Pith. sign in

REVIEW 5 major objections 7 minor 2 cited by

Boosting ViT-based MRI Reconstruction from the Perspectives of Frequency Modulation, Spatial Purification, and Scale Diversification

T0 review · 5 major / 7 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read FPS-Former, a ViT-based MRI reconstruction framework, reports state-of-the-art accuracy on three public datasets at 4x and 8x acceleration by addressing frequency loss, token noise, and missing multi-scale modeling, with lower compute.

desk verdict Solid incremental architecture paper whose own Table 1 undercuts its 'consistently surpasses' claim; worth reviewing after a fix. read the letter →

arxiv 2412.10776 v1 pith:A7W3D3GE submitted 2024-12-14 eess.IV cs.AIcs.CV

Yucong Meng , Zhiwei Yang , Yonghong Shi , Zhijian Song This is my paper · ORCID
classification eess.IVcs.AIcs.CV
keywords acceleratedMRIreconstructionvisiontransformerfrequencymodulationattentionLaplacianpyramidspatialpurificationmulti-scalefeed-forwardnetworksingle-coilandmulti-coilimagerestoration
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

FPS-Former is a ViT-based MRI reconstruction framework built on the claim that the remaining gap in accelerated MRI comes from three architectural habits of vision transformers, not from data or scale. Self-attention acts as a low-pass filter, dense attention mixes content-irrelevant tokens, and the standard feed-forward network cannot represent multi-scale structure. The paper answers each with a targeted module—Laplacian-pyramid frequency recalibration, hash-based grouped attention, and multi-scale depthwise-convolution feed-forward paths—and reports state-of-the-art PSNR/SSIM/NMSE on CC359, fastMRI, and SKM-TEA at 4x and 8x acceleration while using less computation than its strongest rivals. If the reported comparisons are fair, the result is a practical recipe for transformer-based reconstruction that is both sharper and cheaper.

What carries the argument

The load-bearing object is the FPS block, a drop-in replacement for the standard transformer block, composed of three modules. FMAM builds a Laplacian pyramid of the input feature by subtracting progressively stronger Gaussian smoothings, computes a within-frequency attention map for each pyramid level, sums the maps, and applies the summed scores to the value computed from the original input, so high-frequency bands that plain self-attention suppresses are selectively amplified. SPAM flattens tokens, assigns each a hash code via a random projection, sorts by hash, and runs multi-head self-attention only inside each hash group, so unrelated tokens never attend to each other and quadratic cost is reduced. SDFN replaces the naive feed-forward MLP with two parallel depthwise-convolution branches ($3\times3$ and $5\times5$) whose outputs are cross-concatenated and refined, adding multi-scale local modeling. Around these blocks, HEFR, a self-attention-switched set of CNN experts (average pooling, separable convolutions, dilated convolutions), refines features at early and final stages, and a data-consistency layer ties the reconstruction back to the measured k-space.

What would settle it

Reproduce the comparison on CC359 at 4x acceleration under the paper's stated protocol (random 1D Cartesian masks, official data split, L1 loss, AdamW, 300K iterations) by training every baseline from its own official implementation. The central claim collapses if FPS-Former does not at least match its reported PSNR advantage over ReconFormer (34.38 vs 34.16) and AST (34.38 vs 32.78) under those identical conditions.

Watch

Extended reading notes

Core claim

The central claim is that a ViT for MRI reconstruction can outperform current state of the art by repairing three internal weaknesses rather than by scaling the model. The proposed FPS-Former couples frequency modulation attention (FMAM), which computes attention scores per level of a Gaussian Laplacian pyramid and re-weights the original value features with their sum, with spatial purification attention (SPAM), which hashes tokens, sorts them, and restricts multi-head attention to hash-neighbor groups, and a scale diversification feed-forward network (SDFN) that fuses $3\times3$ and $5\times5$ depthwise-convolution branches. On single-coil CC359 and fastMRI and multi-coil SKM-TEA at acceleration factors 4 and 8, and under Cartesian, radial, and random masks, the authors report that FPS-Former beats CNN and transformer baselines including ReconFormer and AST, with 12.51M parameters and 152G FLOPs on CC359. The ablation study attributes the gain to all four components, with SPAM contributing the largest single PSNR drop when removed.

