through a Hybrid Local-Global fusion Module (HLGM) that mutually exchanges features in local windows and through globally reduced attention. The final output is the 3D model's prediction plus the 2D prediction, i.e. residual learning of the whole volume relative to the slice-level baseline. The paper reports state-of-the-art results on Synapse (89.10% average DSC for 8 classes and 87.46% for 13 organs), BraTS (79.5% average DSC), and ACDC (91.88% DSC), with an ablation study showing that the fusion module contributes about 1% and residual learning about 0.9% on Synapse.

Load-bearing premise

The reported margins assume the comparison baselines, especially nnFormer and nnU-Net, were trained under exactly the same data splits, preprocessing, and training schedule as HResFormer, but the paper adopts those baseline numbers from earlier publications rather than rerunning them in its own protocol.

Editorial extensions

If this is right

  • On Synapse multi-organ CT with 8 classes, HResFormer reports 89.10% average DSC, 2.5 points above nnFormer and 2.1 points above nnU-Net.
  • On the 13-organ Synapse split, HResFormer outperforms both CoTR and nnU-Net on average DSC, and beats nnU-Net on all 13 individual organs.
  • HResFormer keeps parameters comparable to other Transformer methods (117M vs. 150M for nnFormer) and lowers FLOPs from 157.9G to 131.7G, with a lightweight variant at 63.4G still surpassing nnFormer.
  • On BraTS brain tumor MRI, HResFormer surpasses nnU-Net by 1.8% and nnFormer by 2.1% average DSC; on ACDC it surpasses nnU-Net by 0.52 points.
  • Ablation studies on Synapse show that both HLGM and residual learning contribute independently to the reported gain, with HLGM adding roughly 1% and residual learning roughly 0.9%.

Reading between the lines

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

  • The hybrid design may generalize outside segmentation: any 3D medical task where a 2D high-resolution view matters, such as small-lesion detection or classification, could benefit from the same 2D-prior-plus-3D-context structure.
  • A direct testable variant would swap the 2D Transformer for a convolutional 2D network and keep the fusion and residual machinery; that would isolate whether the gain comes from the hybrid topology itself or from Transformer attention.
  • Because the paper adopts baseline numbers from external publications rather than rerunning them in its own protocol, the honest comparison requires in-house reproduction of nnFormer and nnU-Net under identical splits and training schedules; until then the true margin may be smaller than reported.
  • Since the 3D attention is quadratic in volume size, the hybrid gain might be preserved at lower cost by substituting a linear-complexity sequence model for the 3D self-attention, a direction the authors themselves flag as future work.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 8 minor

Summary. The paper proposes HResFormer, a hybrid 2D+3D Transformer for volumetric medical image segmentation. The 2D branch processes axial slices to capture fine-grained inner-slice information, and the 3D branch processes volumetric data with a Hybrid Local-Global fusion Module (HLGM) that mutually fuses the two streams. Residual learning adds the 2D predictions to the 3D output. The method is evaluated on Synapse multi-organ CT, BraTS brain tumor MRI, and ACDC cardiac MRI, reporting state-of-the-art DSC compared to published Transformer and CNN baselines.

Significance. The paper introduces a novel hybrid architecture that combines a 2D Transformer for fine-grained axial-slice features with a 3D Transformer for volumetric context, using a mutual fusion module (HLGM) and residual prediction addition. If the reported results are reproducible under controlled conditions, this is a valuable contribution with practical relevance, as it directly addresses a known weakness of 2D and 3D segmentation models. The manuscript provides extensive experiments on three benchmarks, includes parameter and FLOPs comparisons, and reports a lightweight variant with a favorable accuracy/cost trade-off; the authors also commit to releasing code. However, the evidence for the headline SOTA claim is weakened by the reliance on baselines adopted from other papers, by an internally inconsistent ablation row in Table VI, and by a false per-organ claim in Section IV.C. These issues need to be resolved before the results can be fully credited.

