Pith. sign in

REVIEW 2 major objections 5 minor 58 references

Frequency-Aware Dual-Stream Learning for Balanced Realism and Fidelity in Electron Microscopy Imaging

T0 review · 2 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read WaveletEM claims to break the realism-versus-fidelity trade-off in electron microscopy image enhancement by splitting each image into low- and high-frequency bands before reconstructing it.

desk verdict A genuinely new dual-stream wavelet architecture with strong LPIPS and speed numbers, but the 'doubles resolution' claim rests on a metric that can be fooled by the very hallucinations the authors admit to—so the central result needs a different validation. read the letter →

arxiv 2607.22765 v1 pith:PEGAVCCH submitted 2026-07-24 eess.IV cs.CV

Longmi Gao , Zhengkai Zhao , Pan Gao , Manoranjan Paul This is my paper · ORCID
classification eess.IVcs.CV
keywords electronmicroscopyimageenhancementdiscretewavelettransformconditionaldiffusionmodeltransformersuper-resolutiondenoisingresolutionratioperceptualrealismandfidelity
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

WaveletEM claims to break the realism-versus-fidelity trade-off in electron microscopy image enhancement by separating the problem in the frequency domain. A 2D discrete wavelet transform splits each image into a low-frequency structural component and high-frequency detail subbands; a conditional diffusion model reconstructs the structure and a transformer network recovers the details, before an inverse transform reassembles the output. On the EMDiffuse mouse brain cortex test set, the method reports an almost three-fold better LPIPS than PSSR (0.0133 vs. 0.0411) and a resolution ratio of 1.8579 in denoising, interpreted as roughly doubling the recovered resolution. Because the pipeline can start from a 2-second acquisition and adds only about 2.2 seconds of inference, the authors claim a fast, high-quality imaging workflow that was not previously available. A sympathetic reader would care because the claimed speed-up and detail recovery directly address the physical bottleneck that slows electron microscopy for structural biology.

What carries the argument

The load-bearing mechanism is Frequency-Aware Structural Decomposition using the 2D discrete wavelet transform with Haar filters, which separates an image into one low-frequency subband $X_{LL}$ and three high-frequency subbands $X_{LH}$, $X_{HL}$, $X_{HH}$ at half resolution. The low-frequency branch (LFCDM) is a DDPM conditioned on $X_{LL}$ and trained with an L2 denoising objective; the high-frequency branch (HFTBM) is a residual-group transformer whose HFIRB blocks combine depthwise convolutions with spatial-window self-attention through an Interflow feature exchange, trained with L1 loss. An inverse DWT fuses the two outputs. The working hypothesis is that matching global structure is a generative problem while recovering fine detail is a deterministic regression problem, so forcing each objective onto its own frequency band avoids the spectral bias that makes single-stream networks choose one at the expense of the other.

What would settle it

Take the fast (2 s) EM acquisitions, run WaveletEM, and compare its output to ground truth from the slow (100 to 130 s) acquisition on a set of known organelle boundaries; if Resolution Ratio above 1 reflects real detail, downstream boundary segmentation or particle counting should improve, while if it reflects invented textures, those measurements should stay flat or worsen even though LPIPS improves.

Watch

Extended reading notes

Core claim

The paper's central claim is that the apparent conflict between realistic textures and pixel-accurate reconstruction in EM is not intrinsic to the problem, but is a consequence of processing all frequencies in one network. WaveletEM therefore decomposes the input with Haar wavelets, sends the low-frequency approximation to a conditional DDPM trained to synthesize globally consistent structure, and sends the three high-frequency subbands to a transformer-based detail restorer trained with L1 loss. Separate supervision with L2 on the diffusion branch and L1 on the transformer branch keeps the two objectives from interfering. On the EMDiffuse denoising task, WaveletEM reports PSNR 28.327 dB, SSIM 0.6264, LPIPS 0.0133, and a resolution ratio of 1.8579, improving on EMDiffuse's LPIPS of 0.0164 and PSSR's resolution ratio of 0.7632; in 2x super-resolution it reports a resolution ratio of 1.9259. The paper interprets these numbers as evidence that dual-stream frequency decomposition reconciles biological realism with quantitative fidelity.

Load-bearing premise

The load-bearing premise is that a Resolution Ratio above 1 (1.8579 in denoising) means the output contains finer real structure than the ground truth, rather than fine textures that the generative branch simply invented.

