Pith. sign in

REVIEW 5 major objections 5 minor 44 references

Diffusion-based Counterfactual Augmentation: Towards Robust and Interpretable Knee Osteoarthritis Grading

T0 review · 5 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 Counterfactual X-rays generated by a classifier-guided latent SDE improve knee osteoarthritis grading across architectures and datasets, and reveal a latent barrier structure aligned with clinical progression.

desk verdict A solid, well-documented application of latent-diffusion counterfactual augmentation to knee OA grading, but the missing ablations leave the mechanism underdetermined. read the letter →

arxiv 2506.15748 v1 pith:UIECKQDV submitted 2025-06-18 eess.IV cs.CV

Zhe Wang , Yuhua Ru , Aladine Chetouani , Tina Shiang , Fang Chen , Fabian Bauer , Liping Zhang , Didier Hans
show 4 more authors
Rachid Jennane William Ewing Palmer Mohamed Jarraya Yung Hsin Chen
This is my paper · ORCID
classification eess.IVcs.CV
keywords KneeosteoarthritisKellgren-LawrencegradingCounterfactualaugmentationDiffusionmodelsStochasticdifferentialequationsSelf-correctivelearningMedicalimageinterpretabilityRadiographclassification
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

This paper tries to establish that a knee-osteoarthritis classifier can improve its own grading by training on counterfactual radiographs it helped generate: images that start from a real case and are pushed just across the decision boundary into an adjacent Kellgren-Lawrence grade. The push happens in the latent space of a diffusion model, where a stochastic differential equation balances a boundary drive from a frozen classifier against a manifold constraint that keeps images realistic, followed by a short reverse-diffusion refinement. A trainable twin of the frozen classifier is then fine-tuned on the original images plus these counterfactuals, with an alignment loss that keeps its predictions close to the frozen reference on the generated samples. Across the OAI and MOST cohorts, the reported result is consistent, often statistically significant accuracy gains for adjacent-grade tasks on several architectures, together with the claim that the minimal SDE transition time between grades behaves like a clinical energy barrier: hardest for KL-2 to KL-3, and systematically easier for regression than progression. If these claims hold, model uncertainty is converted into a targeted training signal rather than discarded.

What carries the argument

The load-bearing mechanism is a stochastic differential equation in the latent space of a latent diffusion model, with drift $f(z(s)) = (1-\lambda)\, v_{\mathrm{manifold}}(s) + \lambda\, v_{\mathrm{boundary}}(s)$. The manifold term is the normalized score function estimated by the denoising network; the boundary term is the normalized gradient, taken with respect to the decoded image, of $\log P(y' \mid x) - \log P(y \mid x)$ from a frozen reference classifier. Euler-Maruyama simulation of this SDE produces a trajectory of latent states, and a 50-step reverse-diffusion refinement pulls every sampled intermediate state back onto the learned data manifold. The second half of the machinery is self-corrective learning: a trainable classifier is trained with a multi-task loss that combines cross-entropy on originals and counterfactuals, with labels dynamically assigned from the classifier's instantaneous probability distribution during simulation, and an alignment loss $\mathcal{L}_{\mathrm{align}} = \| C(x') - C^*(x') \|_2^2$ that keeps the trainable classifier close to the frozen reference on the counterfactuals.

What would settle it

Blinded expert radiologists grade the DCA-generated counterfactual images: if a substantial share of images generated as KL-3 from KL-2 sources are not rated KL-3, or show no new joint-space narrowing or osteophytes, then the claimed clinical plausibility is unsupported and the accuracy gains need another explanation.

Watch

Extended reading notes

Core claim

The central claim is that Diffusion-based Counterfactual Augmentation (DCA) turns a classifier's own uncertainty into a reliable training signal. The framework samples a trajectory in latent space that crosses the frozen reference classifier's decision boundary between adjacent KL grades, decodes and refines intermediate states with reverse diffusion, and trains a learnable classifier to agree with the frozen one on those boundary-adjacent inputs. The reported evidence is that this self-corrective loop improves binary adjacent-grade accuracy for VGG-11, EfficientNet-B0, Inception-V3, ViT-B/16, and two published KOA classifiers on both OAI and MOST, with many comparisons reaching $p<0.05$ and often $p<0.001$. The paper further claims that the minimum simulation time $T_{\min}^{\mathrm{SDE}}$ needed to cross a boundary is largest for the KL-2 to KL-3 transition and larger for progression than for regression, which it reads as a latent-space topology that mirrors the non-linear, asymmetric difficulty of real disease progression.

