Pith. sign in

REVIEW 4 major objections 5 minor 61 references

Pre-trained Under Noise: A Framework for Robust Bone Fracture Detection in Medical Imaging

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

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

Pith's one-line read This paper claims that under simulated real-world X-ray noise, VGG16 is the only pre-trained model among ResNet50, EfficientNetV2-S, and itself that degrades gracefully, and that this makes it the most reliable architecture for…

desk verdict Plausible VGG16-robustness finding, but Table II's internal inconsistencies undercut the paper's central claim as written. read the letter →

arxiv 2507.09731 v1 pith:LWO75JOR submitted 2025-07-13 eess.IV cs.CV

Robby Hoover , Nelly Elsayed , Zag ElSayed , Chengcheng Li This is my paper · ORCID
classification eess.IVcs.CV
keywords bonefracturedetectiontransferlearningpre-trainedCNNsnoiserobustnessmedicalimagingX-rayclassificationPoisson-Gaussianlow-resourcehealthcare
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper sets out to answer whether pre-trained deep-learning fracture classifiers can be trusted when they receive degraded X-ray images, the kind produced by older or low-cost equipment common in resource-limited clinics. It fine-tunes three pre-trained networks (ResNet50, VGG16, EfficientNetV2-S) on a public bone-fracture dataset and progressively corrupts test images with Poisson, Gaussian, and mixed noise that mimic quantum and electronic noise. The results it reports point to a complexity-robustness tradeoff: the more sophisticated networks score highest on clean images but collapse under noise, while the simpler VGG16 degrades gradually and remains clinically functional the longest. If correct, this means deployment choices for medical AI should weigh robustness under realistic input degradation as heavily as clean benchmark accuracy.

What carries the argument

The load-bearing mechanism is a progressive noise-augmentation protocol that places each test image on a ladder of increasing Poisson noise (modeling quantum mottle), Gaussian noise (modeling electronic noise), and a Poisson-Gaussian combination, then tracks accuracy, AUC, and F1 after each step. The central comparison object is the degradation curve and the critical failure point, defined as a drop of more than 40% accuracy between consecutive noise levels, which separates catastrophic from graceful failure.

What would settle it

Take a set of radiographs acquired on an old or low-cost X-ray machine, add the same progressive Poisson-Gaussian noise to clean copies of the same anatomy, and compare which model keeps clinical accuracy on the real noisy images; if ResNet50 or EfficientNetV2-S outranks VGG16 there, the complexity-robustness claim is false.

Watch

Extended reading notes

Core claim

The central claim is that model complexity is a liability under realistic imaging noise. On clean images ResNet50 reaches 98.62% accuracy and EfficientNetV2-S 95.06%, slightly edging out VGG16's 95.66%, but under progressive Gaussian noise ResNet and EfficientNet cross the paper's catastrophic-failure threshold early and drop to chance-level performance, while VGG16 holds 83.79% accuracy at a noise level where the others have already failed. The same ranking repeats under Poisson and mixed noise, making VGG16 the only pre-trained architecture that never fails catastrophically. The paper interprets this as overfitting to the clean, ImageNet-like feature distribution: skip connections and compound scaling help on clean data but create feature representations that break when the input distribution shifts.

Load-bearing premise

The ranking depends on the assumption that artificially noising clean X-ray images matches what real aging or low-cost X-ray machines do to images; if real degradation differs, VGG16's advantage might not appear in a clinic.

Editorial extensions

If this is right

  • In low-resource imaging settings, a VGG16-based fracture classifier is preferable to ResNet50 or EfficientNetV2-S despite its lower clean-data accuracy.
  • Robustness reporting should include degradation curves and critical failure points, since clean accuracy alone conceals the models' failure modes.
  • Because Gaussian noise proved the most disruptive, efforts to improve robustness in X-ray AI should prioritize electronic noise in the acquisition chain over quantum noise.
  • The observed precision-recall collapse to zero implies that as images degrade, complex models may silently stop detecting fractures, a safety-relevant warning for clinical deployment.