Editorial extensions

If this is right

  • Acquisition-plus-inference time per image drops from roughly 100 to 130 seconds to about 4.2 seconds (2 seconds acquisition plus 2.18 seconds inference), so high-quality EM screening becomes feasible at much higher throughput.
  • The dual-stream design reports a resolution ratio of 1.8579 in denoising and 1.9259 in 2x super-resolution, both well above the highest single-stream baseline, which the paper reads as evidence that recovered detail is not merely perceptual smoothing.
  • WaveletEM transfers to unseen tissue types (bone marrow, heart, liver, HeLa) with modest fine-tuning, suggesting the frequency-split strategy is not over-fit to one sample.
  • The full pipeline needs only about 2.5 seconds per image at 1000 DDPM sampling steps, roughly 3.4x faster than EMDiffuse at a comparable parameter count.
  • Ablations tie the gain to the combination: the full dual-stream configuration beats every single-branch variant on both LPIPS and resolution ratio, and Haar outperforms Daubechies and biorthogonal bases.

Reading between the lines

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

  • Because Resolution Ratio is computed as GT resolution divided by output resolution, a value above 1 means the output is judged sharper than ground truth; the paper leaves open whether that sharpness is real ultrastructure or hallucinated texture, and Section 6 explicitly concedes that the diffusion branch may introduce plausible but false ultrastructural patterns.
  • A direct untested consequence is that biological faithfulness should be validated by downstream tasks such as organelle segmentation or expert review; the authors note the public dataset lacks annotations, so this remains the decisive open check on the double-resolution claim.
  • The same wavelet-split recipe could be applied to other paired slow-fast microscopy settings (cryo-EM, low-dose TEM, or confocal) whenever a generative branch and a deterministic branch can be assigned to different frequency bands; the paper only demonstrates SEM.
  • If the resolution-ratio result is confirmed by ground-truth structure recovery, the approach suggests a general design rule: resolve realism-fidelity conflicts by separating generative and regression objectives in a frequency-transformed space rather than by weighting two losses in pixel space.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 5 minor

Summary. The paper proposes WaveletEM, a dual-stream architecture for electron microscopy image enhancement. Using a 2D discrete wavelet transform, the method decomposes an input EM image into low- and high-frequency subbands, processes the low-frequency component with a conditional diffusion model for structural realism and the high-frequency components with a transformer-based model for detail fidelity, and recombines them via the inverse transform. The authors report extensive experiments on the EMDiffuse dataset, including denoising and 2x super-resolution, claiming superior LPIPS and a resolution ratio that doubles the resolution compared to existing methods, as well as large inference-time savings. The manuscript also includes transfer-learning experiments, ablations over architecture components, wavelet bases, sampling steps, and loss terms, and states that code and data are publicly available.

Significance. The frequency-conscious architectural design is a reasonable engineering contribution: explicitly separating low-frequency structure from high-frequency detail and assigning them to heterogeneous branches is a plausible way to address the realism-fidelity trade-off in EM restoration. The experiments are extensive, the efficiency results are concrete, and the public release of code, pretrained models, and dataset access strengthens reproducibility. However, the headline quantitative claim—that WaveletEM roughly doubles resolution—rests on a resolution metric computed from the output image alone, and the reported ratio above 1.0 has exactly the signature of hallucinated high-frequency content, a failure mode the authors themselves acknowledge in Section 6. As a result, the central claim is not yet established. If that concern is resolved with additional validation, the method would be a solid and practically useful contribution to fast EM imaging workflows.

major comments (2)
  1. [Section 4.1, Eqs. (11)-(13); Table 1] The Resolution Ratio defined in Eq. (13) is GT Resolution divided by Output Resolution, so a value above 1 means the output's decorrelation-based resolution estimate is finer (smaller distance) than the ground truth's. Since Eq. (12) estimates resolution from the output image alone by locating decorrelation peaks in high-pass filtered versions, any model that adds self-consistent high-frequency texture—exactly the 'plausible but false ultrastructural patterns' conceded in Section 6—will shift the peak to higher frequencies and inflate the ratio. The reported 1.8579 therefore has the signature of hallucinated detail rather than recovered true structure, and Section 4.2's claim that WaveletEM 'doubles the resolution' is not supported. Please add evidence that the metric rejects hallucinated texture, for example by reporting cross-correlation between output and ground-truth high-frequency wavelet subbands, or by validating against an independent higher-resolution acquisition.
  2. [Section 4.1, Tables 5-7] The manuscript reports results on a single 'identical held-out test split' and does not state that a separate validation split was used for hyperparameter selection. If the ablation-driven choices (Haar wavelet basis, 1000 sampling steps, L2 for LFCDM with L1 for HFTBM) were made on this same test split, then the headline numbers in Table 1 are selected rather than independent predictions. Please specify the validation protocol, or re-run the analysis on a truly held-out split and report the resulting test metrics.
minor comments (5)
  1. [Section 3.2, Eq. (6)] The text describes 'single-step denoising' but the method uses 1000 DDPM sampling steps during inference; Eq. (6) is the standard iterative DDPM update, so the wording should be clarified to avoid confusion.
  2. [Section 3.3, Eq. (8)] The symbols e_mu and e_sigma are called 'spatial distiller' and 'channel distiller', but these terms are never defined and no reference is given; as written, Eq. (8) is not reproducible.
  3. [Table 3 and Abstract] The abstract claims 'strong generalization', but the HeLa results without fine-tuning are markedly worse (PSNR 24.23, SSIM 0.49, LPIPS 0.071) than the other datasets; the generalization claim should be qualified to acknowledge this limitation.
  4. [Section 4.2 and Table 1] The statement that WaveletEM 'doubles the resolution compared to PSSR' is imprecise: the denoising resolution ratios are 1.8579 versus 0.7632, a factor of about 2.43, so the text should say 'more than doubles' or report the exact ratio.
  5. [Tables 2, 5 and Section 4.2] Inference time is reported as 2.54 s per image in Table 2, 2.18 s per image for 1000 steps in Table 5, and 'just over two seconds' in Section 4.2; please reconcile these values and state the exact configuration used for each reported timing.