major comments (4)
  1. [Section IV.C, Tables I and III] The state-of-the-art comparisons in Tables I and III are adopted from prior publications (nnFormer for Table I, Hatamizadeh et al. for Table III) rather than rerun under the training protocol described in Section IV.B (1000 epochs, SGD with poly decay, specific augmentations, ImageNet-pretrained 2D encoder). Because HResFormer is trained under this specific protocol, the reported margins (e.g., 2.5% DSC over nnFormer on Synapse) may partly reflect differences in training schedules, data preprocessing, and hyperparameters rather than architectural superiority. To support the central SOTA claim, the authors should either rerun at least nnFormer and nnU-Net under the identical protocol (same data splits, preprocessing, optimizer, and training schedule) or substantially temper the claim and clearly state that all baselines are taken from their original papers with potentially different protocols.
  2. [Table VI, Section IV.D] The row labeled 'HResFormer (only 3D)' in Table VI reports per-organ DSC values identical to the nnFormer row in Table I (Ao 92.04, Gb 70.17, Ki(L) 86.57, Ki(R) 86.25, Li 96.84, Pa 83.35, Sp 90.51, St 86.83), yet it lists an average of 85.39 instead of the correct average 86.57. This is internally inconsistent and suggests the row was copied from the nnFormer results rather than being an independently trained HResFormer 3D branch. This undermines the ablation discussion in Section IV.D, where the authors argue that their 3D model alone is comparable to nnFormer and that the hybrid design is responsible for the gains.
  3. [Section IV.C, Table II] The text states that 'our HResFormer consistently outperforms nnUnet on all 13 organs', but Table II shows that nnUnet achieves higher DSC than HResFormer on Spleen (96.64 vs 96.40), right kidney (94.74 vs 94.58), and liver (97.05 vs 96.85). This factual error must be corrected; the average gain over nnUnet is still positive, but the claim of universal per-organ superiority is false.
  4. [Section IV.D, Tables VI and VII] The ablation study would benefit from variance estimates. Several differences used to support the design choices are small (e.g., 0.38 DSC between Ours-FF and Ours-DFF in Table VII; 0.98 DSC between the full model and the model without HLGM in Table VI), and no multiple-seed runs, confidence intervals, or significance tests are reported. Without these, it is difficult to determine whether the improvements are systematic or within run-to-run noise, and the ablation conclusions are not fully supported.
minor comments (8)
  1. [Section IV.D] In the sentence 'With residual learning, we can see that the performance consistently improves from 88.21% and 89.10%', the word 'and' should be 'to'.
  2. [Abstract/Throughout] The module name is given as HLGM, but it appears as 'HGLM' in the configuration column of Table VI and in the ablation text of Section IV.D; please use a consistent acronym.
  3. [Section I] In the third contribution bullet, 'Our method consistently previous outperforms' should read 'Our method consistently outperforms previous'.
  4. [Section IV.C] The typo 'HRFomrer' appears in the comparison text on the Synapse results; it should be 'HResFormer'.
  5. [Section III.D, Equation (10)] The equation P'_3d,i = P_2d,0 + P_3d,i contains a stray 'S2d' after the equation that should be removed.
  6. [Section I / Related Work] The paper claims to be 'the first hybrid Transformer' for 3D medical image segmentation, but PHTrans (Liu et al., 2022a), discussed in the Related Work, is also a hybrid architecture (parallel Transformer and CNN). Please qualify the novelty claim, e.g., as the first hybrid 2D+3D Transformer.
  7. [Section V] The conclusion lists 'a) Limitations.' and 'b) Future work.' as sub-items; this formatting is unconventional and should be integrated into the conclusion or moved to a separate section.
  8. [Table II] The reported average DSC for HResFormer (87.46) differs from the arithmetic mean of the listed per-organ values (approximately 87.61). Please check whether the per-organ values or the average are rounded and clarify.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: HResFormer's reported gains are empirical comparisons on fixed benchmarks, not reductions to fitted inputs or self-citations.

full rationale

This is an empirical architecture paper, not a derivation from first principles. The method is a fixed composition of a 2D Transformer, a 3D Transformer, HLGM fusion, and residual addition, described by explicit equations (Eqs. 1-10) with no fitted constants that are later relabeled as predictions. The central claim, that HResFormer outperforms prior methods on Synapse, BraTS, and ACDC, is supported by training on fixed public benchmarks and comparing with external baseline numbers. The baselines in Tables I and III are explicitly adopted from prior work (nnFormer and Hatamizadeh et al.), which is a legitimate although imperfect experimental practice; adopted external numbers are not derived from HResFormer's own equations, so this is a protocol-fairness concern, not circularity. The ablation table contains a suspicious row: the per-organ values of 'HResFormer (only 3D)' in Table VI are identical to the nnFormer row of Table I while the stated average differs (e.g., Ao 92.04, Gb 70.17, Ki(L) 86.57, ..., average 85.39 vs 86.57). This appears to be a copying or labeling error and weakens the ablation evidence, but it does not constitute a circular derivation: the hybrid method's performance is not defined in terms of nnFormer's numbers. Similarly, the paper's self-citations to prior pyramid-transformer work are used as architectural inspiration rather than as load-bearing justification for the reported improvements, and the text contains some overclaims (e.g., 'consistently outperforms nnUnet on all 13 organs' contradicts Table II for several organs), but these are correctness and reporting issues, not circularity. No step in the paper reduces, by construction or by self-citation, to its own inputs.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