Load-bearing premise

The entire comparison rests on the assumption that every baseline was trained and tested under the exact same protocol as FPS-Former; because the paper does not release code, a third party cannot independently verify that the reported baseline numbers came from that fair protocol.

Editorial extensions

If this is right

  • If the reported comparisons are correct, FPS-Former sets the new state of the art for accelerated MRI reconstruction on single-coil CC359 and fastMRI and multi-coil SKM-TEA, with larger margins at 8x than at 4x acceleration.
  • The ablations indicate each component carries weight independently: removing FMAM, SPAM, SDFN, or HEFR lowers CC359 PSNR at 4x by 0.24, 0.55, 0.43, and 0.40 dB respectively.
  • The hash-grouped attention in SPAM is the main cost saver, allowing FPS-Former to exceed ReconFormer's quality while using fewer FLOPs (152G vs 342G) and fewer parameters than AST (12.51M vs 26.10M).
  • The frequency and grouping modules also improve robustness across undersampling patterns, with FPS-Former leading under radial and random masks at 5x and 10x acceleration on CC359.

Reading between the lines

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

  • The hash-based grouping in SPAM is effectively a locality-sensitive hash; if it transfers, the same grouping trick could cheapen self-attention in other image-restoration and high-resolution vision tasks beyond MRI.
  • Because the method is architectural rather than MRI-specific, a natural testable extension is replacing the data-consistency layer with the appropriate forward operator to evaluate FPS blocks on other inverse problems, such as low-dose CT or super-resolution.
  • FMAM's spectral behavior could be probed directly: at intermediate layers, the frequency response of its attention maps should show measurably less high-frequency attenuation than a matched plain ViT, which is a sharper test than the end-to-end PSNR comparison.
  • The single L1 loss leaves room for extensions with perceptual or adversarial losses, which could improve visual fidelity even if PSNR changes little.
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 / 7 minor

Summary. The paper proposes FPS-Former, a ViT-based MRI reconstruction framework designed to address three issues: high-frequency information loss in self-attention, attention over content-irrelevant tokens, and lack of multi-scale feature modeling in feed-forward networks. It introduces three modules—Frequency Modulation Attention Module (FMAM), Spatial Purification Attention Module (SPAM), and Scale Diversification Feed-forward Network (SDFN)—plus a Hybrid Experts Feature Refinement (HEFR) module. The method is evaluated on CC359, fastMRI, and SKM-TEA under Cartesian, radial, and random undersampling masks at acceleration factors 4x, 5x, 8x, and 10x. The central claim is that FPS-Former outperforms state-of-the-art methods while requiring lower computational costs.

Significance. If the results are reproducible, the paper makes a useful architectural contribution by explicitly targeting known ViT weaknesses in low-level vision and MRI reconstruction. The ablation study in Table 4 supports the contribution of each module, and the efficiency analysis (152G FLOPs vs 342G for ReconFormer in Table 6) is a practical plus. The motivation for FMAM is grounded in prior observations about MSA as a low-pass filter, and the qualitative frequency-response analysis in Figure 4 provides supporting evidence. However, the significance is tempered by unresolved questions about protocol fairness, statistical robustness, and the accuracy of the paper's own performance claims.

