Pith. sign in

REVIEW 5 major objections 5 minor 48 references

Motion Artifact Removal in Pixel-Frequency Domain via Alternate Masks and Diffusion Model

T0 review · 5 major / 5 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 PFAD removes MRI motion artifacts at inference without paired data by locking low-frequency k-space and alternating complementary masks across pixel and frequency domains, outperforming supervised and unsupervised baselines on brain…

desk verdict Clever diffusion-based recipe for MRI motion artifact removal, but its quantitative benchmark leans on the same high-frequency assumption the method makes, so the reported gains are less decisive than the tables suggest. read the letter →

arxiv 2412.07590 v2 pith:QGPIWVR6 submitted 2024-12-10 eess.IV cs.CV

Jiahua Xu , Dawei Zhou , Lei Hu , Jianfeng Guo , Feng Yang , Zaiyi Liu , Nannan Wang , Xinbo Gao This is my paper · ORCID
classification eess.IVcs.CV
keywords MRImotionartifactremovaldiffusionmodelunsupervisedimagerestorationk-spaceguidancealternatecomplementarymaskspixel-frequencydomainDDPMmedicalquality
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes PFAD, an unsupervised inference-time method for removing MRI motion artifacts with a pretrained diffusion model. Its central claim is that a motion-corrupted scan can be purified without paired clean data by copying the corrupted image's low-frequency k-space into every reverse-diffusion step and using alternating complementary masks to blend corrupted and generated content in both frequency and pixel domains. If this claim holds, artifact cleanup becomes practical in the clinic, because a model trained once on clean images of an anatomy can be applied to whatever corrupted scan arrives. The paper reports that PFAD beats supervised (Pix2pix) and unsupervised (CycleGAN, UDDN, DR2, GDP) baselines on quantitative metrics for simulated brain, knee, and abdominal motion artifacts, and receives the highest radiologist Likert scores on real clinical abdominal images.

What carries the argument

The central mechanism is the alternate complementary mask pair $M_t$ and $1-M_t$, with $M_t = \omega_t m_t$ and $m_t$ flipped at every reverse step ($m_{t-1}=1-m_t$). It splits both domains into alternating checkerboard areas, so the artifact-corrupted high-frequency and pixel content is never used wholesale, while useful structure from the actual scan is still fed to the diffusion model; the time-varying $\omega_t=1-\sqrt{\bar{\alpha}_t}$ weakens the corrupted guidance in late reverse steps. The other load-bearing piece is the low-frequency lock $\Phi_l(f_{x'_{t-1}})=\Phi_l(f_{x_{\mathrm{ori}}})$, which freezes k-space below cutoff $\pi/10$ from the corrupted input so tissue texture stays anchored. These are combined by $x'_{t-1}$ in the frequency domain and $x''_{t-1}$ in the pixel domain, then blended by $\gamma_t$ into $\tilde{x}_{t-1}$.

What would settle it

Simulate rigid-motion phase corruption applied only to low-frequency k-space lines with $|k_y| < \pi/10$ (below the cutoff the method treats as clean), with the same amplitude used for high-frequency corruption. If PFAD's output then retains visible ghosting or its PSNR and SSIM drop far below the high-frequency-corruption case, the low-frequency-lock premise is falsified. Alternatively, on retrospectively motion-tracked real data, measure the k-space phase-error spectrum: energy below $\pi/10$ would directly violate the assumption.

Watch

Extended reading notes

Core claim

On its own terms, the discovery is that the diffusion reverse process can be reorganized in two coupled domains to remove motion artifacts without training on corrupted images. At each reverse step the method first lets the pretrained diffusion model predict the previous timestep, then overwrites low-frequency k-space of that prediction with the low-frequency k-space of the corrupted input, and overwrites part of the high-frequency k-space with high-frequency content of the corrupted input through a mask that flips every step. In the pixel domain, it mixes the diffusion prediction with a forward-noised version of the corrupted image under the complementary mask. A step-dependent weight $\omega_t = 1-\sqrt{\bar{\alpha}_t}$ gradually reduces how much corrupted guidance enters as the image becomes cleaner, and a scalar $\gamma_t$ shifts emphasis from frequency-domain texture anchoring to pixel-domain sharpness. The paper claims this preserves tissue textures and destroys artifact structure, yielding top quantitative results on simulated brain, knee, and abdominal data and the best radiologist ratings on real abdominal images.