Reading between the lines

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

  • Beyond the paper, the architecture ranking may not survive exposure to the structural noise, blur, and contrast variation that real low-quality machines also add; a natural next experiment is to repeat the ladder with radiographs from actual older devices.
  • Beyond the paper, fine-tuning ResNet50 and EfficientNetV2-S with noise augmentation could close the robustness gap, testing whether the brittleness is intrinsic to the architectures or an artifact of clean-only fine-tuning.
  • Beyond the paper, a practical evaluation metric suggested by this work would be the area under the degradation curve, letting buyers compare models by how much usable accuracy remains across equipment-quality levels.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The manuscript studies how three pre-trained ImageNet classifiers (ResNet50, VGG16, and EfficientNetV2-S) and a simple CNN baseline degrade on a bone-fracture X-ray classification task when Gaussian, Poisson, or mixed Gaussian-Poisson noise is added to the test images. The authors fine-tune each model on the 'Bone Fracture Multi-Region X-ray Data' dataset, progressively add noise, and track accuracy, AUC, and F1-score. They define 'critical failure' as a greater than 40% accuracy drop between consecutive noise levels, summarize the results in Table II, and conclude that VGG16 is the optimal architecture for low-quality imaging because it degrades gracefully, while more complex models fail catastrophically. The paper also proposes a general methodology for evaluating AI robustness under simulated equipment degradation.

Significance. The paper addresses a practically important question: whether complex pre-trained networks remain usable for fracture detection when X-ray equipment quality is poor. Its strengths are that the experiments are direct accuracy measurements on a publicly available dataset, the Simple CNN baseline is a useful control, and the authors explicitly acknowledge dataset-specific and noise-model limitations. There is no circularity in the main results, which are direct measurements rather than outcomes forced by construction. However, the central claim that VGG16 is optimal currently rests on categorical labels in Table II that are internally inconsistent with the stated failure definition, and on a complexity-robustness tradeoff that is contradicted by the reported clean-data accuracies. If the data presentation and the noise protocol are corrected, the framework would provide modest, reproducible evidence; as written, the main conclusion is not supported.

major comments (4)
  1. [III.F.1 and Table II] The critical-failure labels in Table II are inconsistent with the definition given in Section III.F.1. The definition requires a >40% accuracy drop between consecutive noise levels, but EfficientNet's listed critical failure point of 0.0001 corresponds to 75.69% to 52.77%, a relative drop of about 30%, not >40%. Additionally, the text places thresholds 'somewhere between the noise levels of 0.001 - 0.0005,' whereas both listed points (0.0001 and 5e-5) are below that interval. Since the 'catastrophic vs. graceful' labels are the direct evidence for VGG16's superiority, this inconsistency must be resolved before the main claim can be assessed.
  2. [III.D] The noise augmentation schedule is never specified. The manuscript does not report the noise amplitudes, the standard deviation values for Gaussian noise, the Poisson scaling, the number of progressive noise levels, or the ordering of levels; Section III.F.1 gives only a vague reference to levels 'between 0.001 - 0.0005.' Without this schedule, the experiment cannot be reproduced and the critical-failure comparisons in Table II cannot be verified. This is a load-bearing methodological omission because the entire conclusion depends on where and how suddenly each model's accuracy drops.
  3. [Section IV and Table II] The stated complexity-robustness tradeoff is not supported by the reported clean-data accuracies. Section IV claims that 'the more sophisticated networks, ResNet and EfficientNet, showed better performance on clean data than the other two networks,' but Table II lists ResNet50 at 95.06% clean accuracy, below VGG16's 95.66% and only slightly above Simple CNN's 92.09%. Only EfficientNet (98.62%) actually outperforms all other models on clean data. The conclusion that complexity explains brittleness should either be revised or supported with a characterization that separates EfficientNet from ResNet50.
  4. [VI Conclusion and Study Limitations] The conclusion states that VGG16 is 'the optimal choice of architecture for building a robust bone fracture classification system that can hold up to lower-quality medical imaging,' but the study uses a single dataset, a single binary classification task, and three pre-trained architectures with no repeated runs, confidence intervals, or external validation. The authors themselves note that the results 'are specific to this dataset and may not generalize.' The conclusion should be explicitly scoped to these experimental conditions, or additional evidence (e.g., a second X-ray dataset or real noisy radiographs) is needed to justify the broader optimality claim.
minor comments (5)
  1. [Abstract] The phrase 'Medical Imagings' in the first sentence is ungrammatical and should read 'Medical imaging.'
  2. [III.B and III.C] There is a contradiction in the training protocol: Section III.B says the validation set is used to determine when to stop training via early stopping, while Section III.C says all models are trained for five epochs. The authors should clarify which procedure was actually followed.
  3. [III.D] In the discussion of quantum noise, 'a poison distribution' should be 'a Poisson distribution.'
  4. [III.B] The sentence 'Rotation is not be specifically applied' contains a grammatical error and should read 'Rotation is not specifically applied.'
  5. [Figure 2] The x-axis of Figure 2 is described as log-scale but the axis label is missing; adding explicit labels for the noise levels and units would improve interpretability.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the VGG16 robustness conclusion rests on direct noisy-input measurements, not on definitions, fitted parameters, or load-bearing self-citation.

