Pith. sign in

REVIEW 4 major objections 5 minor 68 references

Revisiting Pose Sensitivity in Splat-based Computed Tomography under Sparse-view Reconstruction

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

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

Pith's one-line read Pose inaccuracies, not view sparsity, cause the streak artifacts in splat-based CT under sparse views, and joint geometric refinement during reconstruction removes them.

desk verdict Solid self-calibrating splat CT with a genuinely useful diagnostic experiment; the pose-attribution claim is slightly stronger than the evidence, but the method works and deserves a serious referee. read the letter →

arxiv 2608.04752 v1 pith:57N3YWZJ submitted 2026-08-05 cs.CV cs.GR

Kiseok Choi , Hyeongjun Cho , Inchul Kim , Min H. Kim This is my paper · ORCID
classification cs.CVcs.GR
keywords cone-beamCTsparse-viewreconstructionGaussiansplattinggeometriccalibrationposerefinementself-calibratingtomographyartifactanalysis
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

Splat-based CT, which reconstructs a volume as a cloud of anisotropic 3D Gaussians, produces pronounced streak and strip artifacts when applied to real sparse-view X-ray data. This paper argues that the artifacts are caused not by having too few projection views, but by small inaccuracies in the camera poses of the acquisition geometry. The argument is supported by a controlled comparison: when 75 projections are re-synthesized from a dense 721-view FDK reconstruction, the streaks vanish, while the same number of real projections still show them. The paper then derives a stable, differentiable refinement of pose parameters that is optimized jointly with the Gaussian volume, showing that this self-calibrating pipeline suppresses the artifacts and improves reconstruction fidelity under synthetic and real sparse-view conditions.

What carries the argument

The central object is the splat-based forward-projection operator: the attenuation volume is a continuous sum of 3D Gaussians, and each ray's intensity is computed analytically by integrating the projected Gaussian density along the ray. The paper's mechanism for stable self-calibration is the explicit gradient flow through the pose parameters: each camera pose is parameterized as a small incremental quaternion and translation relative to its initial value, and the backward pass is tracked through both the perspective projection matrix $\mathbf{P}_k$ and the rotation matrix $\mathbf{W}_k$ via the intermediate matrix $\mathbf{M}_{i,k} = \mathbf{J}_{i,k} \mathbf{W}_k$. The key correction is the term $\partial \mathcal{L}/\partial \mathbf{W}_{k,b}$, which previous splatting-based calibration omitted. This formulation adds only seven pose parameters per view and allows the volume and geometry to be optimized together with a simple L1 plus SSIM loss, making total-variation regularization unnecessary.

What would settle it

Reconstruct a known calibration phantom with a real cone-beam system, obtain a high-precision pose estimate from independent phantom-based calibration, then deliberately add known pose perturbations. If splat-based reconstruction from 75 views with deliberately wrong poses reproduces the needle and streak artifacts while the same views with correct poses do not, the causal role of pose errors is directly confirmed; if artifacts persist when a high-precision independent calibration is used, the paper's attribution would be falsified.

Watch

Extended reading notes

Core claim

The central discovery is that splat-based CT is far more sensitive to acquisition-geometry errors than to view sparsity alone. Under pose perturbations typical of real rotating systems, the anisotropic Gaussian representation propagates small rotations and translations into large errors in the rendered attenuation, producing needle-like streaks that are far more pronounced than in conventional reconstruction methods. The paper identifies the missing piece in prior splat-based CT: the gradient through the rotation matrix of the camera pose was neglected, and once the full Jacobian flow through both the projection matrix and the rotation matrix is tracked, joint pose-and-volume optimization becomes stable and accurate. The resulting method recovers camera poses to about 0.6 degrees and 0.7 voxel of translation on average, removes the artifacts, and does so without the total-variation regularization that prior splat-based methods needed, which had over-smoothed fine structure.

Load-bearing premise

The artifact-attribution experiment assumes that FDK reconstruction from 721 real projections is accurate enough to serve as pseudo ground truth, because all synthetic comparisons are generated from that volume.

