Pith. sign in

REVIEW 3 major objections 5 minor 40 references

Self-supervised Deep Learning for Denoising in Ultrasound Microvascular Imaging

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read A self-supervised framework trains on paired half-angle ultrasound RF frames and, without any clean labels, denoises unseen UMI data enough to exceed 15 dB gains in CNR and SNR.

desk verdict A genuinely useful pairing idea for self-supervised RF denoising in UMI, but the N2N justification is load-bearing and unproven, and the validation is single-shot. read the letter →

arxiv 2507.05451 v1 pith:M33PCCSE submitted 2025-07-07 eess.IV cs.CVeess.SP

Lijie Huang , Jingyi Yin , Jingke Zhang , U-Wai Lok , Ryan M. DeRuiter , Jieyang Jin , Kate M. Knoll , Kendra E. Petersen
show 10 more authors
James D. Krier Xiang-yang Zhu Gina K. Hesley Kathryn A. Robinson Andrew J. Bentall Thomas D. Atwell Andrew D. Rule Lilach O. Lerman Shigao Chen Chengwu Huang
This is my paper · ORCID
classification eess.IVcs.CVeess.SP
keywords self-superviseddenoisingultrasoundmicrovascularimagingplane-wavecompoundingradio-frequencyNoise2NoisepowerDopplercolorSVDclutterfiltering
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

HA2HA is a self-supervised denoising framework for ultrasound microvascular imaging (UMI). It builds training pairs by splitting the steered plane-wave angles of beamformed radio-frequency (RF) blood-flow data into odd and even half-angle groups, compounding and clutter-filtering each group separately, so that the vascular signal is shared while noise differs. The paper claims that a network trained on just 90 in-vivo contrast-free pig kidney frames with a Noise2Noise-style loss removes enough noise in unseen contrast-free and contrast-enhanced pig kidney, human liver, and human kidney datasets to raise contrast-to-noise ratio (CNR) and signal-to-noise ratio (SNR) by more than 15 dB and to lower background noise power below that of the compared methods. Because denoising is done directly on RF data rather than on envelope or image-domain data, the cleaned signal also improves downstream color Doppler imaging by suppressing spurious velocity estimates in vessel-free regions. If correct, this gives a label-free, generalizable route to higher-quality microvascular imaging without contrast agents or clean ground truth.

What carries the argument

The load-bearing object is the paired-input construction: for each acquisition, the steering angles are split into odd and even groups; each group is coherently compounded and SVD clutter-filtered to produce two RF volumes that share the same vascular signal but carry independent noise. The network is trained with a composite Noise2Noise loss made of a forward term, a reverse term, and a consistency term, with an L1 weight penalty. The theoretical engine is the Noise2Noise identity that a zero-mean, signal-independent noise leaves the conditional expectation of one noisy observation equal to the clean signal, which is what makes a clean-target-free regression valid. The encoder-decoder backbone processes full RF frames at inference, so the denoised RF signal retains phase information needed for color Doppler imaging.

What would settle it

Train or test HA2HA on simulated plane-wave RF data with a known ground-truth vascular signal and controlled additive electronic noise plus angle-dependent sidelobe interference; if the network output systematically deviates from the known clean signal by a structured, angle-correlated component rather than by zero-mean residual, then the conditional-zero-mean assumption is violated and the denoised images are biased.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central claim is that complementary angular subsets of a plane-wave acquisition provide statistically valid noisy pairs for self-supervised denoising of UMI blood-flow RF data. Because the same vessel geometry appears in both half-angle compounds but the electronic noise, angle-dependent sidelobes, and residual clutter differ, the network can learn the mapping from one noisy observation to the other and thereby converge to the underlying clean vascular signal. The authors demonstrate empirically that a model trained solely on 90 contrast-free pig kidney frames transfers without fine-tuning to contrast-enhanced pig kidney and to human liver and kidney data, producing power Doppler images with more than 15 dB higher CNR and SNR and the lowest background noise power among conventional, angular-processing, and spatiotemporal non-local means baselines, while HA2HA-denoised RF data also yields cleaner color Doppler velocity maps.

Load-bearing premise

The whole method depends on the assumption that the two half-angle observations are the true vascular signal plus independent noise that averages to zero, so that one observation is an unbiased guide to the other.