Load-bearing premise

The argument assumes motion artifacts live almost entirely in the high-frequency part of k-space above a cutoff of $\pi/10$ along the phase-encoding direction, and the simulated benchmark is generated with phase perturbations placed only above that same cutoff, so the main quantitative evidence inherits the assumption rather than testing it.

Editorial extensions

If this is right

  • Hospitals can clean new motion-corrupted scans with no paired rescans: the same pretrained diffusion model, trained once on clean images of that anatomy, purifies each incoming corrupted image at inference.
  • Because the corrupted image's own low-frequency k-space is preserved, the output stays anchored to the actual scan, reducing the risk that the generative model invents plausible but wrong anatomy.
  • Alternating complementary masks distribute artifact removal over the whole image across reverse steps, so no fixed region is left either fully corrupted or fully hallucinated.
  • The time-varying mask weight and domain-balance parameter give a principled schedule for leaning on the corrupted guidance early and the generated content late, which the ablation studies show is needed.
  • Radiologist evaluation on real clinical images indicates the benefit transfers beyond the simulated benchmark, which matters because simulated artifacts share the method's own high-frequency assumption.

Reading between the lines

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

  • A natural next step the paper does not take is replacing the fixed $\pi/10$ cutoff with a learned or estimated corruption map, since real motion can in principle corrupt low-frequency k-space and would then be locked into the output by the low-frequency guidance.
  • The method still needs a clean-image pretraining set for each anatomy and contrast; a hospital with only corrupted archives would have to borrow clean data or pretrain on a compatible public set before PFAD could run.
  • Because the mechanism attacks structured phase errors rather than image content, it may carry over to other k-space phase artifacts such as respiratory ghosting or EPI Nyquist ghosts, with the same alternate-mask design.
  • The grid size of the checkerboard mask is a free parameter with visible quality effects, so clinical deployment would likely tune it per anatomy rather than using a single size.
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

5 major / 5 minor

Summary. The paper proposes PFAD (Pixel-Frequency Alternate Masks Diffusion), an unsupervised method for removing MRI motion artifacts. A diffusion model pre-trained on clean images is guided during reverse diffusion by the low-frequency k-space of the corrupted image, while alternating complementary masks in the pixel and frequency domains are used to destroy artifact structure and exploit usable information. A time-varying hyperparameter a balances frequency- and pixel-domain guidance. The method is evaluated on simulated motion-corrupted brain (HCP), knee (fastMRI), and abdominal MRI, and on 40 real clinical abdominal images, with quantitative metrics and radiologist Likert scoring.

Significance. If the reported gains are robust, PFAD would be a useful contribution to unsupervised MRI motion-artifact correction: it needs no paired data, preserves a clear role for k-space information, and provides a concrete inference algorithm with released code. The paper is honest about its 'lenient' assumption that artifacts are high-frequency, and it includes ablation studies and a small radiologist study. However, the central quantitative claim is currently supported mainly by a simulation pipeline that embeds the method's own cutoff assumption, and one key hyperparameter is selected on the evaluation set. These issues make the superiority claim (Abstract, Table 1) not yet convincing; the real-image evaluation is too small and too subjective to carry the claim on its own.