full rationale

The paper is an empirical robustness comparison rather than a derivation. The claimed chain is: fine-tune ResNet50, VGG16, EfficientNetV2-S, and a Simple CNN on the Bone Fracture Multi-Region X-ray dataset, apply progressively stronger simulated Poisson, Gaussian, and mixed noise to test inputs, and measure accuracy, AUC, and F1. The central conclusion that VGG16 is most robust is a direct measurement of test-set performance under added noise; no parameter is fitted to the outcome, and the noise levels and noise types are not chosen in a way that forces VGG16 to win. The definition of critical failure as a >40% accuracy drop between consecutive noise levels is arbitrary but not circular, because it does not by construction produce the observed VGG16 advantage. The internal inconsistencies in Table II (e.g., ResNet50 assigned a critical failure point despite a 45.1% drop under the paper's own definition, and Simple CNN labeled both 'No catastrophic failure' and 'Failed') are correctness concerns, not circularity. The paper's self-citations by the same authors appear only in background enumerations of deep learning and medical imaging applications and are not load-bearing for the noise-robustness result. The Study Limitations explicitly acknowledges that the results are dataset-specific and that structural noise, blur, and geometric distortions are not covered, which supports the empirical, non-circular nature of the claim. No equation, fitted parameter, or renamed known result is reused as a prediction; the conclusion rests on externally measured model behavior.

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

The paper does not introduce new theoretical entities or fitted model parameters. Its central conclusions rely on domain assumptions about noise modeling and transfer learning, plus two hand-chosen methodological parameters (the failure threshold and noise schedule). The latter are under-specified and inconsistently applied, which weakens the empirical support.

free parameters (2)
  • Critical failure point threshold = >40% accuracy drop between consecutive noise levels
    This hand-chosen threshold is used to categorize models as 'catastrophic' or 'graceful' failure, but Table II applies it inconsistently (e.g., EfficientNet's listed drop is ~23 points). The choice of 40% directly influences the qualitative conclusions.
  • Noise level schedule = Not specified in text (only references to 'progressive' levels and values like 0.001, 0.0005, 5e-5)
    The exact sequence and parameters of Gaussian/Poisson/mixed noise additions are not reported. These values determine the degradation curves and are essential for replicating the experiment.
assumptions (4)
  • domain assumption Poisson noise is a faithful model for quantum noise in X-ray imaging (Section III.D).
    The paper states that quantum noise follows a Poisson distribution and uses this to justify adding Poisson noise, but does not validate that this synthetic noise matches real X-ray degradation at the tested levels.
  • domain assumption Gaussian noise is a faithful model for electronic noise in X-ray imaging (Section III.D).
    Electronic noise is modeled as Gaussian; the paper does not provide evidence that the chosen Gaussian noise characteristics resemble real equipment noise.
  • domain assumption ImageNet pre-trained weights are a suitable starting point for fine-tuning on X-ray images (Section III.C).
    The transfer learning approach assumes that features learned on natural images generalize to bone X-rays well enough that pre-training is beneficial; this is common but unverified here beyond the final accuracies.
  • domain assumption The single provided dataset split is representative and the test set is sufficiently large for reliable comparisons (Section III.A).
    The paper uses one dataset and one split (87/8/5) without reporting test set size or confidence intervals; conclusions about relative model robustness rest on this single split.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Pre-trained Under Noise: A Framework for Robust Bone Fracture Detection in Medical Imaging." pith.science (2026). https://pith.science/paper/LWO75JOR