Circularity Check

1 steps flagged · score 4.0 of 10

Headline LPIPS and resolution-ratio gains are partly a test-set selection artifact: the paper chooses sampling steps, loss terms, and wavelet basis on the same Mouse Brain Cortex split used for the final comparisons, so the reported superiority is partially forced by that selection.

  1. fitted input called prediction [Section 4.1 (Experimental Setting), Section 4.4 (Ablation Study, Tables 5-7), Section 4.2 (Table 1)]
    "For a fair comparison, all competing methods were retrained from scratch on the same EMDiffuse training set using their official implementations and recommended settings, and evaluated on the identical held-out test split. ... Based on this comprehensive analysis, we selected 1000 sampling steps as the optimal balance between computational efficiency and reconstruction quality for our WaveletEM architecture."

    The ablation tables (sampling steps, loss combinations, wavelet basis) are evaluated on the Mouse Brain Cortex dataset with no separate validation split described, and the best-performing configuration is then selected. The final headline results in Table 1 are reported on the same dataset and split. Thus the hyperparameters (1000 steps, L2/L1 loss, Haar wavelet) are effectively fitted to the test set, and the LPIPS and resolution-ratio scores used as evidence of superiority are the same metrics that drove the configuration choice. This makes the reported gains partly a selection artifact rather than an independent prediction.

full rationale

This is an empirical architecture paper rather than a mathematical derivation, so most classical circularity modes do not apply. There is no load-bearing self-citation: the CDFormer baseline is co-authored by one of the present authors, but it is used as a comparison method, not as justification for the central premise. No uniqueness theorem is imported, and no ansatz is smuggled via citation. The resolution-ratio concern raised in the skeptical reading is a validity risk rather than a circularity: the decorrelation-based metric is externally defined and not fitted, but the paper itself concedes in Section 6 that the diffusion branch may introduce plausible but false ultrastructural patterns, so the metric could be inflated by hallucinated texture. That is a correctness/falsifiability caveat, not an equation that reduces to its own input. The one genuine circularity is the test-set hyperparameter selection: the paper reports no validation split for the ablations, selects the configuration that optimizes the same metrics on the same Mouse Brain Cortex data later used for the final comparisons, and then presents those metrics as evidence of superiority. This partially forces the headline LPIPS and resolution-ratio results, though the dual-stream architecture itself retains independent content.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The paper does not introduce physical entities or derive a law; its claims rest on domain assumptions about wavelet decomposition and the resolution metric, and on several hyperparameters selected by test-set performance.

free parameters (3)
  • DDPM inference sampling steps = 1000
    Selected as optimal in Table 5 based on test-set PSNR, SSIM, and LPIPS; other step counts (800, 900, 1500, 2000) were evaluated on the same test split.
  • Loss function assignment = L2 for LFCDM, L1 for HFTBM
    Chosen through Table 6 ablation on the test set; this assignment produced the best test metrics.
  • Wavelet basis = Haar
    Selected via Table 7 ablation on the test set; Daubechies and Biorthogonal bases were rejected on test metrics.
assumptions (4)
  • domain assumption Haar wavelet provides an orthogonal, non-redundant basis such that low-frequency subbands carry semantic structure and high-frequency subbands carry only texture.
    Motivated in Section 2.1 and used to justify the dual-stream design in Section 3.1; no empirical test is provided for the claim that bands can be optimized independently.
  • domain assumption Single-stream networks fail because of spectral bias (over-smoothing or hallucinating), and frequency decomposition resolves this.
    Section 1 frames the problem this way, but the paper does not isolate this effect from the benefit of having two specialized models.
  • domain assumption Decorrelation analysis (Descloux et al.) yields a valid measure of biological resolution in EM images.
    Section 4.1 adopts the Resolution Ratio as the key 'biological faithfulness' metric; the paper notes in Section 6 that image-level metrics cannot fully capture biological faithfulness.
  • standard math Standard DDPM theory (Equations 4-7) holds for the low-frequency branch.
    The forward and reverse diffusion equations are taken directly from Ho et al. [48] and used without modification.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Frequency-Aware Dual-Stream Learning for Balanced Realism and Fidelity in Electron Microscopy Imaging." pith.science (2026). https://pith.science/paper/PEGAVCCH