major comments (5)
  1. [Supplementary A.2, Eqs. (13)-(14); Related Work] The quantitative benchmark is circular with respect to the method's central assumption. In Related Work the authors adopt the assumption that motion artifacts are confined to |ky| > pi/10, and Eq. (4) freezes all low-frequency content of the corrupted image as guidance. The simulated artifacts in Supplementary Eqs. (13)-(14) are generated by applying phase perturbations only for |ky| > k0, with k0 fixed to pi/10. Thus the test data are produced under exactly the premise that the method's filter uses, and the large gains over baselines in Table 1 may reflect this alignment rather than an ability to remove clinically realistic artifacts. The cutoff experiment in Table 7 does not resolve this because it is also conducted on images simulated with k0 = pi/10; selecting pi/10 as optimal under that generator is an expected outcome, not evidence about real artifact spectra. Please add experiments with artifact perturbations extending below pi/10 (e.g., k0 = pi/20 or with low-frequency leakage), or provide real artifact data with paired ground truth or quantitative spectral characterisation.
  2. [Table 5 and 'Hyperparameter Study'] The balance parameter a is selected on the evaluation set, which inflates the reported metrics. Table 5 lists PSNR/SSIM/LPIPS/... for different a values and states that 'we choose the value of a for the best case of the total metric, where a is equal to 0.7.' These appear to be the same test-set metrics reported in Table 1 for the brain dataset, so the method's hyperparameter has been tuned to the test data. Please either (i) derive a from a separate validation split and report test metrics only for the fixed value, or (ii) clearly state that a was selected on a validation set and provide those validation results. The ad hoc 'Total' metric (sum of some metrics minus others, with no normalisation or justification) should also be justified or replaced with a standard criterion.
  3. [Eq. (4); Related Work] The central mechanism is not robust to low-frequency artifact energy. Because Eq. (4) copies the entire low-frequency component of the corrupted image into every reverse step, any real artifact energy below the pi/10 cutoff is frozen into the output and cannot be removed. The paper explicitly acknowledges the assumption is 'lenient', but it never tests the failure mode. Real patient motion during the initial acquisition of central k-space, or respiratory/bulk motion with low-frequency components, violates this assumption. Please provide an experiment that quantifies performance degradation as artifact energy is progressively moved below pi/10, or a spectral analysis of the real clinical images showing that their artifact energy is indeed confined to |ky| > pi/10. Without such evidence, the claim that PFAD works on 'real clinical images' (Figure 5, Table 2) is not strongly supported.
  4. [Table 1; Comparison Approaches] The closest prior work is not compared. The paper adopts its pi/10 assumption from Oh et al. (2023) and describes that work as using score-based diffusion with measured k-space values, but Oh et al. is absent from Table 1 and from the qualitative comparisons. Since Oh et al. is a recent diffusion-based method designed for the same task (MR motion artifact reduction), omitting it undercuts the claim of 'superior performance.' Please include Oh et al. (and, if practical, one more recent k-space-aware baseline) under the same evaluation protocol.
  5. [Table 1; Supplementary B.2, Table 6] Statistical reporting is incomplete. Table 1 reports only means for each metric, with a footnote that Mann-Whitney U tests were used, but no variances, confidence intervals, or effect sizes are given for the quantitative metrics. With the reported differences being small in several cases (e.g., brain PSNR 27.60 for PFAD vs 26.76 for Pix2pix), the reader cannot judge whether the differences are practically meaningful. Please report standard deviations or 95% confidence intervals, and specify the number of test images for each dataset. For the radiologist study, only 40 real images were scored; this is a small sample and the inclusion of variance in Table 6 is helpful, but the protocol should state whether the two radiologists' scores were averaged or adjudicated per-image, and how inter-rater agreement was quantified.
minor comments (5)
  1. [Introduction, list of contributions] In the fourth contribution the phrase 'demonstrates the of our method in metrics' is missing a word ('effectiveness' or 'superiority'); please fix this typo.
  2. [Eq. (6) and surrounding text] The modulus operation in Eq. (6) discards phase information, and the text states that pixel-domain processing compensates for this. Please make this explanation more precise: what exactly is lost, and how does the pixel-domain branch recover it? A short worked illustration would help.
  3. [Figure 4 caption] The sentence describing the difference heatmap ('The color ranges from blue to red, indicating differences from small to large, with deeper colors representing smaller differences') is self-contradictory as written; please clarify whether deeper colors mean larger or smaller errors.
  4. [Related Work, paragraph on Motion Artifact Removal] The sentence 'Noting that most previous works are based on GANs and remove artifacts only in pixel domain' is inaccurate for UDDN, which is described earlier in the same paragraph as removing artifacts in the frequency domain; please rephrase to acknowledge UDDN's k-space operation.
  5. [Algorithm 1, line 3] The definition of omega_i in line 3 could be confused with the noise schedule; please add a short comment that omega_i is the standard DDPM posterior weight (1 - sqrt(alpha_bar_i)) used in RePaint-style guidance.