@misc{pith2026250709731,
  author       = {Pith},
  title        = {Pith review of: Pre-trained Under Noise: A Framework for Robust Bone Fracture Detection in Medical Imaging},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LWO75JOR}},
  note         = {Machine review of arXiv:2507.09731}
}
read the original abstract

Medical Imagings are considered one of the crucial diagnostic tools for different bones-related diseases, especially bones fractures. This paper investigates the robustness of pre-trained deep learning models for classifying bone fractures in X-ray images and seeks to address global healthcare disparity through the lens of technology. Three deep learning models have been tested under varying simulated equipment quality conditions. ResNet50, VGG16 and EfficientNetv2 are the three pre-trained architectures which are compared. These models were used to perform bone fracture classification as images were progressively degraded using noise. This paper specifically empirically studies how the noise can affect the bone fractures detection and how the pre-trained models performance can be changes due to the noise that affect the quality of the X-ray images. This paper aims to help replicate real world challenges experienced by medical imaging technicians across the world. Thus, this paper establishes a methodological framework for assessing AI model degradation using transfer learning and controlled noise augmentation. The findings provide practical insight into how robust and generalizable different pre-trained deep learning powered computer vision models can be when used in different contexts.

Figures

Figures reproduced from arXiv: 2507.09731 by the authors.

Figure 1
Figure 1. An image from the ’fractured’ class before and after augmentations [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Model accuracy degradation as Gaussian and Poisson noise levels increase over time (log-scale). [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

61 extracted references · 58 canonical work pages

  1. [1]

    Optimizing integrated imaging service delivery by tier in low-resource health systems,

    K. DeStigter, K. Pool, A. Leslie, et al., “Optimizing integrated imaging service delivery by tier in low-resource health systems,” Insights into Imaging, vol. 12, p. 129, 2021

    work page 2021

  2. [2]

    Deep learning in medical imaging: general overview,

    J.-G. Lee, S. Jun, Y .-W. Cho, H. Lee, G. B. Kim, J. B. Seo, and N. Kim, “Deep learning in medical imaging: general overview,” Korean journal of radiology, vol. 18, no. 4, pp. 570–584, 2017

    work page 2017

  3. [3]

    A Convolutional-based Model for Early Prediction of Alzheimer's based on the Dementia Stage in the MRI Brain Images

    S. Pellakur, N. Elsayed, Z. ElSayed, and M. Ozer, “A convolutional- based model for early prediction of alzheimer’s based on the dementia stage in the mri brain images,” arXiv preprint arXiv:2302.01417 , 2023

    work page Pith review arXiv 2023

  4. [4]

    Deep learning for medical image processing: Overview, challenges and the future,

    M. I. Razzak, S. Naz, and A. Zaib, “Deep learning for medical image processing: Overview, challenges and the future,” Classification in BioApps: Automation of decision making , pp. 323–350, 2017

    work page 2017

  5. [5]

    Deep Learning Approach for Early Stage Lung Cancer Detection

    S. Abunajm, N. Elsayed, Z. ElSayed, and M. Ozer, “Deep learn- ing approach for early stage lung cancer detection,” arXiv preprint arXiv:2302.02456, 2023

    work page Pith review arXiv 2023

  6. [6]

    Deep learning-based breast cancer classification through medical imaging modalities: state of the art and research challenges,

    G. Murtaza, L. Shuib, A. W. Abdul Wahab, G. Mujtaba, G. Mujtaba, H. F. Nweke, M. A. Al-garadi, F. Zulfiqar, G. Raza, and N. A. Azmi, “Deep learning-based breast cancer classification through medical imaging modalities: state of the art and research challenges,” Artificial Intelligence Review, vol. 53, pp. 1655–1720, 2020

    work page 2020

  7. [7]

