Pith. sign in

REVIEW 3 major objections 6 minor 43 references

Trans${^2}$-CBCT: A Dual-Transformer Framework for Sparse-View CBCT Reconstruction

T0 review · 3 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read A dual-transformer pipeline sets new accuracy marks for six-view CBCT.

desk verdict A sensible dual-transformer recipe for sparse-view CBCT with clean ablations, but the SOTA claim depends on imported baseline numbers, one of which is non-monotonic and likely a protocol artifact. read the letter →

arxiv 2506.17425 v1 pith:TWT4LM6W submitted 2025-06-20 eess.IV cs.AIcs.CV

Minmin Yang , Huantao Ren , Senem Velipasalar This is my paper · ORCID
classification eess.IVcs.AIcs.CV
keywords cone-beamCTsparse-viewreconstructionTransUNetPointTransformerneighbor-awareattentionimplicitneuralrepresentationlow-doseimagingdigitallyreconstructedradiographs
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 argues that the right architecture choice can overcome the severe information loss of reconstructing 3D cone-beam CT volumes from as few as six X-ray projections. It replaces the UNet or ResNet encoders used by prior learning-based methods with TransUNet, a hybrid CNN-transformer, and then refines the resulting 3D features with a neighbor-aware Point Transformer. On the LUNA16 and ToothFairy benchmarks, the authors report that the full model exceeds the strongest previous baseline by 1.8 dB PSNR and 0.028 SSIM at six views. If the numbers hold under identical evaluation, the work would show that combining multi-scale global image features with explicit 3D positional reasoning is a practical recipe for low-dose CBCT.

What carries the argument

The load-bearing object is the two-stage dual-transformer combination. The first stage is TransUNet, a U-shaped CNN-transformer used as a shared encoder; its four decoder feature maps are concatenated after per-view bilinear sampling and max-pooling across views, giving a 464-channel feature per 3D point. The second stage is a neighbor-aware Point Transformer with two layers, k=3 nearest neighbors, learnable positional encoding added to point features, and multi-head attention whose scores combine scaled dot-product similarity with log(w_ij), where w_ij is a Gaussian decay in Euclidean distance. This mechanism is what enforces local smoothness and spatial coherence that plain per-point MLP prediction lacks.

What would settle it

Run C2RV, DIF-Gaussian, and Trans2-CBCT on the same LUNA16 split with identical DRR generation, the same six view angles, the same sampled points, and the same PSNR and SSIM calculation script. If the PSNR difference between Trans2-CBCT and C2RV falls below the reported 1.8 dB, or if an independent implementation of C2RV outperforms the quoted number, the central superiority claim would be falsified.

Watch

Extended reading notes

Core claim

The central claim is that sparse-view CBCT reconstruction is best modeled as a continuous attenuation field whose per-point prediction uses two complementary transformers. First, TransUNet encodes each projection at multiple scales, and view-specific features are sampled at the projected location of each 3D point and fused across views by max-pooling. Second, a Point Transformer adds learnable 3D positional encodings and a Neighbor-Aware Attention that restricts attention to k nearest neighbors, weighting scores by a Gaussian function of Euclidean distance. The paper reports that Trans-CBCT alone already beats all baselines, and Trans2-CBCT adds further gains, achieving 31.03 dB PSNR and 0.9027 SSIM on LUNA16 at six views versus 29.23 dB and 0.8747 for the strongest baseline, C2RV.

Load-bearing premise

The load-bearing assumption is that the baseline numbers the paper quotes from DIF-Gaussian were produced under exactly the same data splits, preprocessing, view angles, and metric computation as the authors' own runs; if the protocols differ, the reported margins may reflect evaluation mismatch rather than architecture.

Editorial extensions

If this is right

  • Replacing UNet or ResNet encoders with TransUNet yields measurable gains even before 3D refinement: +1.17 dB PSNR over the best baseline at six views on LUNA16.
  • The neighbor-aware Point Transformer contributes an additional 0.63 dB PSNR and 0.0117 SSIM on the same setting, showing that explicit 3D spatial reasoning adds value beyond better 2D features.
  • The improvements persist from six to ten views on both a chest CT dataset and a dental CBCT dataset, indicating the recipe transfers across anatomies and imaging geometries.
  • Using k=3 nearest neighbors beats larger neighborhoods, which means overly broad aggregation over-smooths fine anatomical boundaries and lowers both PSNR and SSIM.
  • Trans-CBCT keeps inference time near prior INR-based methods despite having over three times the parameters, while Trans2-CBCT's KNN search currently costs 53.7 seconds per case, identifying a concrete efficiency target.