This ledger lists the unspecified hyperparameters and comparability assumptions that the reported benchmark numbers rest on. Because the paper's claim is empirical, the main risks are unstated design choices and reliance on adopted baseline results.

free parameters (5)
  • 2D transformer depth and width per stage (Li, Ci) = not reported
    The 2D backbone has 4 stages at 4x, 8x, 16x, 32x downsample, but the number of transformer blocks Li and channel counts Ci are not given; the reported accuracy depends on these choices.
  • 3D local window size wd = not reported
    L3D-MSA in Eq. (3) splits features into non-overlap regions of size H/wd by W/wd by D/wd; wd is never specified.
  • Key and Value spatial reduction ratio r = not reported
    GR-Attention in Eq. (1) and GMF in Eqs. (7)-(8) reduce spatial size by factor r; the per-stage values are omitted.
  • CE and Dice loss weights = not reported
    Deep supervision applies cross-entropy and dice losses on 8 outputs (Sec. III.D), but the relative weighting is not stated.
  • Patch embedding and depthwise convolution settings = not reported
    The convolutional patch embedding and the CPFF layer use depthwise convolutions; kernel sizes and strides are omitted.
assumptions (4)
  • domain assumption Baselines in Tables I and III are correctly adopted from their source papers and are directly comparable to HResFormer under the same data splits and preprocessing.
    Section IV.C: 'The results are adopted from nnFormer (Zhou et al., 2023)' and 'Experimental results of baselines are adopted from (Hatamizadeh et al., 2022a)'. If training protocols differ, the reported gains may not reflect architecture alone.
  • domain assumption The public benchmark datasets are used with the standard splits from the cited settings.
    Section IV.A: 'We follow the same setting of (Zhou et al., 2023)' and 'the settings of nnUnet'. Exact case lists are not given.
  • domain assumption The DSC evaluation is computed consistently with the baseline papers.
    No evaluation code is shared; any difference in metric implementation (e.g., foreground-only vs. all-class Dice, handling of empty labels) could shift scores.
  • domain assumption The radiologist reading process (axial first, then coronal review) is a valid motivation for the architecture.
    Introduction paragraph 3: 'radiologists primarily use the axial plane and then routinely review both axial and coronal planes'. This is plausible but not independently established in the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of HResFormer: Hybrid Residual Transformer for Volumetric Medical Image Segmentation." pith.science (2026). https://pith.science/paper/AAXVJXPR

@misc{pith2026241211458,
  author       = {Pith},
  title        = {Pith review of: HResFormer: Hybrid Residual Transformer for Volumetric Medical Image Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AAXVJXPR}},
  note         = {Machine review of arXiv:2412.11458}
}
read the original abstract

Vision Transformer shows great superiority in medical image segmentation due to the ability in learning long-range dependency. For medical image segmentation from 3D data, such as computed tomography (CT), existing methods can be broadly classified into 2D-based and 3D-based methods. One key limitation in 2D-based methods is that the intra-slice information is ignored, while the limitation in 3D-based methods is the high computation cost and memory consumption, resulting in a limited feature representation for inner-slice information. During the clinical examination, radiologists primarily use the axial plane and then routinely review both axial and coronal planes to form a 3D understanding of anatomy. Motivated by this fact, our key insight is to design a hybrid model which can first learn fine-grained inner-slice information and then generate a 3D understanding of anatomy by incorporating 3D information. We present a novel \textbf{H}ybrid \textbf{Res}idual trans\textbf{Former} \textbf{(HResFormer)} for 3D medical image segmentation. Building upon standard 2D and 3D Transformer backbones, HResFormer involves two novel key designs: \textbf{(1)} a \textbf{H}ybrid \textbf{L}ocal-\textbf{G}lobal fusion \textbf{M}odule \textbf{(HLGM)} to effectively and adaptively fuse inner-slice information from 2D Transformer and intra-slice information from 3D volumes for 3D Transformer with local fine-grained and global long-range representation. \textbf{(2)} a residual learning of the hybrid model, which can effectively leverage the inner-slice and intra-slice information for better 3D understanding of anatomy. Experiments show that our HResFormer outperforms prior art on widely-used medical image segmentation benchmarks. This paper sheds light on an important but neglected way to design Transformers for 3D medical image segmentation.

Figures

Figures reproduced from arXiv: 2412.11458 by the authors.