@misc{pith2026260722765,
  author       = {Pith},
  title        = {Pith review of: Frequency-Aware Dual-Stream Learning for Balanced Realism and Fidelity in Electron Microscopy Imaging},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PEGAVCCH}},
  note         = {Machine review of arXiv:2607.22765}
}
read the original abstract

Electron microscopy enables nanoscale cellular visualization but faces a trade-off between imaging resolution and acquisition speed. Existing learning-based methods rely on single-stream architectures that struggle to balance perceptual realism and quantitative fidelity, either over-smoothing details or generating unrealistic hallucinations. This work introduces a frequency-adaptive dual-stream architecture to resolve this conflict. Using discrete wavelet transform, we decompose images into low-frequency structures and high-frequency details, then employ a conditional diffusion model for realistic global synthesis and a transformer network for precise detail recovery. Experiments on the EMDiffuse dataset show the method achieves superior LPIPS and resolution ratio, substantially outperforming existing approaches. The method also shows strong generalization across diverse biological samples, supporting fast and reliable electron microscopy imaging for structural biology and nanotechnology applications. The source code and associated dataset are publicly available to facilitate further research.

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 33 canonical work pages

  1. [1]

    IEEE Transactions on Image Processing 25(2), 540–552 (2016) https://doi.org/10.1109/TIP.2015.2504901

    Jensen, K.H., Sigworth, F.J., Brandt, S.S.: Removal of vesicle structures from transmission electron microscope images. IEEE Transactions on Image Processing 25(2), 540–552 (2016) https://doi.org/10.1109/TIP.2015.2504901

    arXiv 2016

  2. [2]

    Nature628(8006), 47–56 (2024) https://doi.org/10.1038/ s41586-024-07198-2

    Nogales, E., Mahamid, J.: Bridging structural and cell biology with cryo- electron microscopy. Nature628(8006), 47–56 (2024) https://doi.org/10.1038/ s41586-024-07198-2

    work page 2024

  3. [3]

    Automation in Construction144, 104602 (2022) https://doi.org/10.1016/j.autcon.2022.104602

    Bangaru, S.S., Wang, C., Zhou, X., Hassan, M.: Scanning electron microscopy (sem) image segmentation for microstructure analysis of concrete using u-net convolutional neural network. Automation in Construction144, 104602 (2022) https://doi.org/10.1016/j.autcon.2022.104602

    work page arXiv 2022

  4. [4]

    Nanoscale Horizons7(12), 1427–1477 (2022) https://doi.org/10

    Botifoll, M., Pinto-Huguet, I., Arbiol, J.: Machine learning in electron microscopy for advanced nanocharacterization: current developments, available tools and future outlook. Nanoscale Horizons7(12), 1427–1477 (2022) https://doi.org/10. 1039/d2nh00377e

    work page 2022

  5. [5]

    Crystals15, 192 (2025) https://doi.org/10.3390/cryst15020192

    Gong, T., Chen, L., Wang, X., Qiu, Y., Liu, H., Yang, Z., Walther, T.: Recent developments in transmission electron microscopy for crystallographic charac- terization of strained semiconductor heterostructures. Crystals15, 192 (2025) https://doi.org/10.3390/cryst15020192

    work page doi:10.3390/cryst15020192 2025

  6. [6]

    Cell187(3), 563–584 (2024) https://doi.org/10

    McCafferty, C.L., Klumpe, S., Amaro, R.E., Kukulski, W., Collinson, L., Engel, B.D.: Integrating cellular electron microscopy with multimodal data to explore biology across space and time. Cell187(3), 563–584 (2024) https://doi.org/10. 1016/j.cell.2024.01.005

    work page 2024

  7. [7]

    Measurement Science and Technology11, 1826 (2000) https://doi.org/10.1088/0957-0233/11/12/703

    Reimer, L.: Scanning electron microscopy: Physics of image formation and micro- analysis, second edition. Measurement Science and Technology11, 1826 (2000) https://doi.org/10.1088/0957-0233/11/12/703

    work page doi:10.1088/0957-0233/11/12/703 2000

  8. [8]

    17–43 (2016)

    Inkson, B.J.: Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM) for Materials Characterization, pp. 17–43 (2016). https://doi. org/10.1016/B978-0-08-100040-3.00002-X

    work page doi:10.1016/b978-0-08-100040-3.00002-x 2016

Show all 58 references
  1. [9]