Load-bearing premise

The counterfactual images are labelled by the same frozen classifier that guides their generation, and the new classifier is trained to agree with that frozen one on those images; if the frozen classifier's idea of a KL grade is not clinically accurate, the procedure locks in its mistakes instead of fixing them.

Editorial extensions

If this is right

  • Adjacent-grade classification accuracy improves after DCA augmentation across all evaluated architectures on both OAI and MOST, with the strongest and most consistent gains on distinctions where a real pathological signal exists, such as KL-1 vs KL-2, KL-2 vs KL-3, and KL-3 vs KL-4.
  • The generated counterfactuals are closer to real data and to the intended target grade than those from DiverseCF, GANterfactual, and DiME, as measured by PSNR, SSIM, LPIPS, and RMSE, and visually show structural changes such as joint-space narrowing rather than imperceptible perturbations.
  • The minimum SDE transition time $T_{\min}^{\mathrm{SDE}}$ functions as a quantitative interpretability metric: it identifies KL-2 to KL-3 as the hardest adjacent transition and shows regressive transitions are systematically cheaper than progressive ones, aligning the learned latent space with clinical knowledge of KOA progression.
  • Diffusion-based refinement is necessary: without it, generated counterfactuals have substantially higher MMD and k-NN distances from real data and scatter farther from the real-data cluster in t-SNE.

Reading between the lines

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

  • The paper does not test whether the frozen reference classifier's boundaries correspond to expert consensus; if they do not, the alignment loss will propagate those systematic errors into the improved classifier, so an expert-grading study of the generated counterfactuals is the natural next check.
  • The fixed simulation-time scaling factor of $1.5 \times \bar{T}_{\min}^{\mathrm{SDE}}$ is chosen empirically; an adaptive schedule that stops the SDE at the per-sample boundary crossing could be tested against the fixed rule and would likely change both sample diversity and training signal strength.
  • The same energy-barrier analysis could be applied to other ordinal medical imaging tasks with ambiguous boundaries, such as diabetic retinopathy or vertebral fracture grading, where a latent topology that fails to show the clinically expected asymmetry would flag a misaligned representation.
  • Because labels on counterfactuals are assigned dynamically from the classifier's own probability distribution, the accuracy gains may partly reflect label smoothing or boundary sharpening rather than new anatomical information; comparing DCA against simple mixup or margin-based augmentation on the same latent trajectories would isolate the contribution of counterfactual realism.
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

5 major / 5 minor

Summary. The paper proposes a Diffusion-based Counterfactual Augmentation (DCA) framework for knee osteoarthritis (KOA) grading. It generates counterfactual radiographs by steering a latent-space SDE with a classifier-informed boundary drive and a diffusion manifold constraint, then refines the samples with a few reverse diffusion steps. The generated counterfactuals are used in a 'self-corrective learning' loop: a frozen reference classifier provides pseudo-labels and an alignment target, while a trainable classifier is fine-tuned on the augmented data. The authors evaluate DCA on the OAI and MOST datasets with several architectures, reporting statistically significant accuracy gains and an interpretability analysis based on the minimum SDE time required to cross decision boundaries. They also compare counterfactual image quality against DiverseCF, GANterfactual, and DiME.

Significance. If the central claims are substantiated, DCA would be a meaningful contribution to medical image augmentation and interpretability. The paper has notable strengths: public implementation, evaluation on two large cohorts, multiple backbone architectures, and p-values over repeated trials. However, the main accuracy claim currently lacks crucial control experiments, and the interpretability claim is entangled with the frozen classifier's own decision geometry. The limitations section candidly acknowledges dependence on the frozen classifier and empirical hyperparameters, but the missing ablations prevent the reader from attributing the reported gains specifically to the boundary-targeted counterfactual mechanism.