Reading between the lines

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

  • If the dual-transformer design is genuinely responsible for the margin, similar gains might appear in other inverse problems with sparse angular sampling, such as fan-beam sparse-view CT or limited-angle digital breast tomosynthesis, where global context and local smoothness are both critical.
  • The reported sensitivity to k suggests a testable extension: an adaptive or learned neighbor count, or a re-weighted distance kernel, could push accuracy further or shift the optimal k as voxel size or anatomy changes.
  • Because the paper quotes most baseline numbers from the DIF-Gaussian work rather than re-running them, the cleanest confirmation is an independent re-run under identical data splits, DRR settings, and metrics; until that happens, the architectural conclusion should be treated as conditional on protocol matching.
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

3 major / 6 minor

Summary. The paper proposes two transformer-based models for sparse-view cone-beam CT (CBCT) reconstruction from 6-10 projection views. Trans-CBCT replaces the conventional UNet/ResNet image encoder used in prior multi-view feature querying methods with TransUNet, concatenates multi-scale decoder features, queries them per 3D point via bilinear interpolation, fuses views by max pooling, and regresses attenuation with a small prediction head. Trans2-CBCT adds a Point Transformer with learnable 3D positional encoding and a KNN-restricted multi-head attention whose logits include a Gaussian distance bias. Both models are trained with point-wise MSE on LUNA16 and ToothFairy. Results in Table I report consistent PSNR/SSIM superiority over FDK, SART, NAF, NeRP, FBPConvNet, FreeSeed, BBDM, PixelNeRF, DIF-Net, DIF-Gaussian, and C2RV; ablations in Tables III-V analyze the number of sampled points, KNN size, and multi-scale feature aggregation.

Significance. If the comparative numbers are trustworthy, the paper makes a useful empirical contribution: it demonstrates that a CNN-transformer hybrid encoder with point-based geometric refinement yields large gains in extremely sparse-view CBCT, e.g., 1.8 dB PSNR over C2RV at 6 LUNA16 views. The internal ablations and the consistency of the Trans2-CBCT gains over Trans-CBCT support the design choices, and the efficiency analysis is helpful. However, the significance is capped by reproducibility and comparability issues: no code or weights are released, the headline claim relies on imported baseline numbers, and no uncertainty estimates are provided.

major comments (3)
  1. [Section IV-A and Table I] The central claim that Trans-CBCT and Trans2-CBCT 'consistently outperform all prior methods' (Section V) rests on baseline entries that were not re-run under the authors' protocol. The text states that only DIF-Gaussian* was reproduced and 'the remaining results are taken from DIF-Gaussian [26]'. If the imported rows were produced with different data splits, DRR generation, view angles, or metric computation, the reported margins (e.g., +1.8 dB over C2RV on LUNA16 at 6 views) could be protocol artifacts. A concrete red flag is the DIF-Gaussian ToothFairy row: PSNR/SSIM go from 29.83/88.67 at 8 views to 29.07/85.17 at 10 views, a sharp degradation with more views that is not seen in any other method and that suggests the imported numbers come from inconsistent experimental conditions. Please rerun all baselines under the same pipeline, or explicitly restrict the superiority claim to the conditions that were actually compared, and reconcile or remove the non-monotonic row. C2RV is also missing entirely from ToothFairy, so the 'both datasets' comparison is incomplete.
  2. [Section IV-B and Table I] The reported PSNR/SSIM values are single numbers without standard deviations, confidence intervals, or significance tests. Since the headline is a comparative superiority claim spanning 18 cells (two datasets x three view counts x two metrics), the lack of any uncertainty estimate makes it impossible to judge whether the margins are stable or driven by one seed. At minimum, report mean and standard deviation over multiple training runs or a paired significance test over test volumes for the key comparison (Trans2-CBCT vs C2RV on LUNA16 and vs DIF-Gaussian on ToothFairy).
  3. [Section III-D and Section IV-B] The evaluation protocol for full volumes is not described. Training samples N'=10,000 points from the ground-truth volume (Section III-D), but Table I is said to evaluate at a 'fixed resolution of 256^3'. The paper never states whether inference queries all 16.7M voxels, whether a subset of points is sampled for metric computation, or how the KNN neighborhood is defined at full resolution. This matters because PSNR/SSIM depend on the exact set of evaluated points, and the reported inference time of 53.7 s in Table II is not interpretable without this information. Please specify the full inference procedure and metric evaluation grid.
minor comments (6)
  1. [Abstract and Section V] The phrase 'in term of' should be 'in terms of' in both the abstract and the conclusion.
  2. [Section III-C, Eq. (5)] The dimensions of W1, W2, b1, and b2 in the positional encoding are not specified; please state them explicitly so that PE(p_i) can be added to the 464-dimensional feature vector F^p.
  3. [Section III-C, Eq. (6)] Please clarify whether sigma is learned or fixed and, if fixed, its value. The role of the log(w_ij) bias in the attention score also deserves a brief explanation, since w_ij can be very close to zero for distant neighbors.
  4. [Section IV-A] The text says projection images are 'randomly sample[d] within the range of [0°, 180°)', but it is not stated whether the same set of projection angles is used for all methods and for all test volumes; this should be specified because random per-volume angles can affect fair comparison.
  5. [Section IV-C and Fig. 4] The text says 'the forth column of Fig. 4'; this should be 'the fourth column'.
  6. [Reproducibility] No code or pretrained weights are made available; releasing them, or at least a detailed configuration file, would substantially strengthen the reproducibility of the reported comparisons.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: empirical architecture comparison; imported baselines raise comparability risk but not circularity.