    Applied Physics Letters108(2016) https://doi.org/10.1063/1.4949011

    Zhang, B., Zhang, W., Shen, Z., Chen, Y., Li, J., Zhang, S., Zhang, Z., Wuttig, M., Mazzarello, R., Ma, E., Han, X.: Element-resolved atomic structure imaging of rocksalt gesbte phase change materials. Applied Physics Letters108(2016) https://doi.org/10.1063/1.4949011

    2016 doi

  2. [10]

    Microscopy and Microanalysis27(5), 943–995 (2021) https: //doi.org/10.1017/S1431927621012125 https://academic.oup.com/mam/article- pdf/27/5/943/47962669/mam0943.pdf 19

    Liu, J.J.: Advances and applications of atomic-resolution scanning transmission electron microscopy. Microscopy and Microanalysis27(5), 943–995 (2021) https: //doi.org/10.1017/S1431927621012125 https://academic.oup.com/mam/article- pdf/27/5/943/47962669/mam0943.pdf 19

    2021 doi

  3. [11]

    Scientific Reports8(1), 2676 (2018) https://doi.org/10.1038/ s41598-018-20377-2

    Y¨ ucelen, E., Lazi´ c, I., Bosch, E.G.T.: Phase contrast scanning transmission elec- tron microscopy imaging of light and heavy atoms at the limit of contrast and resolution. Scientific Reports8(1), 2676 (2018) https://doi.org/10.1038/ s41598-018-20377-2

    2018

  4. [12]

    Scientific Reports6(1), 25453 (2016) https://doi.org/10.1038/srep25453

    Zhang, B., Wang, X.-P., Shen, Z.-J., Li, X.-B., Wang, C.-S., Chen, Y.-J., Li, J.-X., Zhang, J.-X., Zhang, Z., Zhang, S.-B., Han, X.-D.: Vacancy structures and melting behavior in rock-salt gesbte. Scientific Reports6(1), 25453 (2016) https://doi.org/10.1038/srep25453

    2016 doi

  5. [13]

    Ultramicroscopy219, 113125 (2020) https://doi

    Kharin, A.Y.: Deep learning for scanning electron microscopy: Synthetic data for the nanoparticles detection. Ultramicroscopy219, 113125 (2020) https://doi. org/10.1016/j.ultramic.2020.113125

    2020

  6. [14]

    Nature Methods12(12), (2015) https://doi

    Huff, J.: The airyscan detector from zeiss: confocal imaging with improved signal- to-noise ratio and super-resolution. Nature Methods12(12), (2015) https://doi. org/10.1038/nmeth.f.388

    2015 doi

  7. [15]

    Nature Methods18(4), 406–416 (2021) https://doi.org/10.1038/s41592-021-01080-z

    Fang, L., Monroe, F., Novak, S.W., Kirk, L., Schiavon, C.R., Yu, S.B., Zhang, T., Wu, M., Kastner, K., Latif, A.A., Lin, Z., Shaw, A., Kubota, Y., Mendenhall, J., Zhang, Z., Pekkurnaz, G., Harris, K., Howard, J., Manor, U.: Deep learning- based point-scanning super-resolution ...

    2021 doi

  8. [16]

    Nature Biomedical Engineering 9, 1907–1919 (2025) https://doi.org/10.1038/s41551-025-01413-9

    Jiang, N., Ji, H., Guan, Z., Pan, Y., Deng, C., Guo, Y., Liu, D., Chen, T., Wang, S., Wu, Y., Yang, D., Ran, A.R., Hamzah, H., Chee, M.L., Yin, C., Thinggaard, B.S., Pedersen, F.N., Peng, Q., Quek, T.C., Goh, J.H.L., Singh, S., Abd Raof, A.S., Lee-Boey, J.W.S., Lu, Y., Huang, ...

    2025

  9. [17]

    The Lancet Digital Health7(5), 100868 (2025) https://doi.org/10.1016/j.landig.2025.02.008

    Meng, Z., Guan, Z., Yu, S., Wu, Y., Zhao, Y., Shen, J., Lim, C.C., Chen, T., Yang, D., Ran, A., He, F., Hamzah, H., Singh, S., Abd Raof, A.S., Lee-Boey, J.-W.S., Lim, S.-K., Sun, X., Ge, S., Xu, G., Su, H., Cheng, Y., Lu, F., Liao, X., Jin, H., Deng, C., Ruan, L., Zhang, C., W...