Editorial extensions

If this is right

  • Real-world sparse-view CT with splat-based reconstruction can become artifact-free by correcting camera poses, instead of relying on TV smoothing that blurs fine detail.
  • The estimated pose corrections are reusable: they can be exported and applied to other tomographic reconstruction methods, not just the Gaussian splatting pipeline.
  • Self-calibration removes the need for offline phantom-based geometric calibration, allowing reconstruction on systems with mechanical imperfections or drift during scanning.
  • The method remains effective as view count drops from 75 to 25 views, extending the usable sparsity range of splat-based CT before regularization is needed.
  • Because pose error, not sparsity, is the dominant failure source, system-design efforts can focus on the geometric stability of the rotation stage rather than on merely acquiring more views.

Reading between the lines

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

  • If pose error dominates sparse-view artifacts in splat-based CT, the same joint-calibration principle should transfer to other explicit differentiable volume representations such as voxel grids or triplane features that currently rely on regularization to hide misalignment.
  • The paper's artifact-attribution diagnostic—comparing reconstructions from real projections against resimulated projections of a pseudo ground truth and looking for a directional bias in the projection-error map—could be adopted as a general calibration-quality test in other CT systems.
  • A natural extension is an analytical model of the pose-error tolerance of a splat volume as a function of Gaussian covariance and ray length, letting practitioners know the maximum mechanical wobble that a scan can tolerate before streaks appear.
  • If confirmed, the results suggest that in clinical and industrial CT, investing in stage mechanical stability may yield larger image-quality gains than adding more views.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper studies splat-based (Gaussian) cone-beam CT reconstruction under sparse-view conditions. The authors observe that real CT scans produce strong streak/strip artifacts that are absent in synthetic sparse-view reconstructions, and they argue that these artifacts are primarily caused by camera pose inaccuracies rather than by view sparsity. They support this with an analysis experiment (Section 3) that compares reconstructions from real 75-view projections against reconstructions from 75 synthetic projections derived from an FDK-based pseudo ground truth, and with a supplementary reproducibility experiment. The method contribution is a self-calibrating splat-based pipeline that jointly optimizes Gaussian volume parameters and per-view camera poses via a stable gradient formulation, without TV regularization. Experiments on 15 synthetic scenes with injected pose noise show consistent PSNR/SSIM improvements over prior joint calibration and reconstruction methods, and qualitative results on a real dataset show artifact reduction. The central claim is that pose error, not sparsity, is the dominant artifact source, and that the proposed joint refinement substantially improves reconstruction fidelity.

Significance. If the causal claim is correct, the paper identifies a practical bottleneck for splat-based CT and offers a lightweight, differentiable self-calibration method that can be integrated into existing pipelines. The synthetic evaluation is broad: 15 scenes, multiple pose-noise levels, view-count ablations at 75/50/25 views, pose-estimation RMSE tables, and a controlled perturbation-generation scheme with an explicit Lie-algebra derivation. The method also demonstrates a clear quantitative advantage on synthetic data over strong baselines (NeAT, Thies et al.). The paper is explicit about its limitation in extreme sparse-view settings and the trade-off with TV regularization, which adds credibility. However, the evidence for the primary causal attribution is weaker than the language of the abstract suggests: the real-data artifact analysis relies on an FDK pseudo ground truth and a reproduction experiment whose estimation procedure is not fully specified. The central contribution is defensible, but the load-bearing artifact-attribution evidence needs strengthening.