full rationale

The paper contains no analytic derivation whose output reduces to its input. The reconstruction function f is defined as mapping projection images and 3D points to attenuation coefficients, then trained with MSE loss against ground-truth attenuation values sampled from CT volumes; all reported PSNR and SSIM improvements are measured on held-out test data after training. The only fitted quantities are network weights and standard hyperparameters, and these are not relabeled as predictions of an analytic result. The TransUNet encoder and neighbor-aware Point Transformer are adopted from prior work and evaluated empirically on separate train/test splits, so their contribution is not definitionally forced. The baseline table imports several numbers from DIF-Gaussian [26]; this is a comparability and reproducibility concern, not circularity, because imported numbers are not constructed from the current model's inputs and no equation reduces the claimed margins to a fitted parameter. Similarly, the non-monotonic DIF-Gaussian entry in Table I may indicate protocol mismatch, but that is an experimental-validity issue outside the scope of circularity analysis. There is no self-citation load-bearing chain, no imported uniqueness theorem, and no ansatz smuggled in via citation: the architecture choices are presented as empirical design hypotheses and tested against independent data. Therefore the derivation chain is self-contained and the circularity score is 0.

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

The central claim rests on comparability of external baseline numbers, on the projection/feature query geometry, on an unspecified full-volume inference procedure, and on the MSE training objective. There are no invented physical entities. Hyperparameters k, N', l, and sigma are selected or unspecified and affect the reported numbers.

free parameters (4)
  • Neighborhood decay sigma (Eq. 6) = Not reported
    Controls the Gaussian distance weighting in neighbor-aware attention; no value, learning rule, or tuning procedure given, yet it directly shapes attention over k-nearest neighbors.
  • Number of nearest neighbors k = 3
    Chosen via ablations on both test datasets; larger k reduces PSNR/SSIM, smaller k claimed to avoid over-smoothing.
  • Sampled points per iteration N' = 10,000
    Selected by ablation; 5,000 underperforms and 20,000 slightly degrades, so 10,000 is the reported optimum on the evaluation sets.
  • Point Transformer depth l = 2
    Set as a fixed trade-off between efficiency and accuracy; no ablation for l is shown.
assumptions (4)
  • ad hoc to paper Imported baseline numbers from DIF-Gaussian [26] are directly comparable to the authors' runs under identical protocol.
    Section IV-A states the remaining results are taken from DIF-Gaussian; if the underlying splits, view angles, or metric code differ, the reported superiority margins may be artifacts.
  • domain assumption Bilinear interpolation of 2D feature maps at projected coordinates faithfully captures the CBCT geometry for each 3D point.
    Eq. 2 is the only link between 3D points and image features; errors in the projection function or interpolation would directly corrupt the attenuation prediction.
  • ad hoc to paper Full 256^3 evaluation is implied to be achievable by point-wise evaluation, although the inference sampling procedure is never described.
    Training samples 10,000 points per iteration (Section III-C), and the paper does not specify how the complete volume or PSNR/SSIM are computed at test time.
  • domain assumption MSE on trilinearly interpolated point labels is a sufficient objective for full-volume reconstruction quality.
    Eq. 10 optimizes per-point MSE; transfer to full-volume PSNR/SSIM is assumed without analysis.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Trans${^2}$-CBCT: A Dual-Transformer Framework for Sparse-View CBCT Reconstruction." pith.science (2026). https://pith.science/paper/TWT4LM6W