major comments (5)
  1. [Abstract; Section 'Comparison with State-of-the-arts'; Table 1] The abstract's unqualified claim that FPS-Former 'outperforms state-of-the-art methods' and the text's claim that it 'consistently surpasses other Transformer-based approaches across different acceleration rates on both datasets' are contradicted by Table 1 on fastMRI at AF=4, where ReconFormer achieves PSNR 32.53 vs FPS-Former's 32.51. The paper should either qualify these claims (e.g., 'generally outperforms' or 'outperforms on most metrics and datasets') or provide statistical evidence (multiple runs with error bars) showing that the 0.02 dB deficit is not noise. This is the central empirical claim, so the overstatement needs correction.
  2. [Experimental Settings; Comparison with State-of-the-arts] The paper does not state whether baseline methods were retrained under the identical experimental protocol (same undersampling masks, data splits, loss function, and training schedule) or whether their reported numbers were taken from the original publications. Since the paper does not release code or detailed per-baseline training configurations, the claimed improvements over SwinMR, ReconFormer, Restormer, and AST could be artifacts of protocol differences. The authors should specify the protocol for each baseline and, ideally, provide code to allow third-party verification.
  3. [Spatial Purification Attention Module, Eq. (7)–(8)] The hash-based grouping in SPAM relies on random variables a and b, but the paper does not specify whether these are sampled once and fixed or re-sampled each forward pass, nor does it state the number of hash functions used. This makes the method's grouping behavior unreproducible. Additionally, this locality-sensitive hashing scheme is essentially the one used in the Reformer architecture (Kitaev et al., 2020), which should be cited. Clarify the hashing details and reference the prior work.
  4. [Frequency Modulation Attention Module, Eq. (3)] The Gaussian kernel in Eq. (3) uses the normalization constant 1/(σ√(2π)), which is the one-dimensional Gaussian normalization, applied to a two-dimensional Gaussian. For a normalized 2D kernel, the constant should be 1/(2πσ^2). If this is a typo, it should be corrected; if the intentionally non-normalized kernel is used, the effect on the Laplacian pyramid and on the attention recalibration should be discussed, because the DC component may not be preserved across scales.
  5. [Experiments; Tables 1 and 2] All quantitative results are reported as single numbers with no error bars, confidence intervals, or multiple runs. Given that the difference with ReconFormer on fastMRI AF=4 is only 0.02 dB in PSNR, the single-run results are insufficient to support the claim of consistent superiority. The authors should report mean ± standard deviation over at least three seeds, or provide a paired statistical test across test slices, to establish that the observed differences are not random variation.
minor comments (7)
  1. [Figure 1 caption] The caption lists '(3) We propose' but the corresponding panel is labeled '(c)'. The numbering should be consistent.
  2. [Training Details] In the sentence 'the number of groups N in SHAM', 'SHAM' should be 'SPAM'.
  3. [Table 4] The checkmark formatting in Table 4 is garbled in the text, making it difficult to see which components are present in each ablation row. Please ensure the table renders with clear ✓/✗ symbols.
  4. [Experimental Settings] The description of the 1D Cartesian undersampling mask is vague ('similar to the fastMRI challenge'). Specify the exact mask type (e.g., variable-density random, center fraction, low-frequency lines) and how masks are generated for each dataset to improve reproducibility.
  5. [Analysis of Hyper-parameters] The paper defers hyperparameter analysis to the Supplementary Materials, but the arXiv submission does not appear to include supplementary materials. Either include the analysis or summarize the key findings for the key hyperparameters (M, N, r, E, D) in the main text.
  6. [Eq. (2) and Fig. 2] The notation 'F ⋇ P' is used for the combined FMAM and SPAM but is not defined before Eq. (2). Define this operator or replace it with explicit notation to avoid confusion.
  7. [Eq. (5)–(6)] The FMAM attention computation sums softmax attention scores over heads and then applies the summed matrix to a single V derived from Fin. This is non-standard; clarify whether the heads share the V projection or whether the summed matrix is used as a single-head attention weight.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: FPS-Former's claims are empirical and self-contained; the self-citations are ancillary and not load-bearing.

full rationale

The paper's derivation chain is architectural: the FMAM, SPAM, SDFN, and HEFR modules are defined by explicit equations (3)-(13) using standard operations, and the reconstruction loss is a conventional L1 data-consistency objective. Performance claims are supported by experiments on three public datasets against external baselines, with ablations in Tables 4-5 attributing gains to the proposed components. No fitted parameter is later relabeled as a prediction, and no same-author uniqueness theorem or prior result is invoked to force the design. The only self-citations (Yang et al. 2024a in Related Work and Yang et al. 2024b in the Introduction) support generic statements about ViTs and attention mechanisms; they are not load-bearing for the central claim. The possible overstatement in the abstract and Section 5 relative to Table 1 (ReconFormer reports higher PSNR than FPS-Former on fastMRI AF=4) is an empirical-claim consistency issue, not a circularity finding. Overall, the derivation is self-contained and not circular.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The paper introduces new neural network modules but no new physical entities, forces, or dimensions. The free parameters listed are architectural hyperparameters chosen by hand or ablation. The axioms are standard assumptions about ViT behavior and image statistics, none of which are derived within the paper.