major comments (4)
  1. [Section 3; Supplement Section 7] The FDK pseudo-ground-truth assumption is load-bearing for the claim that artifacts are due to pose errors rather than view sparsity. The support offered in Supplement Section 2, Eqs. (5)-(10), assumes independent zero-mean Gaussian per-view pose perturbations and an approximately constant Jacobian across views. Real CT calibration errors are often systematic (e.g., detector tilt, source offset, stage wobble), in which case the FDK average over 721 views does not remove the bias, and the pseudo-GT inherits the geometry error. With such systematic bias, the synthetic 75-view projections are generated from a biased volume, so the reduced artifacts in Figure 3(b) do not cleanly demonstrate the absence of sparsity-induced artifacts. The paper should either justify that the real dataset's geometry errors are independent across views, or validate the pseudo-GT against an independent high-quality reference (e.g., a phantom scan with known geometry).
  2. [Supplement Section 7] The reproduction experiment that establishes a 'direct causal link' between pose errors and needle-like artifacts is incompletely specified. The text states that pose errors are 'estimated from real sparse-view data' and applied to generate simulated projections from the ground-truth volume, but it does not state how these pose errors are obtained. If the pose estimates come from the same joint reconstruction objective that produces the artifact-laden reconstruction, then the reproduction is circular: the pose estimates would already be biased by the method's assumptions. To avoid circularity, the pose errors should be measured independently (e.g., from known phantom geometry, from a separate calibration scan, or from a method that does not use the splat reconstruction objective), and the reproduction should demonstrate that the simulated artifacts match the real ones both qualitatively and quantitatively (e.g., by comparing artifact orientation, location, and magnitude).
  3. [Section 3, Figure 4] The directional edge-correlated bias in the real-data error maps is evidence of model mismatch, but it does not by itself identify pose inaccuracy as the cause. Scatter, beam hardening, detector response nonuniformity, and cone-beam approximation also produce edge-correlated directional residuals under sparse-view sampling. The paper's conclusion that the bias 'reveals the inaccuracy of camera poses' (Section 3, final sentence) is stronger than what the evidence supports. A control experiment is needed that injects independently measured pose errors into a forward model with a known ground-truth volume and checks that the resulting error maps match the real-data pattern; alternatively, a phantom with known geometry and no other physical effects (scatter, beam hardening) could be scanned to isolate pose as the sole variable.
  4. [Section 5.3, Figure 7] The real-world evaluation is qualitative only. The central claim concerns behavior under 'real-world sparse-view conditions,' yet no quantitative metric is reported on the real dataset (e.g., against a high-quality reference volume, a calibration phantom, or a consistent image-quality measure). Given that the synthetic results already show strong quantitative gains, the paper would be strengthened by a quantitative real-data evaluation, even if the reference is an FDK volume from dense projections (with the caveat discussed above) or a known phantom. This would also make the 'substantially improves reconstruction fidelity' claim in the abstract measurable rather than visual.
minor comments (5)
  1. [Section 4, Eq. (6)] The mapping function φ(·) is defined as returning a 3D vector, but the projection ψ(·) is described as R^3→R^2; please clarify the relationship between φ and ψ and the role of the third component in the projection of the Gaussian center.
  2. [Section 5.2] The statement that translation noise σ_trans is '1.0 in unit length (corresponding to the voxel size)' should be stated more precisely: is the unit length exactly one voxel of the reconstruction volume, and is the same scaling used across all scenes? The phrase 'may vary with dataset scaling' is ambiguous.
  3. [Table 3] Table 3 reports only two scenes (Beetle and Head), while the text says 'Reconstruction performance with camera noise varies.' Please either report per-scene results for all scenes or state that the table is representative and include the full results in the supplement.
  4. [Supplement Section 4] The description of the Thies et al. comparison states that the score network was not used and instead an L2 loss against an FBP reference was employed. This is a significant deviation from the original method and should be disclosed in the main text, not only in the supplement, to avoid misleading readers about the strength of the baseline.
  5. [Section 5.3, Eq. (12)] The orientation error formula θ = (1/N) sqrt(arccos(tr(R)-1)/2) is missing parentheses; as written, it is ambiguous whether the square root applies to the entire fraction. Please revise to θ = (1/N) sqrt((arccos(tr(R)-1))/2) or an equivalent clear expression.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the central pose-sensitivity claim is grounded in external synthetic experiments with ground-truth pose perturbations, and no fitted parameter or self-citation is load-bearing.