    A Transfer Learning Based Approach for Classification of COVID-19 and Pneumonia in CT Scan Imaging

    G. Desai, N. Elsayed, Z. Elsayed, and M. Ozer, “A transfer learning based approach for classification of covid-19 and pneumonia in ct scan imaging,” arXiv preprint arXiv:2210.09403 , 2022

    work page Pith review arXiv 2022

  8. [8]

    Overview of deep learning in medical imaging,

    K. Suzuki, “Overview of deep learning in medical imaging,” Radiolog- ical physics and technology , vol. 10, no. 3, pp. 257–273, 2017

    work page 2017

Show all 61 references
  1. [9]

    Access to medical devices in low-income countries: Addressing sustainability challenges in medical device donations,

    B. Compton, D. M. Barash, J. Farrington, C. Hall, D. Herzog, V . Meka, E. Rafferty, K. Taylor, and A. Varghese, “Access to medical devices in low-income countries: Addressing sustainability challenges in medical device donations,” NAM Perspectives, 2018

    2018

  2. [10]

    A compara- tive study of x-ray based medical imaging devices,

    M. T. Mustapha, B. Uzun, D. U. Ozsahin, and I. Ozsahin, “A compara- tive study of x-ray based medical imaging devices,” in Applications of Multi-Criteria Decision-Making Theories in Healthcare and Biomedical Engineering, pp. 163–180, Elsevier, 2021

    2021

  3. [11]

    A. B. Wolbarst, Looking within: how X-ray, CT, MRI, ultrasound, and other medical images are created, and how they help physicians save lives. Univ of California Press, 1999

    1999

  4. [12]

    Noise issues prevailing in various types of medical images,

    B. Goyal, A. Dogra, S. Agrawal, and B. S. Sohi, “Noise issues prevailing in various types of medical images,” Biomedical and Pharmacology Journal, vol. 11, no. 3, 2018

    2018

  5. [13]

    Review of medical imaging with emphasis on x-ray de- tectors,

    M. Hoheisel, “Review of medical imaging with emphasis on x-ray de- tectors,” Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment , vol. 563, no. 1, pp. 215–224, 2006

    2006

  6. [14]

    Medical x-ray imaging, current status and some future challenges,

    E. L. Ritman et al. , “Medical x-ray imaging, current status and some future challenges,” Advances in X-ray analysis , vol. 49, pp. 1–12, 2006

    2006

  7. [15]

    Digital radiography image denoising using a generative adversarial network,

    Y . Sun, X. Liu, P. Cong, L. Li, and Z. Zhao, “Digital radiography image denoising using a generative adversarial network,” Journal of X- ray Science and Technology , vol. 26, no. 4, pp. 523–534, 2018

    2018

  8. [16]

    X-ray-based medical imaging and reso- lution,

    W. Huda and R. B. Abrahams, “X-ray-based medical imaging and reso- lution,” American Journal of Roentgenology, vol. 204, no. 4, pp. W393– W397, 2015

    2015

  9. [17]

    Noise issues prevailing in various types of medical images,

    B. Goyal, S. Agrawal, and B. Sohi, “Noise issues prevailing in various types of medical images,” Biomedical & Pharmacology Journal, vol. 11, no. 3, p. 1227, 2018

    2018

  10. [18]

    A method for modeling noise in medical images,

    P. Gravel, G. Beaudoin, and J. A. De Guise, “A method for modeling noise in medical images,” IEEE Transactions on medical imaging , vol. 23, no. 10, pp. 1221–1232, 2004

    2004

  11. [19]

    Denoising of medical images corrupted by poisson noise,

    I. Rodrigues, J. Sanches, and J. Bioucas-Dias, “Denoising of medical images corrupted by poisson noise,” in 2008 15th IEEE International Conference on Image Processing , pp. 1756–1759, 2008

    2008

  12. [20]

    Optimizing medical image contrast, detail and noise in the digital era,

    P. Sprawls, “Optimizing medical image contrast, detail and noise in the digital era,” Medical Physics International , vol. 2, no. 1, 2014

    2014

  13. [21]