major comments (5)
  1. [Section V-D, Table VI] The central accuracy claim is not yet supported because the self-corrective learning strategy is not ablated against simpler alternatives. With the current design, the classifier is trained on counterfactual samples whose labels are assigned from the frozen classifier's instantaneous probability distribution, and Eq. (11) additionally penalizes deviation from that same frozen classifier. The reported gains in Table VI could therefore arise from (i) increased training-set size, (ii) generic manifold-constrained samples with no boundary targeting, or (iii) consistency regularization that merely aligns the trainable classifier to the frozen one. I request the following control experiments: (a) unguided diffusion augmentation with the manifold term only (λ=0) and the same number of generated samples; (b) counterfactual samples labeled with the hard target grade y' and no Lalign term; (c) adding the same number of real target-class images to the training set. Without these, the 'statistically significant improvements' do not establish that boundary-targeted counterfactuals are the causal factor.
  2. [Eq. (11) and Section V-D] There is a potential circularity that is acknowledged but not resolved. The frozen reference classifier C* supplies both the boundary gradient in Eq. (7) and the alignment target in Eq. (11), and the pseudo-labels for counterfactual samples are dynamically derived from C*'s own probability distribution. This means the training signal for the new classifier is generated by the same decision function that defines the augmentation trajectory. If C*'s decision boundary does not align with clinical truth, the loop will reinforce its systematic errors. The paper's limitations section admits this dependence, but the claims in the abstract and conclusion ('converts model uncertainty into a robust training signal') go beyond what the current evidence supports. A concrete test would be to use a clinician-derived grading or an independent classifier to validate a subset of the generated counterfactuals and to measure whether the augmented classifier's errors shift toward or away from C*'s characteristic mistakes.
  3. [Section V-B, Fig. 3] The interpretability claim concerning T_min^SDE is not yet established. The transition time is defined as the point at which the frozen classifier's output probability for the target grade exceeds 0.5. Thus the reported 'energy barriers' and the asymmetry between progression and regression are properties of C*'s decision geometry, not necessarily of the underlying disease manifold or of clinical progression. The claim that 'the learned latent space topology aligns with clinical knowledge of KOA progression' requires a control that does not depend on the classifier's own probability threshold. For example, the authors could correlate T_min^SDE with an independent clinical severity measure (e.g., joint space narrowing grade or expert-assessed progression), or compare the topology obtained with a classifier trained on a different label set. Without such a check, the KL-2→KL-3 peak and the progression/regression asymmetry may simply reflect the trained classifier's class-conditional gradients.
  4. [Table VI] The statistical reporting is insufficient to support 'consistent and statistically significant improvements' as a general claim. The table reports only mean accuracies and p-values from an independent t-test over three trials, with no standard deviations, confidence intervals, or effect sizes. With n=3 per condition, a t-test has very low power, and the paper performs many comparisons without correction for multiple testing. The authors should report per-trial results or at least mean ± std across trials, and should state the exact test procedure (paired or unpaired) and whether the normality assumption is plausible. This is load-bearing for the central accuracy claim, because the largest apparent gains occur in settings where the baseline is already high (e.g., KL-2 vs. KL-3 and KL-3 vs. KL-4), where a few misclassified cases can dominate the difference.
  5. [Section IV-B] The description of the frozen reference classifier's training data is ambiguous and potentially problematic. The text states that the Vision Transformer classifier from [16] was 'pre-trained on the target KOA datasets' and later used to guide generation and to pseudo-label counterfactuals. If this classifier was trained on the same full datasets (including any images that later appear in the test splits), then the generation process has access to test-set information, which would inflate the downstream accuracy gains. The authors must specify the exact train/validation/test splits used for the reference classifier, confirm that those splits are disjoint from the evaluation test sets, and state whether the reference classifier was re-trained in the 'closed-loop design' experiments or used as-is from prior work. This is a necessary detail for assessing both the method's validity and the fairness of the comparisons in Section V-C.
minor comments (5)
  1. [Table VI] The table formatting makes it difficult to read: the header 'Acc Acc ⋄ p-value' is repeated and the VGG-11 rows appear split across columns. Please reformat the table so that each model occupies one row with clear OAI and MOST columns, and ensure the p-value column is aligned with the correct accuracy pair.
  2. [Eq. (7)] The boundary gradient g(s) is computed with respect to the image-space reconstruction x(s)=D(z(s)), but the SDE in Eq. (8) updates the latent state z(s). The paper does not explain how an image-space gradient is transported into the latent update; there is no Jacobian term or change-of-variables correction. Please clarify this step or provide a justification for treating the normalized image-space gradient as a valid latent-space direction.
  3. [Section V-C] The validity criterion for generated samples ('confidences exceeding 0.95') is used to select examples for qualitative display, but there is no reporting of what fraction of generated samples pass this threshold. If the threshold excludes many samples, the qualitative comparison may be biased. Please report the pass rates per method and per transition.
  4. [Section IV-B] Several hyperparameters are fixed empirically (λ, classifier guidance scale, T_SDE scaling factor, confidence threshold, and β) without a sensitivity analysis. Given that the proposed method is sensitive to these choices, a short ablation over λ and the T_SDE scaling factor would strengthen the reproducibility claims.
  5. [General] The paper uses the phrase 'self-corrective learning' throughout, but the mechanism is essentially a form of self-training/distillation from a frozen teacher. The discussion in Section VI would benefit from situating this work relative to the extensive literature on self-training and pseudo-labeling, which also addresses the risk of confirmation bias; this would help the reader understand what is new.