Editorial extensions

If this is right

  • A single HA2HA model, trained once on contrast-free pig kidney RF data, can be applied to unseen contrast-enhanced and human datasets without fine-tuning, so no labeled or clean reference data is needed for new sites.
  • Denoising at the RF level rather than at the envelope or image level preserves phase information, so downstream Doppler processing such as color Doppler imaging inherits the noise suppression.
  • The method's gains exceed 15 dB in CNR and SNR on the tested pig kidney, human liver, and human kidney volumes, with background noise power lower than the conventional, angular processing, and ST-NLM baselines.
  • Because the training pairs come from a single angular sweep rather than temporally adjacent frames, the approach avoids motion-related misalignment between noisy pairs.
  • At the lowest tested transmit energy (DC = 0.1), the method no longer outperforms angular processing, indicating a practical operating range for the self-supervised strategy.

Reading between the lines

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

  • If the independence and zero-mean noise assumption degrades with stronger angle-dependent sidelobes or tissue motion, the learned target would be a biased version of the true vascular signal; a useful test would be to run HA2HA on simulated plane-wave RF with known ground truth and measure residual bias.
  • Because training used a single scanner and probe configuration, the reported cross-dataset generalization leaves open whether the same half-angle construction transfers to other frequencies, probes, or imaging protocols; a multi-scanner study would settle this.
  • The same odd/even angle-splitting recipe could be applied to raw channel data or IQ data rather than beamformed RF, potentially enabling system-level denoising before beamforming and reducing computational cost.
  • If the 15 dB gains reproduce in controlled phantom or clinical studies, the method could enable reduced acoustic output in contrast-free microvascular imaging, at least down to the transmit-energy regime where it still beats the baselines.
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

3 major / 5 minor

Summary. The paper proposes HA2HA, a self-supervised deep learning framework for denoising ultrasound microvascular imaging (UMI) radio-frequency (RF) data. Training pairs are formed by splitting the plane-wave steering angles into odd and even subsets, independently compounding and SVD-clutter-filtering each subset, and using one subset as input and the other as target in a Noise2Noise-style loss with symmetric, consistency, and L1 regularization terms. A single model trained on 90 frames of contrast-free pig kidney RF data is tested on contrast-free and contrast-enhanced pig kidney, human liver, and human kidney with CKD, as well as on pig kidney data acquired at varying transmit duty cycles. The paper reports improvements exceeding 15 dB in CNR and SNR over conventional, angular processing (AP), and spatiotemporal non-local means (ST-NLM) methods, and shows qualitative improvements in color Doppler imaging derived from the denoised RF signals.

Significance. If the reported results are robust, the work is significant: it offers a label-free, RF-domain denoising method that generalizes across species, contrast conditions, and anatomical regions, with potential clinical translation. The use of complementary angular subset pairs is a novel adaptation of Noise2Noise for UMI, and the demonstration of downstream color Doppler improvement is a useful addition. The evaluation is broad, and the supplementary materials include videos and figures that support the qualitative findings. However, the central quantitative claims rest on single measurements without uncertainty quantification, and the theoretical justification relies on a noise-independence assumption that is likely violated by the proposed data construction. These issues need to be addressed before the conclusions can be fully relied upon.