full rationale

The paper's central claim is that real sparse-view splat-based CT artifacts primarily originate from pose inaccuracies rather than view sparsity. The supporting evidence is an artifact-attribution experiment: a pseudo ground-truth volume is formed by FDK from 721 real projections, 75 synthetic projections are generated from it, and the splat reconstruction from synthetic projections lacks the streaks seen with the real 75-view data (Section 3, Figure 2-4). This comparison does not define the conclusion in terms of the method's own outputs: the synthetic branch is generated by an external classical algorithm, and the real-versus-synthetic asymmetry is an empirical observation, not a construction. The joint pose-refinement method is evaluated on synthetic data with known injected rotation and translation perturbations, and pose-estimation accuracy is scored against ground truth (Section 5.3, Table 2), so the calibration evaluation is externally grounded rather than self-referential. No load-bearing result rests on a self-citation: the cited baselines (R2-Gaussian, NeAT, Thies et al.) are all external works, and the paper does not invoke a uniqueness theorem or prior result by the same authors to force its modeling choices. The supplementary reproducibility experiment (Sec. 7) simulates pose errors estimated from real data and reproduces artifact patterns, but the paper does not state that the pose errors were estimated with the same joint objective; while under-specified, this is not definitional circularity. The FDK pseudo-GT faithfulness assumption and the possibility of systematic geometry bias are legitimate threats to the strength of the causal evidence, but they concern experimental validity and generalizability, not whether a prediction is equivalent to its inputs by construction. Therefore no circular step is exhibited, and the appropriate circularity score is 0.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central claim rests on an FDK pseudo-GT assumption, the correctness of the splat projection model, the validity of incremental pose updates, and controlled synthetic noise assumptions. Two loss hyperparameters are incompletely specified. No new physical entities, particles, or mediators are introduced.

free parameters (2)
  • SSIM loss weight lambda = not reported
    Equation (11) combines L1 and SSIM with a weight lambda, but the value is never given in the paper. The balance between the two terms affects the optimization trajectory and exact quantitative results, and no sensitivity analysis is provided.
  • camera pose learning rate schedule = 2e-4 decaying to 2e-5 over 30,000 steps
    Section 5.1 states the schedule as a hand-chosen setting. Pose refinement could be sensitive to this schedule, and no ablation over learning rates is reported.
assumptions (5)
  • domain assumption FDK reconstruction from 721 real projections is an accurate pseudo ground truth
    Section 3, Step 1, and supplement Section 2 rely on this to generate synthetic 75-view projections. The supporting RMSE scaling assumes independent zero-mean pose noise and near-constant Jacobians across views.
  • domain assumption The splat-based forward projection model in Equations (2) and (3) correctly describes cone-beam X-ray attenuation
    The pose-sensitivity analysis and gradient derivation are built on the R2-Gaussian and EWA splatting formulation. If this model is inaccurate, the derived pose gradients and the conclusions about artifact sources inherit that error.
  • domain assumption Initial camera poses are close enough for small incremental quaternion and translation updates
    Section 4, Equation (7) parameterizes pose corrections as small increments relative to the initial geometry. Large initial misalignment could break the local linearization and cause the joint optimization to fail.
  • domain assumption Synthetic rotation noise sampled in the Lie algebra is unbiased
    Supplement Section 3 shows E[R] approximately equals (1 - sigma_rot^2) times identity, which is not exactly unbiased at second order. The 'unbiased' claim holds only for small sigma_rot.
  • domain assumption Jacobians are roughly constant across views in the RMSE scaling derivation
    Supplement Equation (8) requires Ji approximately equal to J for all views to obtain RMSE proportional to 1/sqrt(N). Without this, the pseudo-GT error estimate and the FDK reliability argument weaken.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Revisiting Pose Sensitivity in Splat-based Computed Tomography under Sparse-view Reconstruction." pith.science (2026). https://pith.science/paper/57N3YWZJ

@misc{pith2026260804752,
  author       = {Pith},
  title        = {Pith review of: Revisiting Pose Sensitivity in Splat-based Computed Tomography under Sparse-view Reconstruction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/57N3YWZJ}},
  note         = {Machine review of arXiv:2608.04752}
}
read the original abstract

X-ray computed tomography (CT) reconstructs volumetric representations of objects from projection images obtained by transmitting X-rays through a target. Recent splat-based tomography, which represents a volume as a continuous distribution of 3D Gaussians, has demonstrated both high reconstruction quality and fast convergence in cone-beam sparse-view CT. However, when deployed in real CT systems with limited and non-uniform view distributions, we observe distinctive streak and strip artifacts that are far more pronounced than in conventional reconstruction methods. Through detailed analysis, we show that these artifacts primarily originate from pose inaccuracies in the acquisition geometry rather than from view sparsity itself. We revisit pose sensitivity in the splatting formulation and derive a stable gradient-based framework that jointly refines geometric parameters during reconstruction. Our study not only identifies how pose perturbations propagate through the differentiable projection operator but also reveals why splat-based CT is particularly vulnerable to geometric misalignment. The resulting formulation remains lightweight and easily integrable into existing pipelines while substantially improving reconstruction fidelity under real-world sparse-view conditions.