    Deep learning for computer vision: A brief review,

    A. V oulodimos, N. Doulamis, A. Doulamis, and E. Protopapadakis, “Deep learning for computer vision: A brief review,” Computational intelligence and neuroscience , vol. 2018, no. 1, p. 7068349, 2018

    2018

  14. [22]

    A review on action recognition for accident detection in smart city transportation systems,

    V . A. Adewopo, N. Elsayed, Z. ElSayed, M. Ozer, A. Abdelgawad, and M. Bayoumi, “A review on action recognition for accident detection in smart city transportation systems,” Journal of Electrical Systems and Information Technology, vol. 10, no. 1, p. 57, 2023

    2023

  15. [23]

    Deep learning in computer vision: A critical review of emerging techniques and application scenar- ios,

    J. Chai, H. Zeng, A. Li, and E. W. Ngai, “Deep learning in computer vision: A critical review of emerging techniques and application scenar- ios,” Machine Learning with Applications , vol. 6, p. 100134, 2021

    2021

  16. [24]

    Deep learning for biological image classification,

    C. Affonso, A. L. D. Rossi, F. H. A. Vieira, A. C. P. de Leon Ferreira, et al., “Deep learning for biological image classification,” Expert systems with applications, vol. 85, pp. 114–122, 2017

    2017

  17. [25]

    Smart city transportation: Deep learning ensemble approach for traffic accident detection,

    V . A. Adewopo and N. Elsayed, “Smart city transportation: Deep learning ensemble approach for traffic accident detection,” IEEE Access, vol. 12, pp. 59134–59147, 2024

    2024

  18. [26]

    Reduced-gate convolutional lstm architecture for next-frame video prediction using predictive cod- ing,

    N. Elsayed, A. S. Maida, and M. Bayoumi, “Reduced-gate convolutional lstm architecture for next-frame video prediction using predictive cod- ing,” in 2019 international joint conference on neural networks (IJCNN), pp. 1–9, IEEE, 2019

    2019

  19. [27]

    Medical image classification using synergic deep learning,

    J. Zhang, Y . Xie, Q. Wu, and Y . Xia, “Medical image classification using synergic deep learning,” Medical image analysis, vol. 54, pp. 10– 19, 2019

    2019

  20. [28]

    Brownlee, Deep learning for computer vision: image classification, object detection, and face recognition in python

    J. Brownlee, Deep learning for computer vision: image classification, object detection, and face recognition in python . Machine Learning Mastery, 2019

    2019

  21. [29]

    Baby physical safety mon- itoring in smart home using action recognition system,

    V . Adewopo, N. Elsayed, and K. Anderson, “Baby physical safety mon- itoring in smart home using action recognition system,” in SoutheastCon 2023, pp. 142–149, IEEE, 2023

    2023

  22. [30]

    Data augmentation for improving deep learning in image classification problem,

    A. Mikołajczyk and M. Grochowski, “Data augmentation for improving deep learning in image classification problem,” in 2018 international interdisciplinary PhD workshop (IIPhDW) , pp. 117–122, IEEE, 2018

    2018

  23. [31]

    Dos and ddos attack detection in iot infrastructure using xception model with explainability,

    N. Elsayed, Z. ElSayed, and A. Abdelgawad, “Dos and ddos attack detection in iot infrastructure using xception model with explainability,” The International FLAIRS Conference Proceedings , vol. 38, May 2025

    2025

  24. [32]

    Fingerprint minutiae extraction using deep learning,

    L. N. Darlow and B. Rosman, “Fingerprint minutiae extraction using deep learning,” in 2017 IEEE International Joint Conference on Bio- metrics (IJCB), pp. 22–30, IEEE, 2017

    2017

  25. [33]

    Flowpic: Encrypted internet traffic classi- fication is as easy as image recognition,

    T. Shapira and Y . Shavitt, “Flowpic: Encrypted internet traffic classi- fication is as easy as image recognition,” in IEEE INFOCOM 2019- IEEE conference on computer communications workshops (INFOCOM WKSHPS), pp. 680–687, IEEE, 2019

    2019

  26. [34]

    A compar- ison of deep learning performance against health-care professionals in detecting diseases from medical imaging: a systematic review and meta- analysis,

    X. Liu, L. Faes, A. U. Kale, S. K. Wagner, D. J. Fu, A. Bruynseels, T. Mahendiran, G. Moraes, M. Shamdas, C. Kern, et al. , “A compar- ison of deep learning performance against health-care professionals in detecting diseases from medical imaging: a systematic review and meta- ...