Circularity Check

2 steps flagged · score 5.0 of 10

Self-corrective training loop is self-referential, but the external accuracy evaluation keeps the paper from full circularity.

  1. self definitional [Section III-C, Eq. (11); Section V-D (self-corrective learning)]
    "Lalign =∥C(x′)−C ∗(x′)∥2 2 (Eq. 11); 'Each sample's label was dynamically assigned based on the classifier's instantaneous probability distribution at the corresponding simulation step'; 'the same classifier architecture used to define decision boundaries for counterfactual generation was subsequently fine-tuned on the generated samples and re-evaluated' (Sec. V-D)."

    Counterfactual trajectories are steered by C*'s gradient (Eqs. 6-7); the synthetic samples' labels are C*'s own instantaneous probability assignments; Eq. 11 then minimizes ||C(x')-C*(x')||2 on exactly those samples. The augmented-distribution supervision is therefore, by construction, a projection onto C*'s decision function, so the loop cannot independently correct C*'s systematic errors. The OAI/MOST test accuracy is external and could still improve, but the internal 'self-corrective' signal reduces to self-consistency with the reference classifier; missing ablations (unguided diffusion augmentation, hard target labels, real target-class images) leave the causal mechanism underdetermined.

  2. other [Section V-B (Interpretable latent space dynamics) and Eq. (7)]
    "we introduce the minimum required simulation time, T min SDE, as a surrogate measure for the energy barrier between adjacent KL grades. To quantify this barrier, we defined a successful transition as the point at which the classifier's output probability for the target KL grade surpasses the 0.5 threshold."

    T_min^SDE is defined by C*'s own probability crossing the 0.5 threshold, and the traversed trajectory is driven by C*'s logit gradient (Eq. 7). Claiming this metric 'reveals' a topology that 'faithfully mirrors' clinical KOA progression is therefore an interpretation of C*'s decision geometry, not an independent probe of the disease manifold. Because C* was trained on the same clinical KL labels, the reported nonlinearity and asymmetry are inherited from the training labels; the barrier metric re-describes the classifier's boundary rather than supplying external confirmation.

full rationale

The strongest accuracy claim is evaluated against held-out clinical KL labels on OAI/MOST, so it does not reduce by construction to the frozen classifier's outputs; that part is externally grounded. However, the self-corrective loop is internally closed: C* supplies the gradient for generating counterfactuals (Eq. 7), the pseudo-labels for those counterfactuals (Section V-D), and the alignment target for the trainable classifier (Eq. 11). On the augmented distribution, the model is trained to reproduce C*'s own decisions, so the 'correction' is self-consistency with the reference classifier rather than correction against an independent clinical standard. The paper acknowledges the dependence in Limitations ('entirely dependent on gradients derived from a frozen reference classifier') but does not provide controls that would separate the genuine contribution of boundary-targeted counterfactuals from pseudo-label self-training or mere dataset enlargement. The T_min^SDE interpretability claim is also self-referential because the barrier is defined by C*'s own probability crossing, so the observed clinical alignment is inherited from the classifier's training labels. Overall, this is partial circularity in the internal training loop, not a fully circular derivation of the external accuracy result.