free parameters (7)
  • M (number of frequency pyramid levels in FMAM) = 3
    Chosen by ablation in 'Analysis of FPS block'; affects the range of frequency information captured.
  • N (number of token groups in SPAM) = 4
    Set in Training Details; chosen by hand, no ablation shown for this value.
  • r (channel expansion ratio in SDFN) = 2
    Set in Training Details; standard expansion ratio chosen by hand.
  • E (number of experts in HEFR) = 8
    Set in Training Details; hyperparameter chosen without extensive ablation.
  • D (dimension of weight matrix in HEFR) = 32
    Set in Training Details; hyperparameter chosen by hand.
  • (N0,N1,N2,N3,N4) block counts = (4,1,2,2,1)
    Architecture hyperparameters set in Training Details; not swept in the paper.
  • Attention heads per FPS block = (1,2,4,8)
    Architecture hyperparameters set in Training Details; chosen to match feature resolutions.
assumptions (5)
  • domain assumption The self-attention mechanism of Vision Transformers behaves as a low-pass filter, suppressing high-frequency details.
    Invoked in the Introduction and FMAM section to justify the need for frequency modulation; based on cited works (Park and Kim 2022; Wang et al. 2022).
  • domain assumption MRI images contain clusters of similar patches, so grouping tokens by hash codes captures contextually related content.
    Motivates SPAM in Eq. 7-9; this is a heuristic assumption about natural image statistics, not formally justified.
  • domain assumption Multi-scale feature representations improve image reconstruction.
    Used to justify SDFN; widely accepted in the image restoration literature, but not proven in this paper.
  • domain assumption L1 loss is an appropriate training objective for MRI reconstruction.
    Used in Eq. 1 as the sole loss; standard practice in the field.
  • domain assumption The AdamW optimizer with a fixed schedule converges to a good solution for this architecture.
    Training details rely on standard deep learning practice; no convergence analysis is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Boosting ViT-based MRI Reconstruction from the Perspectives of Frequency Modulation, Spatial Purification, and Scale Diversification." pith.science (2026). https://pith.science/paper/A7W3D3GE

@misc{pith2026241210776,
  author       = {Pith},
  title        = {Pith review of: Boosting ViT-based MRI Reconstruction from the Perspectives of Frequency Modulation, Spatial Purification, and Scale Diversification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/A7W3D3GE}},
  note         = {Machine review of arXiv:2412.10776}
}
read the original abstract

The accelerated MRI reconstruction process presents a challenging ill-posed inverse problem due to the extensive under-sampling in k-space. Recently, Vision Transformers (ViTs) have become the mainstream for this task, demonstrating substantial performance improvements. However, there are still three significant issues remain unaddressed: (1) ViTs struggle to capture high-frequency components of images, limiting their ability to detect local textures and edge information, thereby impeding MRI restoration; (2) Previous methods calculate multi-head self-attention (MSA) among both related and unrelated tokens in content, introducing noise and significantly increasing computational burden; (3) The naive feed-forward network in ViTs cannot model the multi-scale information that is important for image restoration. In this paper, we propose FPS-Former, a powerful ViT-based framework, to address these issues from the perspectives of frequency modulation, spatial purification, and scale diversification. Specifically, for issue (1), we introduce a frequency modulation attention module to enhance the self-attention map by adaptively re-calibrating the frequency information in a Laplacian pyramid. For issue (2), we customize a spatial purification attention module to capture interactions among closely related tokens, thereby reducing redundant or irrelevant feature representations. For issue (3), we propose an efficient feed-forward network based on a hybrid-scale fusion strategy. Comprehensive experiments conducted on three public datasets show that our FPS-Former outperforms state-of-the-art methods while requiring lower computational costs.

Figures

Figures reproduced from arXiv: 2412.10776 by the authors.