Figures

Figures reproduced from arXiv: 2608.04752 by the authors.

Figure 1
Figure 1. Reconstruction of a real walnut from cone-beam CT [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The pipeline of our artifact analysis experiment. (Step 1) [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Comparison between the reconstructions from real 75- [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (10 more)
Figure 4
Figure 4. Figure 4: Comparison of projection images and error maps be [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: The backward gradient computation pipeline with re [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: Comparison of reconstructed volume slices on the synthetic dataset with the pose perturbation. We compare our method against [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 7
Figure 7. Figure 7: Comparison of reconstructed volume slices on the real dataset [ [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 8
Figure 8. Figure 8: Comparison between the baseline [48] and ours from extremely sparse input projections (25 views) with geometric per￾turbations. Our method reconstructs the volume with significantly fewer artifacts even under extreme conditions. while our self-calibration framework con…
Figure 2
Figure 2. Figure 2: TV term impact under extreme sparse-view conditions. [PITH_FULL_IMAGE:figures/full_fig_p013_2.png]
Figure 1
Figure 1. Figure 1: Reproducibility of a real artifact. construction reproduces artifact patterns consistent with the real result ( [PITH_FULL_IMAGE:figures/full_fig_p013_1.png]
Figure 3
Figure 3. Figure 3: Comparison of reconstructed volume slices on the synthetic dataset with the pose perturbation. We compare our method against [PITH_FULL_IMAGE:figures/full_fig_p014_3.png]
Figure 4
Figure 4. Figure 4: Comparison of reconstructed volume slices on the synthetic dataset with the pose perturbation. We compare our method against [PITH_FULL_IMAGE:figures/full_fig_p015_4.png]
Figure 5
Figure 5. Figure 5: Comparison of reconstructed volume slices on the synthetic dataset with the pose perturbation. We compare our method against [PITH_FULL_IMAGE:figures/full_fig_p016_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

68 extracted references · 24 canonical work pages

  1. [1]

    2001 , publisher=

    Principles of computerized tomographic imaging , author=. 2001 , publisher=

    2001

  2. [2]

    2003 , publisher=

    Computed tomography: principles, design, artifacts, and recent advances , author=. 2003 , publisher=

    work page 2003

  3. [3]

    W. C. Röntgen , title =. Sitzungsberichte der Physikalisch-Medizinischen Gesellschaft zu Würzburg , volume =

    work page

  4. [4]

    Computer Graphics and Image Processing , volume=

    The reconstruction of fan-beam data by filtering the back-projection , author=. Computer Graphics and Image Processing , volume=. 1979 , publisher=

    work page 1979

  5. [5]

    Journal of the Optical Society of America A , volume=

    Practical cone-beam algorithm , author=. Journal of the Optical Society of America A , volume=. 1984 , publisher=

    1984

  6. [6]

    IEEE Transactions on Visualization and Computer Graphics , volume=

    EWA splatting , author=. IEEE Transactions on Visualization and Computer Graphics , volume=. 2002 , publisher=

    2002

  7. [7]

    Physics in Medicine & Biology , volume=

    Metal artifact reduction method based on a constrained beam-hardening estimator for polychromatic x-ray CT , author=. Physics in Medicine & Biology , volume=. 2021 , publisher=

    2021

  8. [8]

    International Conference on Medical Image Computing and Computer-Assisted Intervention , pages=

    NAF: neural attenuation fields for sparse-view CBCT reconstruction , author=. International Conference on Medical Image Computing and Computer-Assisted Intervention , pages=. 2022 , organization=

    2022

Show all 68 references
  1. [9]

    CVPR , year=

    Structure-Aware Sparse-View X-ray 3D Reconstruction , author=. CVPR , year=

  2. [10]

    ECCV , year=

    Radiative gaussian splatting for efficient x-ray novel view synthesis , author=. ECCV , year=