    2025

  10. [18]

    IEEE Transactions on Image Processing11(7), 790–801 (2002) https: //doi.org/10.1109/TIP.2002.800888

    Dima, A., Scholz, M., Obermayer, K.: Automatic segmentation and skeletoniza- tion of neurons from confocal microscopy images based on the 3-d wavelet transform. IEEE Transactions on Image Processing11(7), 790–801 (2002) https: //doi.org/10.1109/TIP.2002.800888

    2002

  11. [19]

    Nature Communications15(1), 4677 (2024) https://doi.org/ 10.1038/s41467-024-49125-z

    Lu, C., Chen, K., Qiu, H., Chen, X., Chen, G., Qi, X., Jiang, H.: Diffusion- based deep learning method for augmenting ultrastructural imaging and volume electron microscopy. Nature Communications15(1), 4677 (2024) https://doi.org/ 10.1038/s41467-024-49125-z

    2024 doi

  12. [20]

    IEEE Transactions on Pattern Anal- ysis and Machine Intelligence47(9), 7898–7915 (2025) https://doi.org/10.1109/ TPAMI.2025.3572584

    Wang, J., Lu, X., Bennamoun, M., Sheng, B.: Non-rigid point cloud registration via anisotropic hybrid field harmonization. IEEE Transactions on Pattern Anal- ysis and Machine Intelligence47(9), 7898–7915 (2025) https://doi.org/10.1109/ TPAMI.2025.3572584

    2025

  13. [21]

    Virtual Reality & Intelligent Hardware 7(2), 139–154 (2025) https://doi.org/10.1016/j.vrih.2024.08.005

    Azizi, A., Charalambous, P., Chrysanthou, Y.: Deepsafe:two-level deep learning approach for disaster victims detection. Virtual Reality & Intelligent Hardware 7(2), 139–154 (2025) https://doi.org/10.1016/j.vrih.2024.08.005

    2025 doi

  14. [22]

    https: //arxiv.org/abs/1806.08734

    Rahaman, N., Baratin, A., Arpit, D., Draxler, F., Lin, M., Hamprecht, F.A., Bengio, Y., Courville, A.: On the Spectral Bias of Neural Networks (2019). https: //arxiv.org/abs/1806.08734

    2019 arXiv

  15. [23]

    Society for Industrial and Applied Mathematics, ??? (1992)

    Daubechies, I.: Ten Lectures on Wavelets. Society for Industrial and Applied Mathematics, ??? (1992). https://doi.org/10.1137/1.9781611970104 . eprint: https://epubs.siam.org/doi/pdf/10.1137/1.9781611970104. https://epubs.siam.org/doi/abs/10.1137/1.9781611970104

    1992 doi

  16. [24]

    ACM Transactions on Multimedia Computing, Communica- tions, and Applications21(12), 1–23 (2025) https://doi.org/10.1145/3767732

    Wen, Y., Shen, S., Shi, W., Cao, W., Bi, L., Yang, X., Sheng, B.: A lightweight depthwise separable ConvNet with frequency-domain enhancement for retinal vessel segmentation. ACM Transactions on Multimedia Computing, Communica- tions, and Applications21(12), 1–23 (2025) https:...

    2025 doi

  17. [25]

    Mathematische Annalen69(3), 331–371 (1910) https://doi.org/10.1007/BF01456326

    Haar, A.: Zur theorie der orthogonalen funktionensysteme. Mathematische Annalen69(3), 331–371 (1910) https://doi.org/10.1007/BF01456326

    1910 doi

  18. [26]

    https://arxiv.org/abs/2407.05848

    Finder, S.E., Amoyal, R., Treister, E., Freifeld, O.: Wavelet Convolutions for Large Receptive Fields (2024). https://arxiv.org/abs/2407.05848

    2024 arXiv

  19. [27]

    Sadat, S., Buhmann, J., Bradley, D., Hilliges, O., Weber, R.M.: LiteVAE: Lightweight and Efficient Variational Autoencoders for Latent Diffusion Mod- els. arXiv. arXiv:2405.14477 (2024). https://doi.org/10.48550/arXiv.2405.14477 . http://arxiv.org/abs/2405.14477 Accessed 2024-11-18 21

    2024 doi arXiv

  20. [28]

    Phung, H., Dao, Q., Tran, A.: Wavelet Diffusion Models are fast and scalable Image Generators. arXiv. arXiv:2211.16152 (2023). https://doi.org/10.48550/ arXiv.2211.16152 . http://arxiv.org/abs/2211.16152 Accessed 2024-11-19

    2023 doi arXiv

  21. [29]

    ACM Trans

    Jiang, H., Luo, A., Fan, H., Han, S., Liu, S.: Low-light image enhancement with wavelet-based diffusion models. ACM Trans. Graph.42(6) (2023) https://doi. org/10.1145/3618373

    2023 doi

  22. [30]

    In: Proceedings of the 36th Interna- tional Conference on Neural Information Processing Systems

    Finder, S.E., Zohav, Y., Ashkenazi, M., Treister, E.: Wavelet feature maps compression for image-to-image cnns. In: Proceedings of the 36th Interna- tional Conference on Neural Information Processing Systems. NIPS ’22. Curran Associates Inc., Red Hook, NY, USA (2022)

    2022

  23. [31]