Figure 1
Figure 1. Our main idea. (a) The pipeline of standard ViTs [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. (a) The overall architecture of the proposed FPS-Former. Given an input image [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Qualitative comparison of different methods on (a) the single-coil dataset including CC359 and fastMRI, (b) the [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: (a) Frequency response analysis. (b) Visualization [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. DH-Mamba: Exploring Dual-domain Hierarchical State Space Models for MRI Reconstruction

    eess.IV 2025-01 conditional novelty 6.0 of 10

    DH-Mamba is a dual-domain hierarchical Mamba network that uses circular k-space scanning and local diversity enhancement to outperform prior MRI reconstruction methods on three public datasets.

  2. Continuous K-space Recovery Network with Image Guidance for Fast MRI Reconstruction

    eess.IV 2024-11 conditional novelty 6.0 of 10

    IGKR-Net recovers missing MRI k-space samples with an implicit neural representation transformer guided by image features, outperforming earlier reconstruction networks on CC359, fastMRI, IXI, and SKM-TEA benchmarks.

Reference graph

Works this paper leans on

43 extracted references · 29 canonical work pages · cited by 2 Pith papers

  1. [1]

    , " * 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 mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    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 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Aghabiglou, A. 2021. MR image reconstruction using densely connected residual convolutional networks. Computers in Biology and Medicine, 139: 105010

    work page 2021

  4. [4]

    Aghabiglou, A.; and Eksioglu, E. M. 2021. Projection-Based cascaded U-Net model for MR image reconstruction. Computer Methods and Programs in Biomedicine, 207: 106151

    work page 2021

  5. [6]

    M.; Benjdira, B.; Koubaa, A.; El-Shafai, W.; Khan, Z.; and Boulila, W

    Ali, A. M.; Benjdira, B.; Koubaa, A.; El-Shafai, W.; Khan, Z.; and Boulila, W. 2023. Vision transformers in image restoration: A survey. Sensors, 23(5): 2385

    work page 2023

  6. [7]

    Cai, H.; Li, J.; Hu, M.; Gan, C.; and Han, S. 2023. EfficientViT: Lightweight Multi-Scale Attention for High-Resolution Dense Prediction. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 17302--17313

    work page 2023

  7. [8]

    R.; Fan, Q.; and Panda, R

    Chen, C.-F. R.; Fan, Q.; and Panda, R. 2021. CrossViT: Cross-Attention Multi-Scale Vision Transformer for Image Classification. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 357--366

    2021

  8. [9]

    Chen, X.; Pan, J.; Lu, J.; Fan, Z.; and Li, H. 2023. Hybrid cnn-transformer feature fusion for single image deraining. In Proceedings of the AAAI conference on artificial intelligence, volume 37, 378--386

    work page 2023

Show all 43 references
  1. [10]

    L.; Wang, G.; Rueckert, D.; Firmin, D.; and Yang, G

    Chen, Y.; Sch \"o nlieb, C.-B.; Li \`o , P.; Leiner, T.; Dragotti, P. L.; Wang, G.; Rueckert, D.; Firmin, D.; and Yang, G. 2022. AI-based reconstruction for fast MRI—A systematic review and meta-analysis. Proceedings of the IEEE, 110(2): 224--245

    2022

  2. [11]

    D.; Schmidt, A

    Desai, A. D.; Schmidt, A. M.; Rubin, E. B.; Sandino, C. M.; Black, M. S.; Mazzoli, V.; Stevens, K. J.; Boutin, R.; R \'e , C.; Gold, G. E.; et al. 2022. Skm-tea: A dataset for accelerated mri reconstruction with dense image labels for quantitative clinical evaluation. arXiv pr...