Figure 1
Figure 1. Overview of our HResFormer. It consists of three main parts: a 2D model, a Hybrid Local-Global fusion Module (HLGM), and a 3D model. The 3D [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. The overview of Hybrid Local-Global Fusion Module. We use one [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Examples of multi-organ segmentation results. “GT” refers to the ground-truth. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

20 extracted references · 14 canonical work pages

  1. [3]

    Utnet: a hybrid transformer ar- chitecture for medical image segmentation,

    Y . Gao, M. Zhou, and D. N. Metaxas, “Utnet: a hybrid transformer ar- chitecture for medical image segmentation,” in MICCAI. Springer, 2021, pp. 61–71. H. Cao, Y . Wang, J. Chen, D. Jiang, X. Zhang, Q. Tian, and M. Wang, “Swin-unet: Unet-like pure transformer for medical image segmentation,” in ECCV Workshops,

    work page 2021

  2. [5]

    U-Net: Convolutional net- works for biomedical image segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-Net: Convolutional net- works for biomedical image segmentation,” in MICCAI. Springer, 2015, pp. 234–241. ¨O. C ¸ic ¸ek, A. Abdulkadir, S. S. Lienkamp, T. Brox, and O. Ronneberger, “3d u-net: learning dense volumetric segmentation from sparse annotation,” in MICCAI. Springer, 2016, pp. 424–432. X. Li, Q. Dou, H....

    work page 2015

  3. [9]

    Transbts: Multimodal brain tumor segmentation using transformer,

    W. Wang, C. Chen, M. Ding, H. Yu, S. Zha, and J. Li, “Transbts: Multimodal brain tumor segmentation using transformer,” in JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 10 MICCAI. Springer, 2021, pp. 109–119. D. Karimi, S. D. Vasylechko, and A. Gholipour, “Convolution-free medical image segmentation using transformers,” in MICCAI. Springer, 20...

    arXiv 2021

  4. [10]

    Swin unetr: Swin transformers for semantic segmentation of brain tumors in mri images,

    A. Hatamizadeh, V . Nath, Y . Tang, D. Yang, H. R. Roth, and D. Xu, “Swin unetr: Swin transformers for semantic segmentation of brain tumors in mri images,” in MICCAI Brainlesion Workshop. Springer, 2022, pp. 272–284. Y . Jiang, Y . Zhang, X. Lin, J. Dong, T. Cheng, and J. Liang, “Swinbts: A method for 3d multimodal brain tumor segmentation using swin tra...

    work page 2022

  5. [11]

    Phtrans: Parallelly aggregating global and local representations for medical image segmentation,

    W. Liu, T. Tian, W. Xu, H. Yang, X. Pan, S. Yan, and L. Wang, “Phtrans: Parallelly aggregating global and local representations for medical image segmentation,” in MICCAI. Springer, 2022, pp. 235–244. P. Wang, X. Wang, H. Luo, J. Zhou, Z. Zhou, F. Wang, H. Li, and R. Jin, “Scaled relu matters for training vision transformers,” in AAAI, vol. 36, no. 3, 202...

    work page 2022

  6. [16]

    V-net: Fully convolutional neural networks for volumetric medical image segmentation,

    F. Milletari, N. Navab, and S.-A. Ahmadi, “V-net: Fully convolutional neural networks for volumetric medical image segmentation,” in 3DV. IEEE, 2016, pp. 565–571. K. Kobayashi et al., Mathematics of information and coding . Amer- ican Mathematical Soc., 2002, vol

    work page 2016

  7. [18]

    Rethinking semantic segmentation from a sequence-to-sequence perspective with transformers,

    S. Zheng, J. Lu, H. Zhao, X. Zhu, Z. Luo, Y . Wang, Y . Fu, J. Feng, T. Xiang, P. H. Torr et al., “Rethinking semantic segmentation from a sequence-to-sequence perspective with transformers,” in CVPR, 2021, pp. 6881–6890. A. Gu and T. Dao, “Mamba: Linear-time sequence modeling with selective state spaces,” arXiv preprint arXiv:2312.00752 ,

    arXiv 2021

  8. [22]

    Pyramid vision transformer: A versatile backbone for dense prediction without convolutions,

    W. Wang, E. Xie, X. Li, D.-P. Fan, K. Song, D. Liang, T. Lu, P. Luo, and L. Shao, “Pyramid vision transformer: A versatile backbone for dense prediction without convolutions,” in ICCV, 2021, pp. 568–578. A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez et al., “Attention is all you need,” in NeurIPS, 2017, pp. 5998–6008. J. M. J. Val...

    arXiv 2021

Show all 20 references
  1. [71]

    Segment anything,

    A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.-Y . Lo et al. , “Segment anything,” in ICCV, 2023, pp. 4015–4026. Sucheng Ren is a research assistant in the department of Electronic and Computer Engineering, The Hong Kon...