    ACM Trans

    Gal, R., Hochberg, D.C., Bermano, A., Cohen-Or, D.: Swagan: a style-based wavelet-driven generative model. ACM Trans. Graph.40(4) (2021) https://doi. org/10.1145/3450626.3459836

    2021

  24. [32]

    In: 2017 IEEE International Conference on Computer Vision (ICCV), pp

    Huang, H., He, R., Sun, Z., Tan, T.: Wavelet-srnet: A wavelet-based cnn for multi- scale face super resolution. In: 2017 IEEE International Conference on Computer Vision (ICCV), pp. 1698–1706 (2017). https://doi.org/10.1109/ICCV.2017.187

    2017 doi

  25. [33]

    IEEE transactions on pattern analysis and machine intelligence 38(2), 295–307 (2015)

    Dong, C., Loy, C.C., He, K., Tang, X.: Image super-resolution using deep convolu- tional networks. IEEE transactions on pattern analysis and machine intelligence 38(2), 295–307 (2015)

    2015

  26. [34]

    In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp

    Kim, J., Lee, J.K., Lee, K.M.: Accurate image super-resolution using very deep convolutional networks. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1646–1654 (2016)

    2016

  27. [35]

    IEEE transactions on image processing26(7), 3142–3155 (2017)

    Zhang, K., Zuo, W., Chen, Y., Meng, D., Zhang, L.: Beyond a gaussian denoiser: Residual learning of deep cnn for image denoising. IEEE transactions on image processing26(7), 3142–3155 (2017)

    2017

  28. [36]

    In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp

    Ledig, C., Theis, L., Husz´ ar, F., Caballero, J., Cunningham, A., Acosta, A., Aitken, A., Tejani, A., Totz, J., Wang, Z.,et al.: Photo-realistic single image super-resolution using a generative adversarial network. In: Proceedings of the IEEE Conference on Computer Vision and...

    2017

  29. [37]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp

    Liang, J., Cao, J., Sun, G., Zhang, K., Van Gool, L., Timofte, R.: Swinir: Image restoration using swin transformer. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 1833–1844 (2021)

    2021

  30. [38]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

    Zamir, S.W., Arora, A., Khan, S., Hayat, M., Khan, F.S., Yang, M.-H.: Restormer: Efficient transformer for high-resolution image restoration. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 5728–5739 (2022) 22

    2022

  31. [39]

    Neurocomputing479, 47–59 (2022)

    Li, H., Yang, Y., Chang, M., Chen, S., Feng, H., Xu, Z., Li, Q., Chen, Y.: Srdiff: Single image super-resolution with diffusion probabilistic models. Neurocomputing479, 47–59 (2022)

    2022

  32. [40]

    Advances in Neural Information Processing Systems35, 23593–23606 (2022)

    Kawar, B., Elad, M., Ermon, S., Song, J.: Denoising diffusion restoration models. Advances in Neural Information Processing Systems35, 23593–23606 (2022)

    2022

  33. [41]

    IEEE Transactions on Multimedia27, 6198–6210 (2025) https://doi.org/10.1109/TMM.2025.3565935

    Wen, Y., Luo, B., Shi, W., Ji, J., Cao, W., Yang, X., Sheng, B.: Sat-net: Structure-aware transformer-based attention fusion network for low-quality reti- nal funduimages enhancement. IEEE Transactions on Multimedia27, 6198–6210 (2025) https://doi.org/10.1109/TMM.2025.3565935

    2025

  34. [42]

    Computer Anima- tion and Virtual Worlds36(3), 70030 (2025) https://doi.org/10.1002/cav.70030 https://onlinelibrary.wiley.com/doi/pdf/10.1002/cav.70030

    Lin, C., Zou, C., Xu, H.: Scnet: A dual-branch network for strong noisy image denoising based on swin transformer and convnext. Computer Anima- tion and Virtual Worlds36(3), 70030 (2025) https://doi.org/10.1002/cav.70030 https://onlinelibrary.wiley.com/doi/pdf/10.1002/cav.70030

    2025 doi

  35. [43]

    ACM Transactions on Multimedia Computing, Communications, and Applications21(12), 1–21 (2025) https://doi.org/10.1145/3765745

    Wen, Y., Wang, X.-N., Tang, J., Li, P., Zhu, L., Qin, J., Yang, X., Sheng, B.: CCM-Net: Contrastive and consistent multi-task network for artifact seg- mentation and quality classification of OCTA images. ACM Transactions on Multimedia Computing, Communications, and Applicatio...