    2022 arXiv

  3. [12]

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

    2020 arXiv

  4. [13]

    Eo, T.; Jun, Y.; Kim, T.; Jang, J.; Lee, H.-J.; and Hwang, D. 2018. KIKI-net: cross-domain convolutional neural networks for reconstructing undersampled magnetic resonance images. Magnetic resonance in medicine, 80(5): 2188--2201

    2018

  5. [14]

    Feng, C.-M.; Yang, Z.; Chen, G.; Xu, Y.; and Shao, L. 2021. Dual-octave convolution for accelerated parallel MR image reconstruction. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, 116--124

    2021

  6. [15]

    Guo, P.; Mei, Y.; Zhou, J.; Jiang, S.; and Patel, V. M. 2024. ReconFormer: Accelerated MRI Reconstruction Using Recurrent Transformer. IEEE Transactions on Medical Imaging, 43(1): 582--593

    2024

  7. [16]

    Han, K.; Wang, Y.; Chen, H.; Chen, X.; Guo, J.; Liu, Z.; Tang, Y.; Xiao, A.; Xu, C.; Xu, Y.; Yang, Z.; Zhang, Y.; and Tao, D. 2023. A Survey on Vision Transformer. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(1): 87--110

    2023

  8. [17]

    Huang, J.; Fang, Y.; Wu, Y.; Wu, H.; Gao, Z.; Li, Y.; Del Ser, J.; Xia, J.; and Yang, G. 2022. Swin transformer for fast MRI. Neurocomputing, 493: 281--304

    2022

  9. [18]

    Khan, A.; Sohail, A.; Zahoora, U.; and Qureshi, A. S. 2020. A survey of the recent architectures of deep convolutional neural networks. Artificial intelligence review, 53: 5455--5516

    2020

  10. [19]

    K.; and Akcakaya, M

    Knoll, F.; Hammernik, K.; Zhang, C.; Moeller, S.; Pock, T.; Sodickson, D. K.; and Akcakaya, M. 2020. Deep-learning methods for parallel magnetic resonance imaging reconstruction: A survey of the current approaches, trends, and issues. IEEE signal processing magazine, 37(1): 128--140

    2020

  11. [20]

    U.; Yurt, M.; \"O zbey, M.; and Cukur, T

    Korkmaz, Y.; Dar, S. U.; Yurt, M.; \"O zbey, M.; and Cukur, T. 2022. Unsupervised MRI reconstruction via zero-shot learned adversarial transformers. IEEE Transactions on Medical Imaging, 41(7): 1747--1763

    2022

  12. [21]

    Li, Y.; Zhang, K.; Cao, J.; Timofte, R.; and Van Gool, L. 2021 a . Localvit: Bringing locality to vision transformers. arXiv preprint arXiv:2104.05707

    2021 arXiv

  13. [22]

    Li, Z.; Liu, F.; Yang, W.; Peng, S.; and Zhou, J. 2021 b . A survey of convolutional neural networks: analysis, applications, and prospects. IEEE transactions on neural networks and learning systems, 33(12): 6999--7019

    2021

  14. [23]

    Lin, K.; and Heckel, R. 2022. Vision transformers enable fast and robust accelerated MRI. In International Conference on Medical Imaging with Deep Learning, 774--795. PMLR

    2022

  15. [24]

    Liu, Z.; Lin, Y.; Cao, Y.; Hu, H.; Wei, Y.; Zhang, Z.; Lin, S.; and Guo, B. 2021. Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF international conference on computer vision, 10012--10022

    2021

  16. [25]

    Park, N.; and Kim, S. 2022. How do vision transformers work? arXiv preprint arXiv:2202.06709

    2022 arXiv

  17. [26]

    Sarvamangala, D.; and Kulkarni, R. V. 2022. Convolutional neural networks in medical image understanding: a survey. Evolutionary intelligence, 15(1): 1--22

    2022

  18. [27]

    V.; Price, A

    Schlemper, J.; Caballero, J.; Hajnal, J. V.; Price, A. N.; and Rueckert, D. 2018. A deep cascade of convolutional neural networks for dynamic MR image reconstruction. IEEE transactions on Medical Imaging, 37(2): 491--503

    2018

  19. [28]

    W.; and Zhang, X

    Shen, G.; Li, M.; Anderson, S.; Farris, C. W.; and Zhang, X. 2024. Magnetic Resonance Image Processing Transformer for General Reconstruction. arXiv:2405.15098

    2024 arXiv

  20. [29]

    I.; Ong, F.; Cheng, J

    Tamir, J. I.; Ong, F.; Cheng, J. Y.; Uecker, M.; and Lustig, M. 2016. Generalized magnetic resonance image reconstruction using the Berkeley advanced reconstruction toolbox. In ISMRM Workshop on Data Sampling & Image Reconstruction, Sedona, AZ, volume 7, 8