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

The main contribution rests on a set of tunable hyperparameters and the assumption that a classifier's latent boundary geometry reflects clinical disease progression. No new physical or clinical entities are introduced; T_min^SDE is a derived metric, not a new entity.

free parameters (5)
  • lambda = not reported
    Balances manifold constraint and boundary drive in Eq. (3); the paper does not state the chosen value or how it was selected.
  • Classifier guidance scale = 2.5
    Reported in Section IV-B; influences the strength of boundary guidance in the SDE, but no sensitivity analysis is provided.
  • T_SDE scaling factor = 1.5
    Section V-D: simulation time set to 1.5 times the mean T_min^SDE, selected empirically to push samples confidently into the target class; this directly affects the generated samples and the training signal.
  • Confidence threshold = 0.95
    Section V-C: generated samples are deemed valid only if the frozen classifier assigns them to the intended grade with confidence greater than 0.95; this filters the augmented dataset.
  • beta (beta-VAE) = 0.9
    Section IV-B: beta value of 0.9 for the beta-VAE; impacts latent space structure but is a standard tunable.
assumptions (5)
  • domain assumption The beta-VAE/LDM latent space preserves clinically meaningful semantics for KOA
    The whole method assumes that trajectories in this latent space correspond to pathological changes; this is not independently validated.
  • domain assumption The frozen classifier's decision boundary is a useful proxy for clinical grade boundaries
    Counterfactual generation and pseudo-labels are both derived from this classifier (Sections III-A and V-D).
  • domain assumption The score function approximation via epsilon_theta is accurate enough for manifold guidance
    Eqs. (4)-(5) treat the denoiser output as the data score; no verification of this approximation on the KOA data is provided.
  • standard math Euler-Maruyama discretization of the SDE is adequate
    Standard numerical scheme for SDEs; reasonable but with no convergence check.
  • standard math Three repeated trials justify t-test statistics
    Independent t-tests on three trials assume approximate normality and independence; this is statistically weak but common in the field.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Diffusion-based Counterfactual Augmentation: Towards Robust and Interpretable Knee Osteoarthritis Grading." pith.science (2026). https://pith.science/paper/UIECKQDV

@misc{pith2026250615748,
  author       = {Pith},
  title        = {Pith review of: Diffusion-based Counterfactual Augmentation: Towards Robust and Interpretable Knee Osteoarthritis Grading},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UIECKQDV}},
  note         = {Machine review of arXiv:2506.15748}
}
read the original abstract

Automated grading of Knee Osteoarthritis (KOA) from radiographs is challenged by significant inter-observer variability and the limited robustness of deep learning models, particularly near critical decision boundaries. To address these limitations, this paper proposes a novel framework, Diffusion-based Counterfactual Augmentation (DCA), which enhances model robustness and interpretability by generating targeted counterfactual examples. The method navigates the latent space of a diffusion model using a Stochastic Differential Equation (SDE), governed by balancing a classifier-informed boundary drive with a manifold constraint. The resulting counterfactuals are then used within a self-corrective learning strategy to improve the classifier by focusing on its specific areas of uncertainty. Extensive experiments on the public Osteoarthritis Initiative (OAI) and Multicenter Osteoarthritis Study (MOST) datasets demonstrate that this approach significantly improves classification accuracy across multiple model architectures. Furthermore, the method provides interpretability by visualizing minimal pathological changes and revealing that the learned latent space topology aligns with clinical knowledge of KOA progression. The DCA framework effectively converts model uncertainty into a robust training signal, offering a promising pathway to developing more accurate and trustworthy automated diagnostic systems. Our code is available at https://github.com/ZWang78/DCA.

Figures

Figures reproduced from arXiv: 2506.15748 by the authors.

Figure 1
Figure 1. The flowchart of manifold-constrained counterfactual data generation. Starting from an original data point [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. T-SNE visualization of real data (blue), counterfactual [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Quantitative analysis of the computational effort [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 30 canonical work pages

  1. [16]

    Transformer with selective shuffled position embedding and key-patch exchange strategy for early detection of knee osteoarthritis,

    Z. Wang, A. Chetouani, M. Jarraya, D. Hans, and R. Jennane, “Transformer with selective shuffled position embedding and key-patch exchange strategy for early detection of knee osteoarthritis,”Expert Systems with Applications, vol. 255, p. 124614, 2024

    work page 2024

  2. [1]

    Osteoarthritis: a disease of the joint as an organ,

    R. F. Loeser, S. R. Goldring, C. R. Scanzello, and M. B. Goldring, “Osteoarthritis: a disease of the joint as an organ,”Arthritis and rheumatism, vol. 64, no. 6, p. 1697, 2012

    work page 2012

  3. [2]

    Epidemiology and burden of osteoarthritis,

    A. Litwic, M. H. Edwards, E. M. Dennison, and C. Cooper, “Epidemiology and burden of osteoarthritis,”British Medical Bulletin, vol. 105, no. 1, pp. 185–199, 01 2013. [Online]. Available: https://doi.org/10.1093/bmb/lds038

    work page doi:10.1093/bmb/lds038 2013

  4. [3]