@misc{pith2026250617425,
  author       = {Pith},
  title        = {Pith review of: Trans$^2$-CBCT: A Dual-Transformer Framework for Sparse-View CBCT Reconstruction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TWT4LM6W}},
  note         = {Machine review of arXiv:2506.17425}
}
abstract

Cone-beam computed tomography (CBCT) using only a few X-ray projection views enables faster scans with lower radiation dose, but the resulting severe under-sampling causes strong artifacts and poor spatial coverage. We address these challenges in a unified framework. First, we replace conventional UNet/ResNet encoders with TransUNet, a hybrid CNN-Transformer model. Convolutional layers capture local details, while self-attention layers enhance global context. We adapt TransUNet to CBCT by combining multi-scale features, querying view-specific features per 3D point, and adding a lightweight attenuation-prediction head. This yields Trans-CBCT, which surpasses prior baselines by 1.17 dB PSNR and 0.0163 SSIM on the LUNA16 dataset with six views. Second, we introduce a neighbor-aware Point Transformer to enforce volumetric coherence. This module uses 3D positional encoding and attention over k-nearest neighbors to improve spatial consistency. The resulting model, Trans$^2$-CBCT, provides an additional gain of 0.63 dB PSNR and 0.0117 SSIM. Experiments on LUNA16 and ToothFairy show consistent gains from six to ten views, validating the effectiveness of combining CNN-Transformer features with point-based geometry reasoning for sparse-view CBCT reconstruction.

Figures

Figures reproduced from arXiv: 2506.17425 by the authors.

Figure 1
Figure 1. The architecture of TransUNet. It includes a CNN-Transformer hybrid [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. The architecture of Trans-CBCT and Trans [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. The architecture of neighbor-aware Point Transformer layer. (a) [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Visualization of 6-view reconstruction results for different methods. Top three rows: axial, coronal, and sagittal slices from the LUNA16 dataset. [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 28 canonical work pages

  1. [26]

    Learning 3d gaussians for extremely sparse-view cone-beam ct reconstruction,

    Y . Lin, H. Wang, J. Chen, and X. Li, “Learning 3d gaussians for extremely sparse-view cone-beam ct reconstruction,” in International Conference on Medical Image Computing and Computer-Assisted Inter- vention. Springer, 2024, pp. 425–435

    work page 2024

  2. [1]

    Algebraic reconstruction techniques (art) for three-dimensional electron microscopy and x-ray photography,

    R. Gordon, R. Bender, and G. T. Herman, “Algebraic reconstruction techniques (art) for three-dimensional electron microscopy and x-ray photography,” Journal of theoretical Biology , vol. 29, no. 3, pp. 471– 481, 1970

    work page 1970

  3. [2]

    Simultaneous algebraic reconstruction tech- nique (sart): A superior implementation of the art algorithm,

    A. Andersen and A. Kak, “Simultaneous algebraic reconstruction tech- nique (sart): A superior implementation of the art algorithm,” Ultrasonic Imaging, vol. 6, no. 1, pp. 81–94, 1984

    work page 1984

  4. [3]

    Convergence studies on iterative algorithms for image reconstruction,

    M. Jiang and G. Wang, “Convergence studies on iterative algorithms for image reconstruction,” IEEE Transactions on Medical Imaging , vol. 22, no. 5, pp. 569–579, 2003

    work page 2003

  5. [4]

    Nonlinear total variation based noise removal algorithms,

    L. I. Rudin, S. Osher, and E. Fatemi, “Nonlinear total variation based noise removal algorithms,” Physica D: Nonlinear Phenomena , vol. 60, no. 1, pp. 259–268, 1992. [Online]. Available: https: //www.sciencedirect.com/science/article/pii/016727899290242F

    arXiv 1992

  6. [5]