    2019

  27. [35]

    Deep learning in medical imaging,

    M. Kim, J. Yun, Y . Cho, K. Shin, R. Jang, H.-j. Bae, and N. Kim, “Deep learning in medical imaging,” Neurospine, vol. 16, pp. 657– 668, Dec 2019. Erratum in: Neurospine. 2020 Jun;17(2):471-472. doi: 10.14245/ns.1938396.198.c1

    2019 doi

  28. [36]

    Recruiting doctors from poor countries: the great brain robbery?,

    V . Patel, “Recruiting doctors from poor countries: the great brain robbery?,” BMJ, vol. 327, no. 7420, p. 926, 2003

    2003

  29. [37]

    Deep learning-enabled medical computer vision,

    A. Esteva, K. Chou, S. Yeung, N. Naik, A. Madani, A. Mottaghi, Y . Liu, E. Topol, J. Dean, and R. Socher, “Deep learning-enabled medical computer vision,” NPJ digital medicine , vol. 4, no. 1, p. 5, 2021

    2021

  30. [38]

    Computer vision to enhance healthcare domain: An overview of features, implementation, and opportunities,

    M. Javaid, A. Haleem, R. P. Singh, and M. Ahmed, “Computer vision to enhance healthcare domain: An overview of features, implementation, and opportunities,” Intelligent Pharmacy, 2024

    2024

  31. [39]

    Computer vision and machine learning for medical image analysis: recent advances, challenges, and way forward,

    E. Elyan, P. Vuttipittayamongkol, P. Johnston, K. Martin, K. McPherson, C. F. Moreno-Garc´ıa, C. Jayne, and M. M. K. Sarker, “Computer vision and machine learning for medical image analysis: recent advances, challenges, and way forward,” Artificial Intelligence Surgery , vol. ...

    2022

  32. [40]

    A guide to deep learning in healthcare,

    A. Esteva, A. Robicquet, B. Ramsundar, V . Kuleshov, M. DePristo, K. Chou, C. Cui, G. Corrado, S. Thrun, and J. Dean, “A guide to deep learning in healthcare,” Nature medicine, vol. 25, no. 1, pp. 24–29, 2019

    2019

  33. [41]

    Sensitivity of different machine learning algorithms to noise,

    A. Atla, R. Tada, V . Sheng, and N. Singireddy, “Sensitivity of different machine learning algorithms to noise,” Journal of Computing Sciences in Colleges, vol. 26, no. 5, pp. 96–103, 2011

    2011

  34. [42]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the IEEE conference on computer vision and pattern recognition , pp. 770–778, 2016

    2016

  35. [43]

    Very deep convolutional networks for large-scale image recognition,

    K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” arXiv preprint arXiv:1409.1556 , 2014

    2014 arXiv

  36. [44]

    Efficientnetv2: Smaller models and faster training,

    M. Tan and Q. Le, “Efficientnetv2: Smaller models and faster training,” in International Conference on Machine Learning , pp. 10096–10106, PMLR, 2021

    2021

  37. [45]

    Bone fracture multi-region x-ray data,

    M. Rodrigo, M. Kumar, A. Faramawy, and H. Arya, “Bone fracture multi-region x-ray data,” 2024

    2024

  38. [46]

    Enlarging smaller images before inputting into convolu- tional neural network: zero-padding vs. interpolation,

    M. Hashemi, “Enlarging smaller images before inputting into convolu- tional neural network: zero-padding vs. interpolation,” Journal of Big Data, vol. 6, no. 98, 2019

    2019

  39. [47]

    Imagenet pre-trained models with batch normalization,

    M. Simon, E. Rodner, and J. Denzler, “Imagenet pre-trained models with batch normalization,” arXiv preprint arXiv:1612.01452 , 2016

    2016 arXiv

  40. [48]