Circularity Check

3 steps flagged · score 5.0 of 10

Simulated benchmarks are generated under the same high-frequency phase-perturbation cutoff (k0=π/10) that the method hard-codes as its low-frequency guidance, so the quantitative gains are partly self-confirming; real-image evaluation provides independent support.

  1. self definitional [Related Work and Methodology, Eq. (4)]
    "we adopt a lenient assumption that the motion artifacts in MRI images predominantly manifest in the high-frequency region along the phase encoding lines... set it to be greater than π/10 (Oh et al. 2023). ... Φl(fx′ t−1) = Φl(fxori), (4)"

    The method's central guidance is defined as the corrupted image's low-frequency content below π/10, preserved unchanged at every reverse step (Eq. 4-5). Whether this guidance is artifact-free is exactly the assumed claim, not a derived or tested fact. The assumption is load-bearing: if real artifacts have energy below π/10, Eq. 4 freezes that artifact energy into every step and the method cannot remove it.

  2. other [Supplementary Material A.2, Eqs. (13)-(14)]
    "Φ(ky) = ... |ky| > k0 ... 0, otherwise ... In this simulation, k0 is fixed to π/10 ... For abdominal dataset ... |ky| > k0 ... k0 is fixed to π/10"

    The only quantitative benchmarks (Table 1) are simulated using exactly the same cutoff k0=π/10 that defines the method's low-frequency preservation filter Φl. Consequently, the test images contain no artifact energy in the region that Eq. 4 keeps intact, so the reported PSNR/SSIM/LPIPS superiority is measured entirely inside the method's own premise. The benchmark cannot detect low-frequency artifact leakage and partly confirms the assumption rather than testing the method.

1 more flagged steps
  1. fitted input called prediction [Supplementary Material B.4, Table 7]
    "Experiment on the cutoff frequency of the filter. ... π/5 25.89 ... π/10 27.60 ... π/20 26.34 ..."

    The cutoff hyperparameter is selected as optimal on simulated images whose artifacts were generated with k0=π/10 (Eq. 13-14). Finding π/10 optimal on data built with that same cutoff is expected by construction; it is not independent evidence about the correct cutoff for real clinical motion artifacts, though the paper presents it as guidance for clinical use.

full rationale

PFAD is not a derivation from first principles; it is an inference-time method whose core operation is to freeze the corrupted image's low-frequency k-space as guidance (Eq. 4) under an explicitly 'lenient assumption' that motion artifacts live above π/10. The quantitative evaluation in Table 1 is performed on images simulated with the identical cutoff (Supplementary Eqs. 13-14, k0=π/10), so the artifact model and the method's filter coincide: simulated artifacts are, by construction, absent from the low-frequency region the method preserves. This makes part of the reported improvement a self-consistent outcome of matching the benchmark to the method's premise rather than an independent test of it. The same coupling affects the cutoff ablation (Table 7), where π/10 is selected as optimal on data generated with π/10. Against this must be weighed the real clinical T2-weighted images and radiologist Likert evaluation (Table 2), which are not generated under the simulation assumption and provide independent, though qualitative, evidence that PFAD removes visible artifacts; there is also no self-citation chain or fitted parameter masquerading as a prediction. The circularity is therefore partial and confined to the quantitative simulated benchmark, not total. Score 5 reflects one central quantitative claim that reduces partly by construction while the method retains independent real-image support.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The method rests on a small number of domain assumptions carried from Oh et al. (2023) and on ad hoc design choices such as alternate masks, modulus in frequency reconstruction, and a test-set-tuned balance parameter. No new physical entities are introduced.