    2016

  21. [30]

    Wang, P.; Zheng, W.; Chen, T.; and Wang, Z. 2022. Anti-oversmoothing in deep vision transformers via the fourier domain analysis: From theory to practice. arXiv preprint arXiv:2203.05962

    2022 arXiv

  22. [31]

    Wang, S.; Su, Z.; Ying, L.; Peng, X.; Zhu, S.; Liang, F.; Feng, D.; and Liang, D. 2016. Accelerating magnetic resonance imaging via deep learning. In 2016 IEEE 13th international symposium on biomedical imaging (ISBI), 514--517. IEEE

    2016

  23. [32]

    K.; Zou, K

    Warfield, S. K.; Zou, K. H.; and Wells, W. M. 2004. Simultaneous truth and performance level estimation (STAPLE): an algorithm for the validation of image segmentation. IEEE transactions on medical imaging, 23(7): 903--921

    2004

  24. [33]

    Yang, G.; Zhang, L.; Zhou, M.; Liu, A.; Chen, X.; Xiong, Z.; and Wu, F. 2022. Model-guided multi-contrast deep unfolding network for mri super-resolution reconstruction. In Proceedings of the 30th ACM International Conference on Multimedia, 3974--3982

    2022

  25. [34]

    Yang, Z.; Fu, K.; Duan, M.; Qu, L.; Wang, S.; and Song, Z. 2024 a . Separate and conquer: Decoupling co-occurrence via decomposition and representation for weakly supervised semantic segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni...

    2024

  26. [35]

    Yang, Z.; Meng, Y.; Fu, K.; Wang, S.; and Song, Z. 2024 b . Tackling Ambiguity from Perspective of Uncertainty Inference and Affinity Diversification for Weakly Supervised Semantic Segmentation. ArXiv, abs/2404.08195

    2024 arXiv

  27. [36]

    W.; Arora, A.; Khan, S.; Hayat, M.; Khan, F

    Zamir, S. W.; Arora, A.; Khan, S.; Hayat, M.; Khan, F. S.; and Yang, M.-H. 2022. Restormer: Efficient transformer for high-resolution image restoration. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 5728--5739

    2022

  28. [37]

    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

  29. [38]

    Zeng, G.; Guo, Y.; Zhan, J.; Wang, Z.; Lai, Z.; Du, X.; Qu, X.; and Guo, D. 2021. A review on deep learning MRI reconstruction without fully sampled k-space. BMC Medical Imaging, 21(1): 195

    2021

  30. [39]

    Zeng, W.; Peng, J.; Wang, S.; and Liu, Q. 2020. A comparative study of CNN-based super-resolution methods in MRI reconstruction and its beyond. Signal Processing: Image Communication, 81: 115701

    2020

  31. [40]

    Zhao, H.; Gou, Y.; Li, B.; Peng, D.; Lv, J.; and Peng, X. 2023. Comprehensive and Delicate: An Efficient Transformer for Image Restoration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 14122--14132

    2023

  32. [41]

    Zheng, M.; Xu, J.; Shen, Y.; Tian, C.; Li, J.; Fei, L.; Zong, M.; and Liu, X. 2022. Attention-based CNNs for image classification: A survey. In Journal of Physics: Conference Series, volume 2171, 012068. IOP Publishing

    2022

  33. [42]

    Zhou, B.; Dey, N.; Schlemper, J.; Salehi, S. S. M.; Liu, C.; Duncan, J. S.; and Sofka, M. 2023. DSFormer: A Dual-Domain Self-Supervised Transformer for Accelerated Multi-Contrast MRI Reconstruction. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer V...

    2023

  34. [43]

    Zhou, B.; and Zhou, S. K. 2020. DuDoRNet: learning a dual-domain recurrent network for fast MRI reconstruction with deep T1 prior. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 4273--4282

    2020

  35. [44]

    Zhou, S.; Chen, D.; Pan, J.; Shi, J.; and Yang, J. 2024. Adapt or perish: Adaptive sparse transformer with attentive feature refinement for image restoration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2952--2963

    2024

Pith tools

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