    Does image normal- ization and intensity resolution impact texture classification?,

    M. Kociolek, M. Strzelecki, and R. Obuchowicz, “Does image normal- ization and intensity resolution impact texture classification?,” Comput- erized Medical Imaging and Graphics , vol. 81, p. 101716, 2020

    2020

  41. [49]

    V . U, H. Whitehead, R. Sharma, et al., “Noise.” Radiopaedia.org, 2025. Accessed on 21 Apr 2025

    2025

  42. [50]

    K. M. Hanson, Noise and computed contrast discrimination in tomog- raphy, vol. 5, ch. 113. St. Louis: C. V . Mosby, 1981. Technical Aspects of Computed Tomography

    1981

  43. [51]

    G. P, B. D, M. C, et al., “Noise.” Radiopaedia.org, 2025. Accessed on 21 Apr 2025

    2025

  44. [52]

    Poisson-gaussian noise analysis and estimation for low-dose x-ray images in the nsct domain,

    S. Lee, M. S. Lee, and M. G. Kang, “Poisson-gaussian noise analysis and estimation for low-dose x-ray images in the nsct domain,” Sensors, vol. 18, no. 4, p. 1019, 2018

    2018

  45. [53]

    The classification of noise-afflicted remotely sensed data using three machine-learning techniques: effect of different levels and types of noise on accuracy,

    S. Boonprong, C. Cao, W. Chen, X. Ni, M. Xu, and B. K. Acharya, “The classification of noise-afflicted remotely sensed data using three machine-learning techniques: effect of different levels and types of noise on accuracy,” ISPRS International Journal of Geo-Information , vol...

    2018

  46. [54]

    A review of image denoising algorithms, with a new one,

    A. Buades, B. Coll, and J.-M. Morel, “A review of image denoising algorithms, with a new one,” Multiscale modeling & simulation , vol. 4, no. 2, pp. 490–530, 2005

    2005

  47. [55]

    Relative contributions of anatomical and quantum noise in signal detection and perception of tomographic digital breast images,

    A. Kavuri and M. Das, “Relative contributions of anatomical and quantum noise in signal detection and perception of tomographic digital breast images,” IEEE transactions on medical imaging , vol. 39, no. 11, pp. 3321–3330, 2020

    2020

  48. [56]

    Modeling quantum noise of phosphors used in medical x-ray imaging detectors,

    N. Kalivas, I. Kandarakis, D. Cavouras, L. Costaridou, C. Nomicos, and G. Panayiotakis, “Modeling quantum noise of phosphors used in medical x-ray imaging detectors,” Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Asso...

    1999

  49. [57]

    Iniewski, Medical imaging: principles, detectors, and electronics

    K. Iniewski, Medical imaging: principles, detectors, and electronics . John Wiley & Sons, 2009

    2009

  50. [58]

    A variational method to remove the combination of poisson and gaussian noises,

    D. N. Thanh and S. D. Dvoenko, “A variational method to remove the combination of poisson and gaussian noises,” in International Workshop on Image Mining. Theory and Applications, vol. 2, pp. 38–45, SCITEPRESS, 2015

    2015

  51. [59]

    A method of total variation to remove the mixed poisson-gaussian noise,

    D. Thanh and S. Dvoenko, “A method of total variation to remove the mixed poisson-gaussian noise,” Pattern Recognition and Image Analysis, vol. 26, no. 2, pp. 285–293, 2016

    2016

  52. [60]

    Image restoration with poisson–gaussian mixed noise,

    A. Lanza, S. Morigi, F. Sgallari, and Y .-W. Wen, “Image restoration with poisson–gaussian mixed noise,” Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, vol. 2, no. 1, pp. 12– 24, 2014

    2014

  53. [61]

    An unbiased risk estimator for image denoising in the presence of mixed poisson– gaussian noise,

    Y . Le Montagner, E. D. Angelini, and J.-C. Olivo-Marin, “An unbiased risk estimator for image denoising in the presence of mixed poisson– gaussian noise,” IEEE Transactions on Image processing, vol. 23, no. 3, pp. 1255–1268, 2014

    2014

Pith tools

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