major comments (3)
  1. [Section II-B and II-F, Eq. (3)] The central assumption E[Y2|Y1] = X is not established for the proposed pairing scheme. In the construction, Y1 and Y2 are formed by coherently compounding disjoint angle subsets and independently applying SVD clutter filters. Coherent half-angle compounding changes the point-spread function and sidelobe structure, and independent SVD filtering can select different subspaces in each subset; hence the clean signal components in Y1 and Y2 are not identical X, and the structured interference (sidelobes, clutter variations) counted as "noise" in Section II-B is signal-dependent and not zero-mean conditional on Y1. Under MAE optimization, the network output tends toward the conditional median of Y2 given Y1, which may differ from X by a flow-dependent bias. The reported >15 dB gains may therefore partly reflect suppression of half-angle artifacts rather than unbiased denoising. To substantiate the claim, I recommend a simulation or phantom study with known ground truth X, using the same odd/even pairing, to measure the bias of the HA2HA output relative to X; additionally, comparing a network trained with the proposed pairs against one trained with the full-angle compounded data as the target would help separate artifact suppression from denoising.
  2. [Tables I and II, Section III] All quantitative results are single measurements per dataset and condition. For each dataset, the power Doppler image appears to be computed from a single acquisition, with manually selected ROIs (Fig. 7d1 and Supplementary Fig. S3) and per-method dynamic-range optimization. No error bars, repeated acquisitions, or statistical tests are reported for CNR, SNR, or BNP. The claim of an "improvement exceeding 15 dB" and "consistently the best" performance is therefore not substantiated with uncertainty. The authors should report mean ± standard deviation over multiple frames, subjects, or independent ROI selections, and provide a statistical comparison (e.g., paired tests or confidence intervals) against the best baseline.
  3. [Section III-E and Fig. 8] The duty-cycle experiment is presented as a systematic SNR variation, but it is unclear how many independent measurements underlie each curve in Fig. 8. The text states that the four DC levels were acquired sequentially within each frame to share the same imaging section and motion status, which is a repeated-measures design, but it is not stated whether the quantitative values are computed from a single frame, an average over frames, or a summary over multiple acquisitions. Without this information and without variance estimates, the conclusion that HA2HA is best for DC ≥ 0.2 and that AP outperforms at DC = 0.1 is not reliable. Please specify the number of independent samples, report variability, and justify the absence of error bars.
minor comments (5)
  1. [Section II-F] The text says the training data construction is "described in detail in Section 2.6," but the manuscript uses Roman numeral section numbering; this should be Section II-F.
  2. [Eq. (4)] The normalization factor (2 + λc) in the combined loss is not explained; please clarify why this particular denominator is chosen and how the three loss terms are weighted relative to each other.
  3. [Section III-B] The phrase "The single frame MB image is shown" should be rephrased as "The single-frame blood flow image is shown" for grammatical clarity.
  4. [Fig. 7 and Section II-H] The ROI definitions used for the quantitative metrics appear only in the Fig. 7 caption and Supplementary Fig. S3; they should be stated in the main text near Section II-H, since they are essential for interpreting the reported CNR, SNR, and BNP values.
  5. [Section III-F] The claim that HA2HA effectively suppresses spurious velocity artifacts in color Doppler is supported only by visual inspection and line profiles; consider adding a quantitative metric, such as velocity variance or the fraction of colored pixels in a noise-only region, to the CDI evaluation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the N2N target is an explicit assumption and reported gains are held-out measurements.

full rationale

The paper's derivation chain is a standard Noise2Noise argument: paired observations Y1 and Y2 are constructed from complementary half-angle subsets, and Eq. (3) assumes E[Y2|Y1]=X. That equation is an explicit modeling assumption, not a consequence of the method's own definitions, and it is not derived from the metrics being predicted. The central claims (>15 dB CNR/SNR gains, lowest BNP) are evaluated on held-out datasets rather than fitted to those metrics. The only tuned quantities are hyperparameters lambda_c and lambda_1, which are not the predicted outputs. Self-citations to the authors' AP and ST-NLM baselines are used as comparison methods, not as load-bearing justification for the proposed framework. Even if Eq. (3) is physically questionable (since 'noise' includes structured, potentially signal-dependent sidelobes and clutter variations), that is a validity concern, not circularity. No step reduces by construction to its own inputs.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claim rests on the N2N conditional-independence assumption applied to half-angle subsets, on the assumption that SVD clutter filtering leaves the vascular signal consistent across subsets, and on empirically tuned loss weights. No new physical entities are introduced.

free parameters (2)
  • lambda_c (consistency loss weight) = 0.5
    Selected based on empirical tuning reported in Supplementary Fig. S1; controls the strength of the consistency loss between the two network outputs.
  • lambda_1 (L1 regularization weight) = 1e-5
    Selected based on empirical performance in Supplementary Fig. S2; controls the sparsity regularization on network parameters.
assumptions (4)
  • domain assumption Noise components N1 and N2 across half-angle subsets are independent and zero-mean conditional on clean signal X.
    Invoked in Eq. (3) and Section II-B; this is the core Noise2Noise premise and is not empirically verified for SVD-filtered half-angle RF data.
  • domain assumption SVD clutter filtering removes tissue but preserves vascular signals identically in both odd and even angle subsets.
    Used in Section II-F; if clutter filtering behaves differently across subsets, X differs between Y1 and Y2.
  • domain assumption Vascular RF signals are consistent across odd/even steering-angle subsets; only noise varies.
    Stated in Section II-B; actual angular dependence of flow echoes, sidelobes, and grating artifacts may violate this.
  • standard math A 2D U-Net with MAE objectives can approximate the optimal denoising map for this data distribution.
    Standard deep learning approximation assumption; unproven for this specific data distribution but consistent with common practice.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Self-supervised Deep Learning for Denoising in Ultrasound Microvascular Imaging." pith.science (2026). https://pith.science/paper/M33PCCSE