    Image reconstruction in circular cone-beam computed tomography by constrained, total-variation minimization,

    E. Y . Sidky and X. Pan, “Image reconstruction in circular cone-beam computed tomography by constrained, total-variation minimization,” Physics in Medicine & Biology , vol. 53, no. 17, p. 4777, 2008

    2008

  7. [6]

    Learning deep intensity field for extremely sparse-view cbct reconstruction,

    Y . Lin, Z. Luo, W. Zhao, and X. Li, “Learning deep intensity field for extremely sparse-view cbct reconstruction,” in International Conference on Medical Image Computing and Computer-Assisted Intervention . Springer, 2023, pp. 13–23

    work page 2023

  8. [7]

    Cˆ2rv: Cross- regional and cross-view learning for sparse-view cbct reconstruction,

    Y . Lin, J. Yang, H. Wang, X. Ding, W. Zhao, and X. Li, “Cˆ2rv: Cross- regional and cross-view learning for sparse-view cbct reconstruction,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2024, pp. 11 205–11 214

    work page 2024

Show all 43 references
  1. [8]

    Naf: neural attenuation fields for sparse- view cbct reconstruction,

    R. Zha, Y . Zhang, and H. Li, “Naf: neural attenuation fields for sparse- view cbct reconstruction,” inInternational Conference on Medical Image Computing and Computer-Assisted Intervention . Springer, 2022, pp. 442–452

    2022

  2. [9]

    R 2-gaussian: Rectifying radiative gaussian splatting for tomographic reconstruction,

    R. Zha, T. J. Lin, Y . Cai, J. Cao, Y . Zhang, and H. Li, “R 2-gaussian: Rectifying radiative gaussian splatting for tomographic reconstruction,” in Advances in Neural Information Processing Systems (NeurIPS), 2024

    2024

  3. [10]

    Intensity field decomposition for tissue-guided neural tomography,

    M.-X. Li, J.-G. Yu, Y . Gao, C. Huang, and G.-S. Xia, “Intensity field decomposition for tissue-guided neural tomography,” arXiv preprint arXiv:2411.00900, 2024

    2024 arXiv

  4. [11]

    Snaf: Sparse-view cbct reconstruction with neural attenuation fields,

    Y . Fang, L. Mei, C. Li, Y . Liu, W. Wang, Z. Cui, and D. Shen, “Snaf: Sparse-view cbct reconstruction with neural attenuation fields,” arXiv preprint arXiv:2211.17048, 2022

    2022 arXiv

  5. [12]

    Geometry- aware attenuation learning for sparse-view cbct reconstruction,

    Z. Liu, Y . Fang, C. Li, H. Wu, Y . Liu, D. Shen, and Z. Cui, “Geometry- aware attenuation learning for sparse-view cbct reconstruction,” IEEE Transactions on Medical Imaging , 2024

    2024

  6. [13]

    U-net: Convolutional networks for biomedical image segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-net: Convolutional networks for biomedical image segmentation,” in Medical image computing and computer-assisted intervention–MICCAI 2015: 18th international con- ference, Munich, Germany, October 5-9, 2015, proceedings, part III 18 ...

    2015

  7. [14]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2016, pp. 770–778

    2016

  8. [15]

    Transunet: Transformers make strong encoders for medical image segmentation,

    J. Chen, Y . Lu, Q. Yu, X. Luo, E. Adeli, Y . Wang, L. Lu, A. L. Yuille, and Y . Zhou, “Transunet: Transformers make strong encoders for medical image segmentation,” arXiv preprint arXiv:2102.04306 , 2021

    2021 arXiv

  9. [16]

    Validation, comparison, and combination of algorithms for automatic detection of pulmonary nodules in computed tomography images: the luna16 challenge,

    A. A. A. Setio, A. Traverso, T. De Bel, M. S. Berens, C. Van Den Bo- gaard, P. Cerello, H. Chen, Q. Dou, M. E. Fantacci, B. Geurts et al. , “Validation, comparison, and combination of algorithms for automatic detection of pulmonary nodules in computed tomography images: the lu...