free parameters (3)
  • a (dual-domain balance hyperparameter) = 0.7
    Tuned on the HCP test set by maximizing an ad hoc total metric in Table 5, with no separate validation split.
  • Low/high frequency cutoff = pi/10
    Adopted from Oh et al. 2023; the same value is used as k0 to simulate motion artifacts in the test data, so the evaluation is aligned with the method's core assumption.
  • Checkerboard mask grid size = 16
    Selected via ablation on the HCP test set (Table 8), with no separate validation split.
assumptions (6)
  • domain assumption Motion artifacts in MRI predominantly appear in high-frequency k-space along phase encoding lines.
    Adopted from Oh et al. (2023) in Related Work; motivates keeping the low-frequency component untouched in Eq. 4. The same assumption is used to generate simulated test data in Eq. (13)/(14), so it is not tested independently on simulated images.
  • domain assumption A cutoff frequency of pi/10 separates low and high frequency for artifact removal.
    Used in filters Phi_l and Phi_h; the simulation uses k0 = pi/10 as the perturbation threshold, so the chosen cutoff is matched to the test data.
  • domain assumption A diffusion model trained on clean images represents the clean MRI distribution well enough for reverse-step guidance.
    The whole inference relies on D_theta (Eq. 2-3) generating realistic clean images; no calibration against the corrupted domain is performed.
  • ad hoc to paper Alternating complementary masks over reverse steps remove artifacts while preserving useful information.
    Proposed in Eq. (8)-(9); the mechanism is described qualitatively and supported only by ablation results, not by formal analysis.
  • ad hoc to paper Taking the modulus in Eq. (6) loses phase information, but pixel-domain processing compensates for this loss.
    The Methodology section claims pixel-domain removal handles slight phase loss; no quantitative support is provided.
  • domain assumption Simulated motion artifacts with k0 = pi/10 are representative of real clinical motion artifacts.
    All quantitative evaluation uses this simulation; only a small qualitative radiologist evaluation on 40 real images tests the assumption.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Motion Artifact Removal in Pixel-Frequency Domain via Alternate Masks and Diffusion Model." pith.science (2026). https://pith.science/paper/QGPIWVR6

@misc{pith2026241207590,
  author       = {Pith},
  title        = {Pith review of: Motion Artifact Removal in Pixel-Frequency Domain via Alternate Masks and Diffusion Model},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QGPIWVR6}},
  note         = {Machine review of arXiv:2412.07590}
}
read the original abstract

Motion artifacts present in magnetic resonance imaging (MRI) can seriously interfere with clinical diagnosis. Removing motion artifacts is a straightforward solution and has been extensively studied. However, paired data are still heavily relied on in recent works and the perturbations in k-space (frequency domain) are not well considered, which limits their applications in the clinical field. To address these issues, we propose a novel unsupervised purification method which leverages pixel-frequency information of noisy MRI images to guide a pre-trained diffusion model to recover clean MRI images. Specifically, considering that motion artifacts are mainly concentrated in high-frequency components in k-space, we utilize the low-frequency components as the guide to ensure correct tissue textures. Additionally, given that high-frequency and pixel information are helpful for recovering shape and detail textures, we design alternate complementary masks to simultaneously destroy the artifact structure and exploit useful information. Quantitative experiments are performed on datasets from different tissues and show that our method achieves superior performance on several metrics. Qualitative evaluations with radiologists also show that our method provides better clinical feedback. Our code is available at https://github.com/medcx/PFAD.

Figures

Figures reproduced from arXiv: 2412.07590 by the authors.

Figure 1
Figure 1. Schematic of our research problem and solution. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The framework of PFAD. The green area is the artifact removal process in frequency domain. This process is guided [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Alternate complementary masks. In each iteration, [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Comparison of results on simulated brain images: (a) Artifact images, (b) Pix2pix, (c) CycleGAN, (d) UDDN, (e) [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Comparison of results on real images: (a) Images with real motion artifacts, (b) Pix2pix, (c) CycleGAN, (d) UDDN, (e) [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Comparison of results on simulated knee images: (a) Artifact images, (b) Pix2pix, (c) CycleGAN, (d) UDDN, (e) [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]
Figure 7
Figure 7. Figure 7: Comparison of results on simulated abdominal images: (a) Artifact images, (b) Pix2pix, (c) CycleGAN, (d) UDDN, [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: Comparison of visualization results on ablation study of pixel-frequency domain: From top to bottom, each row [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 27 canonical work pages

  1. [1]

    Armanious, K.; Tanwar, A.; Abdulatif, S.; K \"u stner, T.; Gatidis, S.; and Yang, B. 2020. Unsupervised adversarial correction of rigid MR motion artifacts. In 2020 IEEE 17th International Symposium on Biomedical Imaging (ISBI), 1494--1498. IEEE

    work page 2020

  2. [2]

    Budde, J.; Shajan, G.; Scheffler, K.; and Pohmann, R. 2014. Ultra-high resolution imaging of the human brain using acquisition-weighted imaging at 9.4 T. Neuroimage, 86: 592--598

    work page 2014

  3. [3]