@misc{pith2026250705451,
  author       = {Pith},
  title        = {Pith review of: Self-supervised Deep Learning for Denoising in Ultrasound Microvascular Imaging},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/M33PCCSE}},
  note         = {Machine review of arXiv:2507.05451}
}
read the original abstract

Ultrasound microvascular imaging (UMI) is often hindered by low signal-to-noise ratio (SNR), especially in contrast-free or deep tissue scenarios, which impairs subsequent vascular quantification and reliable disease diagnosis. To address this challenge, we propose Half-Angle-to-Half-Angle (HA2HA), a self-supervised denoising framework specifically designed for UMI. HA2HA constructs training pairs from complementary angular subsets of beamformed radio-frequency (RF) blood flow data, across which vascular signals remain consistent while noise varies. HA2HA was trained using in-vivo contrast-free pig kidney data and validated across diverse datasets, including contrast-free and contrast-enhanced data from pig kidneys, as well as human liver and kidney. An improvement exceeding 15 dB in both contrast-to-noise ratio (CNR) and SNR was observed, indicating a substantial enhancement in image quality. In addition to power Doppler imaging, denoising directly in the RF domain is also beneficial for other downstream processing such as color Doppler imaging (CDI). CDI results of human liver derived from the HA2HA-denoised signals exhibited improved microvascular flow visualization, with a suppressed noisy background. HA2HA offers a label-free, generalizable, and clinically applicable solution for robust vascular imaging in both contrast-free and contrast-enhanced UMI.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 27 canonical work pages

  1. [1]

    Spatiotemporal clutter filtering of ultrafast ultrasound data highly increases Doppler and fUltrasound sensitivity,

    C. Demene et al., "Spatiotemporal clutter filtering of ultrafast ultrasound data highly increases Doppler and fUltrasound sensitivity," IEEE Trans. Med. Imaging, vol. 34, no. 11, pp. 2271-2285, Nov 2015, doi: 10.1109/tmi.2015.2428634

    work page arXiv 2015

  2. [2]

    Ultrasound small vessel imaging with block-Wise adaptive local clutter filtering,

    P. Song, A. Manduca, J. D. Trzasko, and S. Chen, "Ultrasound small vessel imaging with block-Wise adaptive local clutter filtering," IEEE Trans. Med. Imaging, vol. 36, no. 1, pp. 251 -262, Jan 2017, doi: 10.1109/tmi.2016.2605819

    work page arXiv 2017

  3. [3]

    Ultrafast ultrasound localization microscopy for deep super-resolution vascular imaging,

    C. Errico et al., "Ultrafast ultrasound localization microscopy for deep super-resolution vascular imaging," Nature, vol. 527, no. 7579, pp. 499-502, Nov 26 2015, doi: 10.1038/nature16066

    doi:10.1038/nature16066 2015

  4. [4]

    In Vivo Acoustic Super -Resolution and Super - Resolved Velocity Mapping Using Microbubbles,

    K. Christensen-Jeffries, R. J. Browning, M.-X. Tang, C. Dunsby, and R. J. Eckersley, "In Vivo Acoustic Super -Resolution and Super - Resolved Velocity Mapping Using Microbubbles," IEEE Trans. Med. Imaging, vol. 34, no. 2, pp. 433 -440, Feb 2015, doi: 10.1109/tmi.2014.2359650

    arXiv 2015

  5. [5]

    Deep Power-Aware Tunable Weighting for Ultrasound Microvascular Imaging,

    H. Lan et al., "Deep Power-Aware Tunable Weighting for Ultrasound Microvascular Imaging," IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 71, no. 12: Breaking the Resolution Barrier in Ultrasound, pp. 1701 -1713, 2024, doi: 10.1109/TUFFC.2024.3488729

    work page arXiv 2024

  6. [6]

    Improved ultrafast power Doppler imaging using united spatial-angular adaptive scaling wiener postfilter,

    Y. Wang et al., "Improved ultrafast power Doppler imaging using united spatial-angular adaptive scaling wiener postfilter," IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 70, no. 9, pp. 1118 -1134, Sept. 2023, doi: 10.1109/TUFFC.2023.3297571

    work page arXiv 2023

  7. [7]