    2017

  10. [17]

    Deep segmentation of the mandibular canal: a new 3d annotated dataset of cbct volumes,

    M. Cipriano, S. Allegretti, F. Bolelli, M. Di Bartolomeo, F. Pollastri, A. Pellacani, P. Minafra, A. Anesi, and C. Grana, “Deep segmentation of the mandibular canal: a new 3d annotated dataset of cbct volumes,” Ieee Access, vol. 10, pp. 11 500–11 510, 2022

    2022

  11. [18]

    Practical cone-beam algorithm,

    L. A. Feldkamp, L. C. Davis, and J. W. Kress, “Practical cone-beam algorithm,” J. Opt. Soc. Am. A , vol. 1, no. 6, pp. 612–619, Jun 1984

    1984

  12. [19]

    A tutorial on art (algebraic reconstruction techniques),

    R. Gordon, “A tutorial on art (algebraic reconstruction techniques),” IEEE Transactions on Nuclear Science , vol. 21, no. 3, pp. 78–93, 1974

    1974

  13. [20]

    Sparse-view cone-beam computed tomography iterative reconstruction based on new multi-gradient direction total variation,

    J. Cui, Y . Hou, Z. Jiang, G. Yu, L. Ye, Q. Cao, and Q. Sun, “Sparse-view cone-beam computed tomography iterative reconstruction based on new multi-gradient direction total variation,” Journal of Cancer Research and Therapeutics, vol. 20, no. 2, pp. 615–624, 2024

    2024

  14. [21]

    Sparse-view cbct reconstruction via weighted schatten p-norm minimization,

    C. Xu, B. Yang, F. Guo, W. Zheng, and P. Poignet, “Sparse-view cbct reconstruction via weighted schatten p-norm minimization,” Opt. Express, vol. 28, no. 24, pp. 35 469–35 482, Nov 2020. [Online]. Available: https://opg.optica.org/oe/abstract.cfm?URI=oe-28-24-35469

    2020

  15. [22]

    Deep back projection for sparse-view ct reconstruction,

    D. H. Ye, G. T. Buzzard, M. Ruby, and C. A. Bouman, “Deep back projection for sparse-view ct reconstruction,” in 2018 IEEE Global Conference on Signal and Information Processing (GlobalSIP) . IEEE, 2018, pp. 1–5

    2018

  16. [23]

    Deep con- volutional neural network for inverse problems in imaging,

    K. H. Jin, M. T. McCann, E. Froustey, and M. Unser, “Deep con- volutional neural network for inverse problems in imaging,” IEEE Transactions on Image Processing, vol. 26, no. 9, pp. 4509–4522, 2017

    2017

  17. [24]

    pixelnerf: Neural radiance fields from one or few images,

    A. Yu, V . Ye, M. Tancik, and A. Kanazawa, “pixelnerf: Neural radiance fields from one or few images,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021, pp. 4578– 4587

    2021

  18. [25]

    Nerp: implicit neural representation learning with prior embedding for sparsely sampled image reconstruc- 10 tion,

    L. Shen, J. Pauly, and L. Xing, “Nerp: implicit neural representation learning with prior embedding for sparsely sampled image reconstruc- 10 tion,” IEEE Transactions on Neural Networks and Learning Systems , vol. 35, no. 1, pp. 770–782, 2022

    2022

  19. [27]

    An image is worth 16x16 words: Transformers for image recognition at scale,

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

    2010 arXiv

  20. [28]

    Tl-med: A two-stage transfer learning recognition model for medical images of covid-19,

    J. Meng, Z. Tan, Y . Yu, P. Wang, and S. Liu, “Tl-med: A two-stage transfer learning recognition model for medical images of covid-19,” Biocybernetics and Biomedical Engineering , vol. 42, no. 3, pp. 842– 855, 2022. [Online]. Available: https://www.sciencedirect.com/science/ a...

    2022

  21. [29]

    Imagenet: A large-scale hierarchical image database,

    J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei, “Imagenet: A large-scale hierarchical image database,” in 2009 IEEE Conference on Computer Vision and Pattern Recognition , 2009, pp. 248–255

    2009

  22. [30]

    Transmed: Transformers advance multi- modal medical image classification,

    Y . Dai, Y . Gao, and F. Liu, “Transmed: Transformers advance multi- modal medical image classification,” Diagnostics, vol. 11, no. 8, p. 1384, 2021

    2021

  23. [31]