  3. [11]

    Advances in Neural Information Processing Systems (NeurIPS) , year=

    R ^2 -Gaussian: Rectifying Radiative Gaussian Splatting for Tomographic Reconstruction , author=. Advances in Neural Information Processing Systems (NeurIPS) , year=

  4. [12]

    arXiv preprint arXiv:2307.05801 , year=

    Differentiable forward projector for X-ray computed tomography , author=. arXiv preprint arXiv:2307.05801 , year=

    arXiv

  5. [13]

    Ultramicroscopy , volume=

    The ASTRA Toolbox: A platform for advanced algorithm development in electron tomography , author=. Ultramicroscopy , volume=. 2015 , publisher=

    2015

  6. [14]

    Optics express , volume=

    Fast and flexible X-ray tomography using the ASTRA toolbox , author=. Optics express , volume=. 2016 , publisher=

    2016

  7. [15]

    Biomedical Physics & Engineering Express , volume=

    TIGRE: a MATLAB-GPU toolbox for CBCT image reconstruction , author=. Biomedical Physics & Engineering Express , volume=. 2016 , publisher=

    2016

  8. [16]

    arXiv preprint arXiv:2404.03126 , year=

    Gaspct: Gaussian splatting for novel ct projection view synthesis , author=. arXiv preprint arXiv:2404.03126 , year=

    arXiv

  9. [17]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Intratomo: self-supervised learning-based tomography via sinogram synthesis and prediction , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  10. [18]

    GS-CPR: Efficient Camera Pose Refinement via 3D Gaussian Splatting , author=

  11. [19]

    NeurIPS , year=

    Sparse-view Pose Estimation and Reconstruction via Analysis by Generative Synthesis , author=. NeurIPS , year=

  12. [20]

    arXiv preprint arXiv:2502.09563 , year=

    Self-Calibrating Gaussian Splatting for Large Field of View Reconstruction , author=. arXiv preprint arXiv:2502.09563 , year=

    arXiv

  13. [21]

    arXiv preprint arXiv:2502.19800 , year=

    No Parameters, No Problem: 3D Gaussian Splatting without Camera Intrinsics and Extrinsics , author=. arXiv preprint arXiv:2502.19800 , year=

  14. [22]

    , author=

    3d gaussian splatting for real-time radiance field rendering. , author=. ACM Trans. Graph. , volume=

  15. [23]

    and Slaney, Malcolm , title =

    Kak, Avinash C. and Slaney, Malcolm , title =. 2001 , doi =

    2001

  16. [24]

    Online Geometric Calibration of Cone-Beam Computed Tomography for Arbitrary Imaging Objects , year=

    Meng, Yuanzheng and Gong, Hui and Yang, Xiaoquan , journal=. Online Geometric Calibration of Cone-Beam Computed Tomography for Arbitrary Imaging Objects , year=

  17. [25]

    A. H. Andersen and A. C. Kak , title =. Ultrasonic Imaging , volume =. 1984 , doi =

    1984

  18. [26]

    Physics in Medicine & Biology , volume=

    Image reconstruction in circular cone-beam computed tomography by constrained, total-variation minimization , author=. Physics in Medicine & Biology , volume=. 2008 , publisher=

    2008

  19. [27]

    ICCV , year =

    Yoonwoo Jeong and Seokjun Ahn and Christopher Choy and Animashree Anandkumar and Minsu Cho and and Jaesik Park , title =. ICCV , year =

  20. [28]

    2020 , booktitle=

    NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis , author=. 2020 , booktitle=

    2020

  21. [29]

    ACM Trans

    Thomas M\"uller and Alex Evans and Christoph Schied and Alexander Keller , title =. ACM Trans. Graph. , issue_date =. 2022 , pages =. doi:10.1145/3528223.3530127 , publisher =

    2022

  22. [30]

    Attention is All you Need , url =

    Vaswani, Ashish and Shazeer, Noam and Parmar, Niki and Uszkoreit, Jakob and Jones, Llion and Gomez, Aidan N and Kaiser, ukasz and Polosukhin, Illia , booktitle =. Attention is All you Need , url =

  23. [31]