    Fast 3D Super -Resolution Ultrasound With Adaptive Weight-Based Beamforming,

    J. Yan et al., "Fast 3D Super -Resolution Ultrasound With Adaptive Weight-Based Beamforming," IEEE. Trans. Biomed. Eng., pp. 1 -10, 2023, doi: 10.1109/tbme.2023.3263369

    arXiv 2023

  8. [8]

    Super-resolution imaging using conventional and non-conventional beamforming,

    F. Santibanez et al., "Super-resolution imaging using conventional and non-conventional beamforming," J. Acoust. Soc. Am., vol. 151, pp. A53-A53, 2022

    work page 2022

Show all 40 references
  1. [9]

    ASAP: Super -contrast vasculature imaging using coherence analysis and high frame -rate contrast enhanced ultrasound,

    A. Stanziola, C. H. Leow, E. Bazigou, P. D. Weinberg, and M. -X. Tang, "ASAP: Super -contrast vasculature imaging using coherence analysis and high frame -rate contrast enhanced ultrasound," IEEE Trans. Med. Imaging, vol. 37, no. 8, pp. 1847 -1856, Aug 2018, doi: 10.1109/tmi.2...

    2018

  2. [10]

    Ultrafast Power Doppler Imaging Using Frame -Multiply-and-Sum-Based Nonlinear Compounding,

    J. Kang, D. Go, I. Song, and Y. Yoo, "Ultrafast Power Doppler Imaging Using Frame -Multiply-and-Sum-Based Nonlinear Compounding," IEEE Trans. Ultrason., Ferroelectr., Freq. Control, Article vol. 68, no. 3, pp. 453 -464, Mar 2021, doi: 10.1109/tuffc.2020.3011708

    2021

  3. [11]

    Adaptive noise reduction for power Doppler imaging using SVD filtering in the channel domain and coherence weighting of pixels,

    B. Pialot et al., "Adaptive noise reduction for power Doppler imaging using SVD filtering in the channel domain and coherence weighting of pixels," Phys Med Biol, vol. 68, no. 2, Jan 2 2023, doi: 10.1088/1361-6560/acac5d

    2023 doi

  4. [12]

    Blood Flow Imaging in the Neonatal Brain Using Angular Coherence Power Doppler,

    M. Jakovljevic et al., "Blood Flow Imaging in the Neonatal Brain Using Angular Coherence Power Doppler," IEEE Trans Ultrason Fig. 10. Intensity and velocity distribution profiles corresponding to the two vessel locations marked in Fig. 9 . (a) Intensity and (b) velocity profil...

  5. [13]

    High-quality ultrafast power Doppler imaging based on spatial angular coherence factor,

    L. Huang et al., "High-quality ultrafast power Doppler imaging based on spatial angular coherence factor," IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 70, no. 5, pp. 378 -392, 2023, doi: 10.1109/TUFFC.2023.3253257

    2023

  6. [14]

    Simultaneous noise suppression and incoherent artifact reduction in ultrafast ultrasound vascular imaging,

    C. Huang et al., "Simultaneous noise suppression and incoherent artifact reduction in ultrafast ultrasound vascular imaging," IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 68, no. 6, pp. 2075- 2085, 2021, doi: 10.1109/TUFFC.2021.3055498

    2021

  7. [15]

    Randomized spatial downsampling -based Cauchy - RPCA clutter filtering for high -resolution ultrafast ultrasound microvasculature imaging and functional imaging,

    Y. Sui et al., "Randomized spatial downsampling -based Cauchy - RPCA clutter filtering for high -resolution ultrafast ultrasound microvasculature imaging and functional imaging," IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 69, no. 8, pp. 2425 -2436, 2022, doi: 10....