    2023

  2. [203]

    Generalised dice overlap as a deep learning loss function for highly unbalanced segmentations,

    C. H. Sudre, W. Li, T. Vercauteren, S. Ourselin, and M. Jorge Cardoso, “Generalised dice overlap as a deep learning loss function for highly unbalanced segmentations,” in Deep learning in medical image analysis and multimodal learning for clinical decision support . Springer, ...

    2017 arXiv

  3. [235]

    10 041–10

    PMLR, 21–27 Jul 2024, pp. 10 041–10

    2024

  4. [357]

    Shunted self-attention via multi-scale token aggregation,

    S. Ren, D. Zhou, S. He, J. Feng, and X. Wang, “Shunted self-attention via multi-scale token aggregation,” in CVPR, 2022, pp. 10 853– 10

    2022

  5. [464]

    Sg-former: Self-guided transformer with evolving token reallocation,

    S. Ren, X. Yang, S. Liu, and X. Wang, “Sg-former: Self-guided transformer with evolving token reallocation,” in ICCV, 2023, pp. 6003–6014. C.-Y . Lee, S. Xie, P. Gallagher, Z. Zhang, and Z. Tu, “Deeply- supervised nets,” in Artificial intelligence and statistics . PMLR, 2015, ...

    2023 arXiv

  6. [862]

    Dlformer: Discrete latent transformer for video inpainting. 2022 ieee,

    J. Ren, Q. Zheng, Y . yu Zhao, X. Xu, and C. Li, “Dlformer: Discrete latent transformer for video inpainting. 2022 ieee,” in CVPR, 2022, pp. 3501–3510. J. Ren, H. Chen, T. Ye, H. Wu, and L. Zhu, “Triplane-smoothed video dehazing with clip-enhanced generalization,” Internationa...

    2022

  7. [2017]

    Lamp: Large deep nets with automated model parallelism for image segmentation,

    W. Zhu, C. Zhao, W. Li, H. Roth, Z. Xu, and D. Xu, “Lamp: Large deep nets with automated model parallelism for image segmentation,” in MICCAI. Springer, 2020, pp. 374–384. S. Kim, I. Kim, S. Lim, W. Baek, C. Kim, H. Cho, B. Yoon, and T. Kim, “Scalable neural architecture searc...

    2020

  8. [2018]

    More than encoder: Introducing transformer decoder to upsample,

    Y . Li, W. Cai, Y . Gao, C. Li, and X. Hu, “More than encoder: Introducing transformer decoder to upsample,” in BIBM. IEEE, 2022, pp. 1597–1602. X. Xiao, S. Lian, Z. Luo, and S. Li, “Weighted res-unet for high- quality retina vessel segmentation,” in ITME. IEEE, 2018, pp. 327–...

    2022 arXiv

  9. [2021]

    Swin transformer: Hierarchical vision transformer using shifted windows,

    Z. Liu, Y . Lin, Y . Cao, H. Hu, Y . Wei, Z. Zhang, S. Lin, and B. Guo, “Swin transformer: Hierarchical vision transformer using shifted windows,” in ICCV, 2021, pp. 10 012–10

    2021

  10. [2022]

    Multi- compound transformer for accurate biomedical image segmentation,

    Y . Ji, R. Zhang, H. Wang, Z. Li, L. Wu, S. Zhang, and P. Luo, “Multi- compound transformer for accurate biomedical image segmentation,” in MICCAI. Springer, 2021, pp. 326–336. S. Huang, J. Li, Y . Xiao, N. Shen, and T. Xu, “Rtnet: Relation trans- former network for diabetic r...

    2021

  11. [2023]

    Cotr: Efficiently bridging cnn and transformer for 3d medical image segmentation,

    Y . Xie, J. Zhang, C. Shen, and Y . Xia, “Cotr: Efficiently bridging cnn and transformer for 3d medical image segmentation,” in MICCAI. Springer, 2021, pp. 171–180. A. Hatamizadeh, Y . Tang, V . Nath, D. Yang, A. Myronenko, B. Land- man, H. R. Roth, and D. Xu, “Unetr: Transfor...

    2021

  12. [2024]

    Reciprocal transformations for unsupervised video object segmentation,

    S. Ren, W. Liu, Y . Liu, H. Chen, G. Han, and S. He, “Reciprocal transformations for unsupervised video object segmentation,” in CVPR, 2021, pp. 15 455–15

    2021

Pith tools

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