    Chen, W.-K. 1990. Linear Networks And Systems: Algorithms And Computer-Aided Implementations (In 2 Volumes), volume 3. World Scientific

    work page 1990

  4. [4]

    Simultaneous super-resolution and motion artifact removal in diffusion-weighted MRI using unsupervised deep learning

    Chung, H.; Kim, J.; Yoon, J. H.; Lee, J. M.; and Ye, J. C. 2021. Simultaneous super-resolution and motion artifact removal in diffusion-weighted MRI using unsupervised deep learning. arXiv preprint arXiv:2105.00240

    work page Pith review arXiv 2021

  5. [5]

    Cruz, G.; Atkinson, D.; Buerger, C.; Schaeffter, T.; and Prieto, C. 2016. Accelerated motion corrected three-dimensional abdominal MRI using total variation regularized SENSE reconstruction. Magnetic resonance in medicine, 75(4): 1484--1498

    work page 2016

  6. [6]

    M.; and Prieto, C

    Cruz, G.; Atkinson, D.; Henningsson, M.; Botnar, R. M.; and Prieto, C. 2017. Highly efficient nonrigid motion-corrected 3D whole-heart coronary vessel wall imaging. Magnetic resonance in medicine, 77(5): 1894--1908

    work page 2017

  7. [7]

    Dhariwal, P.; and Nichol, A. 2021. Diffusion models beat gans on image synthesis. Advances in neural information processing systems, 34: 8780--8794

    2021

  8. [8]

    Fei, B.; Lyu, Z.; Pan, L.; Zhang, J.; Yang, W.; Luo, T.; Zhang, B.; and Dai, B. 2023. Generative diffusion prior for unified image restoration and enhancement. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 9935--9946

    work page 2023

Show all 48 references
  1. [9]

    S.; S rensen, T

    Hansen, M. S.; S rensen, T. S.; Arai, A. E.; and Kellman, P. 2012. Retrospective reconstruction of high temporal resolution cine images from real-time MRI using iterative motion correction. Magnetic Resonance in Medicine, 68(3): 741--750

    2012

  2. [10]

    Ho, J.; Jain, A.; and Abbeel, P. 2020. Denoising diffusion probabilistic models. Advances in neural information processing systems, 33: 6840--6851

    2020

  3. [11]

    Ho, J.; and Salimans, T. 2022. Classifier-free diffusion guidance. arXiv preprint arXiv:2207.12598

    2022 arXiv

  4. [12]

    Hu, L.; Guo, X.; Zhou, D.; Wang, Z.; Dai, L.; Li, L.; Li, Y.; Zhang, T.; Long, H.; Yu, C.; et al. 2024 a . Development and Validation of a Deep Learning Model to Reduce the Interference of Rectal Artifacts in MRI-based Prostate Cancer Diagnosis. Radiology: Artificial Intellige...

    2024

  5. [13]

    Hu, L.; Zhou, D.; Xu, J.; Lu, C.; Han, C.; Shi, Z.; Zhu, Q.; Gao, X.; Wang, N.; and Liu, Z. 2024 b . Protecting Prostate Cancer Classification from Rectal Artifacts via Targeted Adversarial Training. IEEE Journal of Biomedical and Health Informatics

    2024

  6. [14]

    Isola, P.; Zhu, J.-Y.; Zhou, T.; and Efros, A. A. 2017. Image-to-image translation with conditional adversarial networks. In Proceedings of the IEEE conference on computer vision and pattern recognition, 1125--1134

    2017

  7. [15]

    u stner, T.; W \

    K \"u stner, T.; W \"u rslin, C.; Schwartz, M.; Martirosian, P.; Gatidis, S.; Brendle, C.; Seith, F.; Schick, F.; Schwenzer, N. F.; Yang, B.; et al. 2017. Self-navigated 4D cartesian imaging of periodic motion in the body trunk using partial k-space compressed sensing. Magneti...