    Current evidence for osteoarthri- tis treatments,

    A. Anandacoomarasamy and L. March, “Current evidence for osteoarthri- tis treatments,”Therapeutic Advances in Musculoskeletal Disease, vol. 2, no. 1, pp. 17–28, 2010

    work page 2010

  5. [4]

    Classifications in brief: Kellgren-lawrence classification of osteoarthritis,

    M. D. Kohn, A. A. Sassoon, and N. D. Fernando, “Classifications in brief: Kellgren-lawrence classification of osteoarthritis,”Clinical Orthopaedics and Related Research®, vol. 474, pp. 1886–1893, 2016

    work page 2016

  6. [5]

    Radiological Assessment of Osteo- Arthrosis,

    J. H. Kellgren and J. S. Lawrence, “Radiological Assessment of Osteo- Arthrosis,”Annals of the Rheumatic Diseases, vol. 16, no. 4, p. 494, 1957. 10

    work page 1957

  7. [6]

    Discriminative regularized auto-encoder for early detection of knee osteoarthritis: Data from the Osteoarthritis Initiative,

    Y . Nasser, R. Jennane, A. Chetouani, E. Lespessailles, and M. E. Hassouni, “Discriminative regularized auto-encoder for early detection of knee osteoarthritis: Data from the Osteoarthritis Initiative,”IEEE Transactions on Medical Imaging, vol. 39, no. 9, pp. 2976–2984, 2020

    work page 2020

  8. [7]

    Fully automatic knee osteoarthritis severity grading using deep neural networks with a novel ordinal loss,

    P. Chen, L. Gao, X. Shi, K. Allen, and L. Yang, “Fully automatic knee osteoarthritis severity grading using deep neural networks with a novel ordinal loss,”Computerized Medical Imaging and Graphics, vol. 75, pp. 84–92, 2019

    2019

Show all 44 references
  1. [8]

    Siamese-gap network for early detection of knee osteoarthritis,

    Z. Wang, A. Chetouani, D. Hans, E. Lespessailles, and R. Jennane, “Siamese-gap network for early detection of knee osteoarthritis,” in2022 IEEE 19th International Symposium on Biomedical Imaging (ISBI), 2022, pp. 1–4

    2022

  2. [9]

    Guest editorial deep learning in medical imaging: Overview and future promise of an exciting new technique,

    H. Greenspan, B. Van Ginneken, and R. M. Summers, “Guest editorial deep learning in medical imaging: Overview and future promise of an exciting new technique,”IEEE transactions on medical imaging, vol. 35, no. 5, pp. 1153–1159, 2016

    2016

  3. [10]

    Imagenet classification with deep convolutional neural networks,

    A. Krizhevsky, I. Sutskever, and G. E. Hinton, “Imagenet classification with deep convolutional neural networks,”Advances in neural information processing systems, vol. 25, pp. 1097–1105, 2012

    2012

  4. [11]

    Attention is all you need,

    A. Vaswani, “Attention is all you need,”Advances in Neural Information Processing Systems, 2017

    2017

  5. [12]

    Quantifying Radiographic Knee Osteoarthritis Severity Using Deep Convolutional Neural Networks,

    J. Antony, K. McGuinness, N. E. O’Connor, and K. Moran, “Quantifying Radiographic Knee Osteoarthritis Severity Using Deep Convolutional Neural Networks,”2016 23rd International Conference on Pattern Recognition (ICPR), pp. 1195–1200, 2016

    2016

  6. [13]

    Automatic detection of knee joints and quantification of knee osteoarthritis severity using convolutional neural networks,

    J. Antony, K. McGuinness, K. Moran, and N. E. O’Connor, “Automatic detection of knee joints and quantification of knee osteoarthritis severity using convolutional neural networks,” inInternational conference on machine learning and data mining in pattern recognition. Springer,...