    2025 doi

  36. [44]

    Yu, Y., Li, W., Bai, L., Duan, J., Zhang, X.: Utdm: a universal transformer-based diffusion model for multi-weather-degraded images restoration. Vis. Comput. 41(6), 4269–4285 (2024) https://doi.org/10.1007/s00371-024-03659-x

    2024 doi

  37. [45]

    In: 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp

    Liu, Q., Zhuang, C., Gao, P., Qin, J.: Cdformer: When degradation prediction embraces diffusion model for blind image super-resolution. In: 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 7455–7464 (2024). https://doi.org/10.1109/CVPR52733.2024.00712

    2024

  38. [46]

    In: 2017 IEEE International Con- ference on Computer Vision (ICCV), pp

    Zhu, J.-Y., Park, T., Isola, P., Efros, A.A.: Unpaired image-to-image translation using cycle-consistent adversarial networks. In: 2017 IEEE International Con- ference on Computer Vision (ICCV), pp. 2242–2251 (2017). https://doi.org/10. 1109/ICCV.2017.244

    2017

  39. [47]

    In: Pro- ceedings of the 35th International Conference on Neural Information Processing Systems

    Dhariwal, P., Nichol, A.: Diffusion models beat gans on image synthesis. In: Pro- ceedings of the 35th International Conference on Neural Information Processing Systems. NIPS ’21. Curran Associates Inc., Red Hook, NY, USA (2021)

    2021

  40. [48]

    CoRR abs/2006.11239(2020) 2006.11239

    Ho, J., Jain, A., Abbeel, P.: Denoising diffusion probabilistic models. CoRR abs/2006.11239(2020) 2006.11239

    2020 arXiv

  41. [49]

    In: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp

    Chollet, F.: Xception: Deep learning with depthwise separable convolutions. In: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1800–1807 (2017). https://doi.org/10.1109/CVPR.2017.195 23

    2017 doi

  42. [50]

    https://arxiv.org/abs/2308

    Chen, Z., Zhang, Y., Gu, J., Kong, L., Yang, X., Yu, F.: Dual Aggrega- tion Transformer for Image Super-Resolution (2023). https://arxiv.org/abs/2308. 03364

    2023

  43. [51]

    https: //arxiv.org/abs/1412.6980

    Kingma, D.P., Ba, J.: Adam: A Method for Stochastic Optimization (2017). https: //arxiv.org/abs/1412.6980

    2017 arXiv

  44. [52]

    https://doi.org/10

    Chixiang, L., Xiaojuan, Q., Haibo, J.: EMDiffuse Dataset. https://doi.org/10. 5281/zenodo.10205819 . https://doi.org/10.5281/zenodo.10205819

    doi

  45. [53]

    IEEE Transactions on Pattern Analysis and Machine Intelligence45(4), 4713–4726 (2023) https://doi.org/10.1109/TPAMI

    Saharia, C., Ho, J., Chan, W., Salimans, T., Fleet, D.J., Norouzi, M.: Image super- resolution via iterative refinement. IEEE Transactions on Pattern Analysis and Machine Intelligence45(4), 4713–4726 (2023) https://doi.org/10.1109/TPAMI. 2022.3204461

    2023

  46. [54]

    IEEE Transactions on Image Processing13(4), 600–612 (2004) https://doi.org/10.1109/TIP.2003.819861

    Wang, Z., Bovik, A.C., Sheikh, H.R., Simoncelli, E.P.: Image quality assess- ment: from error visibility to structural similarity. IEEE Transactions on Image Processing13(4), 600–612 (2004) https://doi.org/10.1109/TIP.2003.819861

    2004

  47. [55]

    https://arxiv.org/ abs/1801.03924

    Zhang, R., Isola, P., Efros, A.A., Shechtman, E., Wang, O.: The Unreasonable Effectiveness of Deep Features as a Perceptual Metric (2018). https://arxiv.org/ abs/1801.03924

    2018 arXiv

  48. [56]

    IEEE Transactions on Image Processing20(8), 2378– 2386 (2011) https://doi.org/10.1109/TIP.2011.2109730

    Zhang, L., Zhang, L., Mou, X., Zhang, D.: Fsim: A feature similarity index for image quality assessment. IEEE Transactions on Image Processing20(8), 2378– 2386 (2011) https://doi.org/10.1109/TIP.2011.2109730

    2011

  49. [57]

    Nature Methods16(9), 918–924 (2019) https://doi.org/10.1038/s41592-019-0515-7

    Descloux, A., Grußmayer, K.S., Radenovic, A.: Parameter-free image resolu- tion estimation based on decorrelation analysis. Nature Methods16(9), 918–924 (2019) https://doi.org/10.1038/s41592-019-0515-7

    2019 doi

  50. [58]

    Medical Image Analysis61, 101654 (2020) https://doi.org/10.1016/j.media.2020.101654 24

    Shen, Y., Sheng, B., Fang, R., Li, H., Dai, L., Stolte, S., Qin, J., Jia, W., Shen, D.: Domain-invariant interpretable fundus image quality assessment. Medical Image Analysis61, 101654 (2020) https://doi.org/10.1016/j.media.2020.101654 24

    2020

Pith tools

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