    2017

  8. [16]

    Likert, R. 1932. A technique for the measurement of attitudes. Archives of psychology

    1932

  9. [17]

    Lugmayr, A.; Danelljan, M.; Romero, A.; Yu, F.; Timofte, R.; and Van Gool, L. 2022. Repaint: Inpainting using denoising diffusion probabilistic models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 11461--11471

    2022

  10. [18]

    Nichol, A.; Dhariwal, P.; Ramesh, A.; Shyam, P.; Mishkin, P.; McGrew, B.; Sutskever, I.; and Chen, M. 2021. Glide: Towards photorealistic image generation and editing with text-guided diffusion models. arXiv preprint arXiv:2112.10741

    2021 arXiv

  11. [19]

    E.; and Ye, J

    Oh, G.; Jung, S.; Lee, J. E.; and Ye, J. C. 2023. Annealed score-based diffusion model for mr motion artifact reduction. IEEE Transactions on Computational Imaging

    2023

  12. [20]

    E.; and Ye, J

    Oh, G.; Lee, J. E.; and Ye, J. C. 2021. Unpaired MR motion artifact deep learning using outlier-rejecting bootstrap aggregation. IEEE Transactions on Medical Imaging, 40(11): 3125--3139

    2021

  13. [21]

    Ramesh, A.; Dhariwal, P.; Nichol, A.; Chu, C.; and Chen, M. 2022. Hierarchical text-conditional image generation with clip latents. arXiv preprint arXiv:2204.06125, 1(2): 3

    2022 arXiv

  14. [22]

    Rombach, R.; Blattmann, A.; Lorenz, D.; Esser, P.; and Ommer, B. 2022. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 10684--10695

    2022

  15. [23]

    Ronneberger, O.; Fischer, P.; and Brox, T. 2015. U-net: Convolutional networks for biomedical image segmentation. In Medical image computing and computer-assisted intervention--MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, part II...

    2015

  16. [24]

    Saharia, C.; Chan, W.; Chang, H.; Lee, C.; Ho, J.; Salimans, T.; Fleet, D.; and Norouzi, M. 2022 a . Palette: Image-to-image diffusion models. In ACM SIGGRAPH 2022 conference proceedings, 1--10

    2022

  17. [25]

    L.; Ghasemipour, K.; Gontijo Lopes, R.; Karagol Ayan, B.; Salimans, T.; et al

    Saharia, C.; Chan, W.; Saxena, S.; Li, L.; Whang, J.; Denton, E. L.; Ghasemipour, K.; Gontijo Lopes, R.; Karagol Ayan, B.; Salimans, T.; et al. 2022 b . Photorealistic text-to-image diffusion models with deep language understanding. Advances in neural information processing sy...

    2022

  18. [26]

    Schulz, J.; Siegert, T.; Reimer, E.; Labadie, C.; Maclaren, J.; Herbst, M.; Zaitsev, M.; and Turner, R. 2012. An embedded optical tracking system for motion-corrected magnetic resonance imaging at 7T. Magnetic Resonance Materials in Physics, Biology and Medicine, 25: 443--453

    2012

  19. [27]

    R.; and Bovik, A

    Sheikh, H. R.; and Bovik, A. C. 2006. Image information and visual quality. IEEE Transactions on image processing, 15(2): 430--444

    2006

  20. [28]

    Si, C.; Huang, Z.; Jiang, Y.; and Liu, Z. 2023. Freeu: Free lunch in diffusion u-net. arXiv preprint arXiv:2309.11497

    2023 arXiv

  21. [29]

    Song, J.; Meng, C.; and Ermon, S. 2020. Denoising diffusion implicit models. arXiv preprint arXiv:2010.02502

    2020 arXiv

  22. [30]

    Tamada, D.; Kromrey, M.-L.; Ichikawa, S.; Onishi, H.; and Motosugi, U. 2020. Motion artifact reduction using a convolutional neural network for dynamic contrast enhanced MR imaging of the liver. Magnetic resonance in medical sciences, 19(1): 64--76

    2020

  23. [31]

    C.; Ugurbil, K.; Auerbach, E.; Barch, D.; Behrens, T

    Van Essen, D. C.; Ugurbil, K.; Auerbach, E.; Barch, D.; Behrens, T. E.; Bucholz, R.; Chang, A.; Chen, L.; Corbetta, M.; Curtiss, S. W.; et al. 2012. The Human Connectome Project: a data acquisition perspective. Neuroimage, 62(4): 2222--2231