    Vision transformers for classification of breast ultrasound images,

    B. Gheflati and H. Rivaz, “Vision transformers for classification of breast ultrasound images,” in 2022 44th annual international conference of the IEEE Engineering in Medicine & Biology Society (EMBC) . IEEE, 2022, pp. 480–483

    2022

  24. [32]

    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 International MICCAI brainlesion workshop . Springer, 2021, pp. 272–284

    2021

  25. [33]

    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 Proceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 10 012–10 022

    2021

  26. [34]

    A robust volumetric transformer for accurate 3d tumor segmentation,

    H. Peiris, M. Hayat, Z. Chen, G. Egan, and M. Harandi, “A robust volumetric transformer for accurate 3d tumor segmentation,” in Interna- tional conference on medical image computing and computer-assisted intervention. Springer, 2022, pp. 162–172

    2022

  27. [35]

    D- former: A u-shaped dilated transformer for 3d medical image segmen- tation,

    Y . Wu, K. Liao, J. Chen, J. Wang, D. Z. Chen, H. Gao, and J. Wu, “D- former: A u-shaped dilated transformer for 3d medical image segmen- tation,” Neural Computing and Applications , vol. 35, no. 2, pp. 1931– 1944, 2023

    1931

  28. [36]

    Mbst-driven 4d-cbct reconstruction: Leveraging swin transformer and masking for robust performance,

    N. Cao, Q. Li, K. Sun, H. Zhang, J. Ding, Z. Wang, W. Chen, L. Gao, J. Sun, K. Xie, and X. Ni, “Mbst-driven 4d-cbct reconstruction: Leveraging swin transformer and masking for robust performance,” Computer Methods and Programs in Biomedicine , vol. 262, p. 108637, 2025. [Onlin...

    2025

  29. [37]

    Wutrans: Whole-spectrum unilateral-query-secured transformer for 4d cbct reconstruction,

    P. Yuan, T. Lyu, F. Lyu, Y . Zhang, C. Yang, W. Zhu, Z. Gao, Z. Wu, Y . Chen, W. Zhao, and J. L. Coatrieux, “Wutrans: Whole-spectrum unilateral-query-secured transformer for 4d cbct reconstruction,” Biomedical Signal Processing and Control , vol. 102, p. 107197, 2025. [Online]...

    2025

  30. [38]

    Cbctformer: Transformer model for sparse-view cbct artifact reduction in image- guided radiotherapy,

    X. Zhao, K. Chen, Y . Du, J. Li, L. Sui, and Y . Peng, “Cbctformer: Transformer model for sparse-view cbct artifact reduction in image- guided radiotherapy,” in 2024 17th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP- BMEI)...

    2024

  31. [39]

    Removing ring artifacts in cbct images via transformer with unidirectional vertical gradient loss,

    J. Sha and J. Li, “Removing ring artifacts in cbct images via transformer with unidirectional vertical gradient loss,” Medical Physics , vol. 51, no. 9, pp. 6149–6160, 2024

    2024

  32. [40]

    A transformer-based dual-domain network for reconstructing fov extended cone-beam ct images from truncated sinograms in radiation therapy,

    L. Gao, K. Xie, J. Sun, T. Lin, J. Sui, G. Yang, and X. Ni, “A transformer-based dual-domain network for reconstructing fov extended cone-beam ct images from truncated sinograms in radiation therapy,” Computer Methods and Programs in Biomedicine , vol. 241, p. 107767, 2023

    2023

  33. [41]

    Freeseed: Frequency- band-aware and self-guided network for sparse-view ct reconstruc- tion,

    C. Ma, Z. Li, J. Zhang, Y . Zhang, and H. Shan, “Freeseed: Frequency- band-aware and self-guided network for sparse-view ct reconstruc- tion,” in International Conference on Medical Image Computing and Computer-Assisted Intervention. Springer, 2023, pp. 250–259

    2023

  34. [42]

    Bbdm: Image-to-image trans- lation with brownian bridge diffusion models,

    B. Li, K. Xue, B. Liu, and Y .-K. Lai, “Bbdm: Image-to-image trans- lation with brownian bridge diffusion models,” in Proceedings of the IEEE/CVF conference on computer vision and pattern Recognition , 2023, pp. 1952–1961

    2023

  35. [43]

    Decoupled weight decay regularization,

    I. Loshchilov and F. Hutter, “Decoupled weight decay regularization,” arXiv preprint arXiv:1711.05101 , 2017

    2017 arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.