    2022

  8. [16]

    Multidimensional Clutter Filtering of Aperture Domain Data for Improved Blood Flow Sensitivity,

    K. A. Ozgun and B. C. Byram, "Multidimensional Clutter Filtering of Aperture Domain Data for Improved Blood Flow Sensitivity," IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 68, no. 8, pp. 2645- 2656, Apr 14 2021, doi: 10.1109/TUFFC.2021.3073292

    2021

  9. [17]

    Spatially segmented SVD clutter filtering in cardiac blood flow imaging with diverging waves,

    E. Jafarzadeh, C. E. M. Démoré, P. N. Burns, and D. E. Goertz, "Spatially segmented SVD clutter filtering in cardiac blood flow imaging with diverging waves," Ultrasonics, vol. 132, p. 107006, 2023/07/01/ 2023, doi: https://doi.org/10.1016/j.ultras.2023.107006

    2023

  10. [18]

    Ultrafast Online Clutter Filtering for Ultrasound Microvascular Imaging,

    Y. Chen, B. Fang, H. Li, L. Huang, and J. Luo, "Ultrafast Online Clutter Filtering for Ultrasound Microvascular Imaging," IEEE Trans. Med. Imaging, vol. 44, no. 6, pp. 2477 -2491, June 2025, doi: 10.1109/TMI.2025.3535550

    2025

  11. [19]

    Real -time Adaptive and Localized Spatiotemporal Clutter Filtering for Ultrasound Small Vessel Imaging,

    C. Huang, U. -W. Lok, J. Zhang, H. Liu, and S. Chen, "Real -time Adaptive and Localized Spatiotemporal Clutter Filtering for Ultrasound Small Vessel Imaging," arXiv:2405.11105. [Online]. Available: https://arxiv.org/abs/2405.11105

    arXiv

  12. [20]

    Improved ultrafast power Doppler imaging by using spatiotemporal non -local means filtering,

    L. Huang et al., "Improved ultrafast power Doppler imaging by using spatiotemporal non -local means filtering," IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 69, no. 5, pp. 1610 -1624, Mar 10 2022, doi: 10.1109/TUFFC.2022.3158611

    2022

  13. [21]

    Non -Local Based Denoising Framework for In Vivo Contrast -Free Ultrasound Microvessel Imaging,

    S. Adabi, S. Ghavami, M. Fatemi, and A. Alizad, "Non -Local Based Denoising Framework for In Vivo Contrast -Free Ultrasound Microvessel Imaging," Sensors (Basel), vol. 19, no. 2, Jan 10 2019, doi: 10.3390/s19020245

    2019 doi

  14. [22]

    In Vivo Ultrasound Localization Microscopy Imaging of the Kidney's Microvasculature With Block -Matching 3 -D Denoising,

    S. Lei et al., "In Vivo Ultrasound Localization Microscopy Imaging of the Kidney's Microvasculature With Block -Matching 3 -D Denoising," IEEE Trans Ultrason Ferroelectr Freq Control, vol. 69, no. 2, pp. 523-533, Feb 2022, doi: 10.1109/TUFFC.2021.3125010

    2022

  15. [23]

    Background Removal and Vessel Filtering of Noncontrast Ultrasound Images of Microvasculature,

    M. Bayat, M. Fatemi, and A. Alizad, "Background Removal and Vessel Filtering of Noncontrast Ultrasound Images of Microvasculature," IEEE. Trans. Biomed. Eng., vol. 66, no. 3, pp. 831-842, Mar 2019, doi: 10.1109/TBME.2018.2858205

    2019

  16. [24]

    Noise equalization for ultrafast plane wave microvessel imaging,

    P. Song, A. Manduca, J. D. Trzasko, and S. Chen, "Noise equalization for ultrafast plane wave microvessel imaging," IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 64, no. 11, pp. 1776-1781, Nov 2017, doi: 10.1109/TUFFC.2017.2748387

    2017

  17. [25]

    Noise2void-learning denoising from single noisy images,

    A. Krull, T.-O. Buchholz, and F. Jug, "Noise2void-learning denoising from single noisy images," in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 2129-2137

    2019

  18. [26]

    Self2Self With Dropout: Learning Self -Supervised Denoising From Single Image,

    Y. a. C. Quan, Mingqin and Pang, Tongyao and Ji, Hui, "Self2Self With Dropout: Learning Self -Supervised Denoising From Single Image," in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2020

    2020

  19. [27]

    Noise2self: Blind denoising by self - supervision,

    J. Batson and L. Royer, "Noise2self: Blind denoising by self - supervision," in International conference on machine learning, 2019: PMLR, pp. 524-533

    2019

  20. [28]

    Attention U -Net: Learning Where to Look for the Pancreas,

    O. Oktay et al., "Attention U -Net: Learning Where to Look for the Pancreas," doi: https://arxiv.org/abs/1804.03999

    arXiv

  21. [29]