    2012

  24. [32]

    S.; Alley, M

    Vasanawala, S. S.; Alley, M. T.; Hargreaves, B. A.; Barth, R. A.; Pauly, J. M.; and Lustig, M. 2010 a . Improved pediatric MR imaging with compressed sensing. Radiology, 256(2): 607--616

    2010

  25. [33]

    S.; Iwadate, Y.; Church, D

    Vasanawala, S. S.; Iwadate, Y.; Church, D. G.; Herfkens, R. J.; and Brau, A. C. 2010 b . Navigated abdominal T1-W MRI permits free-breathing image acquisition with less motion artifact. Pediatric radiology, 40: 340--344

    2010

  26. [34]

    C.; Sheikh, H

    Wang, Z.; Bovik, A. C.; Sheikh, H. R.; and Simoncelli, E. P. 2004. Image quality assessment: from error visibility to structural similarity. IEEE transactions on image processing, 13(4): 600--612

    2004

  27. [35]

    Wang, Z.; Zhang, Z.; Zhang, X.; Zheng, H.; Zhou, M.; Zhang, Y.; and Wang, Y. 2023. Dr2: Diffusion-based robust degradation remover for blind face restoration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 1704--1713

    2023

  28. [36]

    Wu, B.; Li, C.; Zhang, J.; Lai, H.; Feng, Q.; and Huang, M. 2023. Unsupervised dual-domain disentangled network for removal of rigid motion artifacts in MRI. Computers in Biology and Medicine, 165: 107373

    2023

  29. [37]

    Xue, W.; Zhang, L.; Mou, X.; and Bovik, A. C. 2013. Gradient magnitude similarity deviation: A highly efficient perceptual image quality index. IEEE transactions on image processing, 23(2): 684--695

    2013

  30. [38]

    Yang, Z.; Zhang, C.; and Xie, L. 2013. Sparse MRI for motion correction. In 2013 IEEE 10th International Symposium on Biomedical Imaging, 962--965. IEEE

    2013

  31. [39]

    Zaitsev, M.; Dold, C.; Sakas, G.; Hennig, J.; and Speck, O. 2006. Magnetic resonance imaging of freely moving objects: prospective real-time motion correction using an external optical motion tracking system. Neuroimage, 31(3): 1038--1050

    2006

  32. [40]

    Zaitsev, M.; Maclaren, J.; and Herbst, M. 2015. Motion artifacts in MRI: A complex problem with many partial solutions. Journal of Magnetic Resonance Imaging, 42(4): 887--901

    2015

  33. [41]

    J.; Defazio, A.; Stern, R.; Johnson, P.; Bruno, M.; et al

    Zbontar, J.; Knoll, F.; Sriram, A.; Murrell, T.; Huang, Z.; Muckley, M. J.; Defazio, A.; Stern, R.; Johnson, P.; Bruno, M.; et al. 2018. fastMRI: An open dataset and benchmarks for accelerated MRI. arXiv preprint arXiv:1811.08839

    2018 arXiv

  34. [42]

    Zhang, L.; Rao, A.; and Agrawala, M. 2023. Adding conditional control to text-to-image diffusion models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 3836--3847

    2023

  35. [43]

    Zhang, L.; Shen, Y.; and Li, H. 2014. VSI: A visual saliency-induced index for perceptual image quality assessment. IEEE Transactions on Image processing, 23(10): 4270--4281

    2014

  36. [44]

    Zhang, L.; Zhang, L.; Mou, X.; and Zhang, D. 2011. FSIM: A feature similarity index for image quality assessment. IEEE transactions on Image Processing, 20(8): 2378--2386

    2011

  37. [45]

    A.; Shechtman, E.; and Wang, O

    Zhang, R.; Isola, P.; Efros, A. A.; Shechtman, E.; and Wang, O. 2018. The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE conference on computer vision and pattern recognition, 586--595

    2018

  38. [46]

    Zhu, J.-Y.; Park, T.; Isola, P.; and Efros, A. A. 2017. Unpaired image-to-image translation using cycle-consistent adversarial networks. In Proceedings of the IEEE international conference on computer vision, 2223--2232

    2017

  39. [47]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all...

  40. [48]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

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