    2017

  7. [14]

    Automatic grading of individual knee osteoarthritis features in plain radiographs using deep convolutional neural networks,

    A. Tiulpin and S. Saarakkala, “Automatic grading of individual knee osteoarthritis features in plain radiographs using deep convolutional neural networks,”Diagnostics, vol. 10, no. 11, p. 932, 2020

    2020

  8. [15]

    Automatic knee osteoarthritis diagnosis from plain radiographs: A deep learning-based approach,

    A. Tiulpin, J. Thevenot, E. Rahtu, P. Lehenkari, and S. Saarakkala, “Automatic knee osteoarthritis diagnosis from plain radiographs: A deep learning-based approach,”Scientific Reports, 2018

    2018

  9. [17]

    Adversarially robust generalization requires more data,

    L. Schmidt, S. Santurkar, D. Tsipras, K. Talwar, and A. Madry, “Adversarially robust generalization requires more data,”Advances in neural information processing systems, vol. 31, 2018

    2018

  10. [18]

    Understanding deep learning via decision boundary,

    S. Lei, F. He, Y . Yuan, and D. Tao, “Understanding deep learning via decision boundary,”IEEE Transactions on Neural Networks and Learning Systems, 2023

    2023

  11. [19]

    D. L. Dawson and H. L. MacMillan,Relationship management of the borderline patient: from understanding to treatment. Routledge, 2013

    2013

  12. [20]

    A survey on image data augmen- tation for deep learning,

    C. Shorten and T. M. Khoshgoftaar, “A survey on image data augmen- tation for deep learning,”Journal of big data, vol. 6, no. 1, pp. 1–48, 2019

    2019

  13. [21]

    A survey of methods for explaining black box models,

    R. Guidotti, A. Monreale, S. Ruggieri, F. Turini, F. Giannotti, and D. Pedreschi, “A survey of methods for explaining black box models,” ACM computing surveys (CSUR), vol. 51, no. 5, pp. 1–42, 2018

    2018

  14. [22]

    Clinical decision support systems for improving diagnostic accuracy and achieving precision medicine,

    C. Castaneda, K. Nalley, C. Mannion, P. Bhattacharyya, P. Blake, A. Pecora, A. Goy, and K. S. Suh, “Clinical decision support systems for improving diagnostic accuracy and achieving precision medicine,” Journal of clinical bioinformatics, vol. 5, pp. 1–16, 2015

    2015

  15. [23]

    Gan-based synthetic medical image augmentation for increased cnn performance in liver lesion classification,

    M. Frid-Adar, I. Diamant, E. Klang, M. Amitai, J. Goldberger, and H. Greenspan, “Gan-based synthetic medical image augmentation for increased cnn performance in liver lesion classification,”Neurocomputing, vol. 321, pp. 321–331, 2018

    2018

  16. [24]

    Biomedical data augmentation using generative adversarial neural networks,

    F. Calimeri, A. Marzullo, C. Stamile, and G. Terracina, “Biomedical data augmentation using generative adversarial neural networks,” in International conference on artificial neural networks. Springer, 2017, pp. 626–634

    2017

  17. [25]

    Deepfake knee osteoarthritis x-rays from generative adversarial neural networks deceive medical experts and offer augmentation potential to automatic classification,

    F. Prezja, J. Paloneva, I. P ¨ol¨onen, E. Niinim ¨aki, and S. ¨Ayr¨am¨o, “Deepfake knee osteoarthritis x-rays from generative adversarial neural networks deceive medical experts and offer augmentation potential to automatic classification,”Scientific Reports, vol. 12, no. 1, p...

    2022

  18. [26]

    Key-exchange convolutional auto-encoder for data augmentation in early knee osteoarthritis classifi- cation,

    Z. Wang, A. Chetouani, and R. Jennane, “Key-exchange convolutional auto-encoder for data augmentation in early knee osteoarthritis classifi- cation,”arXiv preprint arXiv:2302.13336, 2023

    2023 arXiv

  19. [27]

    Tem- poral evolution of knee osteoarthritis: A diffusion-based morphing model for x-ray medical image synthesis,

    Z. Wang, A. Chetouani, R. Jennane, Y . Ru, W. Issa, and M. Jarraya, “Tem- poral evolution of knee osteoarthritis: A diffusion-based morphing model for x-ray medical image synthesis,”arXiv preprint arXiv:2408.00891, 2024

    2024 arXiv

  20. [28]

    Denoising diffusion probabilistic models,

    J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” Advances in neural information processing systems, vol. 33, pp. 6840– 6851, 2020

    2020

  21. [29]

    Counterfactual explana- tions for medical image classification and regression using diffusion autoencoder,

    M. Atad, D. Schinz, H. Moeller, R. Graf, B. Wiestler, D. Rueckert, N. Navab, J. S. Kirschke, and M. Keicher, “Counterfactual explana- tions for medical image classification and regression using diffusion autoencoder,”arXiv preprint arXiv:2408.01571, 2024