    Self - Supervised Deep Unrolled Reconstruction Using Regularization by Denoising,

    P. Huang, C. Zhang, X. Zhang, X. Li, L. Dong, and L. Ying, "Self - Supervised Deep Unrolled Reconstruction Using Regularization by Denoising," IEEE Trans. Med. Imaging, vol. 43, no. 3, pp. 1203-1213, Mar 2024, doi: 10.1109/TMI.2023.3332614

    2024

  22. [30]

    Emulating Low-Dose PCCT Image Pairs With Independent Noise for Self-Supervised Spectral Image Denoising,

    S. Wang, Y. Yang, G. M. Stevens, Z. Yin, and A. S. Wang, "Emulating Low-Dose PCCT Image Pairs With Independent Noise for Self-Supervised Spectral Image Denoising," IEEE Trans. Med. Imaging, vol. 44, no. 1, pp. 530 -542, Jan 2025, doi: 10.1109/TMI.2024.3449817

    2025

  23. [31]

    Consensus Neural Network for Medical Imaging Denoising with Only Noisy Training Samples,

    D. Wu, K. Gong, K. Kim, X. Li, and Q. Li, "Consensus Neural Network for Medical Imaging Denoising with Only Noisy Training Samples," presented at the Medical Image Computing and Computer Assisted Intervention – MICCAI 2019: 22nd International Conference, Shenzhen, China, Octob...

    2019 doi

  24. [32]

    Self -inspired learning for denoising live -cell super - resolution microscopy,

    L. Qu et al., "Self -inspired learning for denoising live -cell super - resolution microscopy," Nat. Methods, vol. 21, no. 10, pp. 1895-1908, 2024, doi: 10.1038/s41592-024-02400-9

    1908 doi

  25. [33]

    Weakly Supervised Deep Learning -Based Optical Coherence Tomography Angiography,

    Z. Jiang et al., "Weakly Supervised Deep Learning -Based Optical Coherence Tomography Angiography," IEEE Trans. Med. Imaging, vol. 40, no. 2, pp. 688 -698, Feb 2021, doi: 10.1109/TMI.2020.3035154

    2021

  26. [34]

    High sensitivity photoacoustic imaging by learning from noisy data,

    X. Tang et al., "High sensitivity photoacoustic imaging by learning from noisy data," IEEE Trans. Med. Imaging, vol. PP, Mar 19 2025, doi: 10.1109/TMI.2025.3552692

    2025

  27. [35]

    Self-supervised PET Denoising,

    S. Y. Yie, S. K. Kang, D. Hwang, and J. S. Lee, "Self-supervised PET Denoising," Nuclear Medicine and Molecular Imaging, vol. 54, no. 6, pp. 299-304, 2020/12/01 2020, doi: 10.1007/s13139-020-00667-2

    2020 doi

  28. [36]

    Speckle2Speckle: Unsupervised Learning of Ultrasound Speckle Filtering Without Clean Data

    C. H. Rüdiger Göbl, Nassir Navab, "Speckle2Speckle: Unsupervised Learning of Ultrasound Speckle Filtering Without Clean Data." [Online]. Available: https://arxiv.org/abs/2208.00402

    arXiv

  29. [37]

    IA-Noise2Noise: An Image Alignment Strategy for Echocardiography Despeckling,

    Y. Li, W. Lu, P. Monkam, and Y. Wang, "IA-Noise2Noise: An Image Alignment Strategy for Echocardiography Despeckling," presented at the 2023 IEEE International Ultrasonics Symposium (IUS), Montreal, QC, Canada, 2023

    2023

  30. [38]

    Deep Ultrasound Denoising Without Clean Data,

    S. Goudarzi and H. Rivaz, "Deep Ultrasound Denoising Without Clean Data," arXiv:2205.03425. [Online]. Available: https://arxiv.org/abs/2205.03425

    arXiv

  31. [39]

    Decoupled Weight Decay Regularization

    I. L. a. F. Hutter, "Decoupled Weight Decay Regularization." [Online]. Available: https://arxiv.org/abs/1711.05101

    arXiv

  32. [40]

    Contextual Transformer Networks for Visual Recognition,

    Y. Li, T. Yao, Y. Pan, and T. Mei, "Contextual Transformer Networks for Visual Recognition," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 45, no. 2, pp. 1489 -1500, 2023, doi: 10.1109/TPAMI.2022.3164083

    2023

Pith tools

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