    MedNeRF: Medical Neural Radiance Fields for Reconstructing 3D-aware CT-Projections from a Single X-ray , booktitle =

    Abril Corona. MedNeRF: Medical Neural Radiance Fields for Reconstructing 3D-aware CT-Projections from a Single X-ray , booktitle =. 2022 , url =. doi:10.1109/EMBC48229.2022.9871757 , timestamp =

    2022

  24. [32]

    proceedings of Medical Image Computing and Computer Assisted Intervention -- MICCAI 2024 , year =

    Liu, Jiachen and Bai, Xiangzhi , title =. proceedings of Medical Image Computing and Computer Assisted Intervention -- MICCAI 2024 , year =

    2024

  25. [33]

    2015 , month =

    Ferrucci, Massimiliano and Leach, Richard K and Giusca, Claudiu and Carmignato, Simone and Dewulf, Wim , title =. 2015 , month =. doi:10.1088/0957-0233/26/9/092003 , url =

    2015 doi

  26. [34]

    Yang, Kai and Kwan, Alexander L. C. and Miller, DeWitt F and Boone, John M , title =. Medical Physics , issn =. 2006 , month =. doi:10.1118/1.2198187 , url =

    2006 doi

  27. [35]

    2023 , issn =

    Geometry calibration for a dental cone-beam CT system with an offset detector , journal =. 2023 , issn =. doi:https://doi.org/10.1016/j.precisioneng.2022.11.004 , url =

    2023 doi

  28. [36]

    Xu, Jingyan and Tsui, Benjamin M. W. , journal=. An Analytical Geometric Calibration Method for Circular Cone-Beam Geometry , year=

  29. [37]

    2016 , month =

    Ouadah, S and Stayman, J W and Gang, G J and Ehtiati, T and Siewerdsen, J H , title =. 2016 , month =. doi:10.1088/0031-9155/61/7/2613 , url =

    2016 doi

  30. [38]

    2008 , month =

    Kyriakou, Y and Lapp, R M and Hillebrand, L and Ertel, D and Kalender, W A , title =. 2008 , month =. doi:10.1088/0031-9155/53/22/001 , url =

    2008 doi

  31. [39]

    and Pfister, H

    Zwicker, M. and Pfister, H. and van Baar, J. and Gross, M. , journal=. EWA splatting , year=

  32. [40]

    arXiv preprint arXiv:1812.01537 , year=

    A micro lie theory for state estimation in robotics , author=. arXiv preprint arXiv:1812.01537 , year=

    arXiv

  33. [41]

    Kingma and Jimmy Ba , editor =

    Diederik P. Kingma and Jimmy Ba , editor =. Adam:. 3rd International Conference on Learning Representations,. 2015 , url =

    2015

  34. [42]

    PyTorch: An Imperative Style, High-Performance Deep Learning Library , author =

  35. [43]

    IEEE Transactions on Medical Imaging , year=

    A gradient-based approach to fast and accurate head motion compensation in cone-beam CT , author=. IEEE Transactions on Medical Imaging , year=

  36. [44]

    Physics in Medicine & Biology , volume=

    Gradient-based geometry learning for fan-beam CT reconstruction , author=. Physics in Medicine & Biology , volume=. 2023 , publisher=

    2023

  37. [45]

    IEEE transactions on medical imaging , volume=

    A fully differentiable framework for 2D/3D registration and the projective spatial transformers , author=. IEEE transactions on medical imaging , volume=. 2023 , publisher=

    2023

  38. [46]

    ACM Transactions on Graphics (TOG) , volume=

    Neat: Neural adaptive tomography , author=. ACM Transactions on Graphics (TOG) , volume=. 2022 , publisher=

    2022

  39. [47]

    International Conference on Medical Image Computing and Computer-Assisted Intervention , pages=

    Differentiable Score-Based Likelihoods: Learning CT Motion Compensation from Clean Images , author=. International Conference on Medical Image Computing and Computer-Assisted Intervention , pages=. 2024 , organization=

    2024

  40. [48]

    13th conference on industrial computed tomography (iCT) , year=

    Uncalibrated CT reconstruction for one-shot scanning of arbitrary trajectories , author=. 13th conference on industrial computed tomography (iCT) , year=

  41. [49]