    2024 arXiv

  22. [30]

    Pairwise feature-based generative adversarial network for incomplete multi-modal alzheimer’s disease diagnosis,

    H. Ye, Q. Zhu, Y . Yao, Y . Jin, and D. Zhang, “Pairwise feature-based generative adversarial network for incomplete multi-modal alzheimer’s disease diagnosis,”The Visual Computer, vol. 39, no. 6, pp. 2235–2244, 2023

    2023

  23. [31]

    Mitigating attribute amplification in counterfactual image generation,

    T. Xia, M. Roschewitz, F. De Sousa Ribeiro, C. Jones, and B. Glocker, “Mitigating attribute amplification in counterfactual image generation,” in International Conference on Medical Image Computing and Computer- Assisted Intervention. Springer, 2024, pp. 546–556

    2024

  24. [32]

    High- resolution image synthesis with latent diffusion models,

    R. Rombach, A. Blattmann, D. Lorenz, P. Esser, and B. Ommer, “High- resolution image synthesis with latent diffusion models,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 10 684–10 695

    2022

  25. [33]

    P. E. Kloeden, E. Platen, P. E. Kloeden, and E. Platen,Stochastic differential equations. Springer, 1992

    1992

  26. [34]

    Wiener processes,

    D. Shale and W. F. Stinespring, “Wiener processes,”Journal of Functional Analysis, vol. 2, no. 4, pp. 378–394, 1968

    1968

  27. [35]

    J. R. Norris,Markov chains. Cambridge university press, 1998, no. 2

    1998

  28. [36]

    The Osteoarthritis Initiative: A NIH Public–Private Partner- ship,

    G. Lester, “The Osteoarthritis Initiative: A NIH Public–Private Partner- ship,”HSS Journal: The Musculoskeletal Journal of Hospital for Special Surgery, vol. 8, no. 1, pp. 62–63, 2011

    2011

  29. [37]

    The multicenter osteoarthritis study (most): opportunities for rehabilitation research,

    N. A. Segal, M. C. Nevitt, K. D. Gross, J. Hietpas, N. A. Glass, C. E. Lewis, and J. C. Torner, “The multicenter osteoarthritis study (most): opportunities for rehabilitation research,”PM & R: the journal of injury, function, and rehabilitation, vol. 5, no. 8, 2013

    2013

  30. [38]

    Understanding disentangling in β-vae,

    C. P. Burgess, I. Higgins, A. Pal, L. Matthey, N. Watters, G. Desjardins, and A. Lerchner, “Understanding disentangling in β-vae,”arXiv preprint arXiv:1804.03599, 2018

    2018 arXiv

  31. [39]

    Adam: A method for stochastic optimization,

    D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” CoRR, vol. abs/1412.6980, 2015

    2015 arXiv

  32. [40]

    Pytorch: An imperative style, high- performance deep learning library,

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Desmaison, A. K ¨opf, E. Yang, Z. DeVito, M. Raison, A. Tejani, S. Chilamkurthy, B. Steiner, L. Fang, J. Bai, and S. Chintala, “Pytorch: An imperative style, high-...

    2019

  33. [41]

    A primer on kernel methods,

    J.-P. Vert, K. Tsuda, and B. Sch ¨olkopf, “A primer on kernel methods,” Kernel methods in computational biology, vol. 47, pp. 35–70, 2004

    2004

  34. [42]

    Explaining machine learning classifiers through diverse counterfactual explanations,

    R. K. Mothilal, A. Sharma, and C. Tan, “Explaining machine learning classifiers through diverse counterfactual explanations,” inProceedings of the 2020 conference on fairness, accountability, and transparency, 2020, pp. 607–617

    2020

  35. [43]

    Gan- terfactual—counterfactual explanations for medical non-experts using generative adversarial learning,

    S. Mertes, T. Huber, K. Weitz, A. Heimerl, and E. Andr ´e, “Gan- terfactual—counterfactual explanations for medical non-experts using generative adversarial learning,”Frontiers in artificial intelligence, vol. 5, p. 825565, 2022

    2022

  36. [44]

    Diffusion models for counterfactual explanations,

    G. Jeanneret, L. Simon, and F. Jurie, “Diffusion models for counterfactual explanations,” inProceedings of the Asian conference on computer vision, 2022, pp. 858–876

    2022

Pith tools

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