    2023 IEEE 20th International Symposium on Biomedical Imaging (ISBI) , pages=

    Joint rigid motion correction and sparse-view ct via self-calibrating neural field , author=. 2023 IEEE 20th International Symposium on Biomedical Imaging (ISBI) , pages=. 2023 , organization=

    2023

  42. [50]

    NVIDIA, July , volume=

    Cuda c programming guide , author=. NVIDIA, July , volume=

  43. [51]

    Radiographics , volume=

    Artifacts in CT: recognition and avoidance , author=. Radiographics , volume=. 2004 , publisher=

    2004

  44. [52]

    Nordic Conference on Digital Health and Wireless Solutions , pages=

    Computed Tomography Artefact Detection Using Deep Learning—Towards Automated Quality Assurance , author=. Nordic Conference on Digital Health and Wireless Solutions , pages=. 2024 , organization=

    2024

  45. [53]

    Medical Imaging 2019: Physics of Medical Imaging , volume=

    Deep learning-based artifact detection for diagnostic CT images , author=. Medical Imaging 2019: Physics of Medical Imaging , volume=. 2019 , organization=

    2019

  46. [54]

    Medical Physics , volume=

    Deep learning-based conditional inpainting for restoration of artifact-affected 4D CT images , author=. Medical Physics , volume=. 2024 , publisher=

    2024

  47. [55]

    Quaternions and spatial rotation

    Wikipedia contributors. Quaternions and spatial rotation. 2025 , url =

    2025

  48. [56]

    Journal of Nondestructive Evaluation , volume=

    Feature-sensitive ring artifact correction for computed tomography , author=. Journal of Nondestructive Evaluation , volume=. 2023 , publisher=

    2023

  49. [57]

    Optics express , volume=

    Ring artifact suppression in X-ray computed tomography using a simple, pixel-wise response correction , author=. Optics express , volume=. 2019 , publisher=

    2019

  50. [58]

    Sensors , volume=

    Complete ring artifacts reduction procedure for lab-based x-ray nano ct systems , author=. Sensors , volume=. 2021 , publisher=

    2021

  51. [59]

    Physics in Medicine & Biology , volume=

    Iterative image-domain ring artifact removal in cone-beam CT , author=. Physics in Medicine & Biology , volume=. 2017 , publisher=

    2017

  52. [60]

    Medical physics , volume=

    A streak artifact reduction algorithm in sparse-view CT using a self-supervised neural representation , author=. Medical physics , volume=. 2022 , publisher=

    2022

  53. [61]

    Proceedings of the Asian Conference on Computer Vision , pages=

    Improving the quality of sparse-view cone-beam computed tomography via reconstruction-friendly interpolation network , author=. Proceedings of the Asian Conference on Computer Vision , pages=

  54. [62]

    IEEE Transactions on Computational Imaging , volume=

    Sparse-view cone beam CT reconstruction using data-consistent supervised and adversarial learning from scarce training data , author=. IEEE Transactions on Computational Imaging , volume=. 2023 , publisher=

    2023

  55. [63]

    ANNALI DELL'UNIVERSITA'DI FERRARA , volume=

    A comparison of regularization models for few-view CT image reconstruction , author=. ANNALI DELL'UNIVERSITA'DI FERRARA , volume=. 2022 , publisher=

    2022

  56. [64]

    arXiv preprint arXiv:2410.07552 , year=

    Methods for Few-View CT Image Reconstruction , author=. arXiv preprint arXiv:2410.07552 , year=

    arXiv

  57. [65]

    X-Field: A Physically Informed Representation for 3D X-ray Reconstruction , author=

  58. [66]

    , author=

    Reduction of CT artifacts caused by metallic implants. , author=. Radiology , volume=

  59. [67]

    Medical physics , volume=

    Normalized metal artifact reduction (NMAR) in computed tomography , author=. Medical physics , volume=. 2010 , publisher=

    2010

  60. [68]

    Unsupervised Polychromatic Neural Representation for

    Qing Wu and Lixuan Chen and Ce Wang and Hongjiang Wei and S Kevin Zhou and Jingyi Yu and Yuyao Zhang , booktitle=. Unsupervised Polychromatic Neural Representation for

Pith tools

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