Pith. sign in

REVIEW 4 major objections 5 minor 52 references

Enhancing Low Back Pain Assessment with Diffusion Models for Lumbar Spine MRI Segmentation

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 ↗ Review JSON

Pith's one-line read A diffusion model matches nnU-Net on lumbar spine MRI segmentation and outperforms it on degenerated intervertebral discs.

desk verdict A competent, incremental diffusion-segmentation paper for lumbar spine MRI; the central comparison holds, but the abstract's degenerated-IVD claim outruns the evidence. read the letter →

arxiv 2608.04906 v1 pith:WRKNOFEQ submitted 2026-08-05 cs.CV

Maria Monzon , Thomas Iff , Ender Konukoglu , Catherine R. Jutzeler This is my paper · ORCID
classification cs.CV
keywords diffusionmodelslumbarspineMRIsemanticsegmentationlowbackpainnnU-Netuncertaintymapsintervertebraldiscspresegmentation
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 claims that a diffusion-based segmentation model, SpineSegDiff, can match or slightly beat the established nnU-Net baseline on lumbar spine MRI segmentation while also producing uncertainty heatmaps. The strongest gain is on intervertebral discs, where SpineSegDiff reports a Dice of 0.90 versus 0.84 for the 2D nnU-Net baseline on combined T1w and T2w data. If the claim holds, diffusion models become a practical alternative for contrast-independent spine segmentation in low back pain assessment, with the extra benefit of per-pixel uncertainty for clinical review. The authors frame the work as a new application of generative diffusion models to a pathology-heavy imaging setting, where conventional segmentation methods often struggle with distorted anatomy.

What carries the argument

The central machinery is a denoising diffusion probabilistic model adapted for segmentation: rather than starting from pure noise, SpineSegDiff takes nnU-Net's initial presegmentation, partially noises it with a cosine scheduler, and learns to recover the ground-truth mask directly over a shortened diffusion process. At inference, DDIM produces several samples at each of the last 15 timesteps, the samples are averaged, and the final mask is an exponentially time-weighted sum; the per-timestep softmax entropy, maximized across timesteps, forms the uncertainty heatmap. This design is what lets the paper claim both accuracy and uncertainty visualization from one model, with the presegmentation step cutting the diffusion time steps needed from 1000 to as few as 30.

What would settle it

Run SpineSegDiff on the 18 oblique MRI series that were excluded from the test set and on non-central slices from the SPIDER dataset; if Dice on these routine acquisitions falls well below the reported 0.90 to 0.92 range, the central-slice evaluation overstates real-world segmentation quality.

Watch

Extended reading notes

Core claim

SpineSegDiff is a 2D diffusion-based segmentation framework that directly predicts the segmentation mask for the central slice of sagittal lumbar spine MRI, conditioned on either T1- or T2-weighted contrast. It uses a dual-encoder U-shaped denoising network, a composite MSE plus Dice plus binary cross-entropy loss, and DDIM sampling with multiple stochastic samples combined by an exponentially time-weighted average. On the SPIDER dataset with 5-fold cross-validation, it achieves mean Dice comparable to nnU-Net across the spinal canal, vertebrae, and intervertebral discs, and notably higher IVD Dice (0.90 versus 0.84). The same stochastic sampling yields entropy-based uncertainty heatmaps that highlight boundaries of degenerated structures. The authors conclude that the model approaches state-of-the-art performance while adding uncertainty quantification, and that the presegmentation strategy, where nnU-Net produces an initial mask and the diffusion model refines it, reduces the number of diffusion steps needed to reach that accuracy.

Load-bearing premise

The evaluation assumes the central slice of each MRI is the informative slice for low back pain assessment and that excluding the 18 oblique MRI series from the test set does not bias the reported Dice scores.

Editorial extensions

If this is right

  • A single 2D diffusion model can segment both T1- and T2-weighted lumbar spine MRI scans without contrast-specific training, reaching mean Dice 0.913 on combined data.
  • The IVD Dice gap (0.90 versus 0.84 over the 2D nnU-Net baseline) suggests diffusion models handle degenerated disc morphology better than the CNN baseline.
  • Presegmentation with nnU-Net reduces the number of diffusion steps from 1000 to as low as 30 while keeping accuracy, which lowers the computational cost of diffusion-based segmentation.
  • The uncertainty heatmaps provide a qualitative quality-assurance signal for clinicians reviewing segmentation of pathological spines, particularly at structure boundaries.
  • The statistical analysis identifies spondylolisthesis and disc narrowing as the pathologies that most degrade SpineSegDiff's Dice, indicating where future improvement is still needed.

Reading between the lines

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

  • A natural next test is whether the uncertainty heatmaps quantitatively correlate with segmentation error per region; the paper shows qualitative examples but does not calibrate the uncertainty against errors.
  • Because the model is trained and evaluated only on the central slice, applying it to full-volume MRI would require a policy for off-center slices, and the 18 oblique series excluded from testing represent an untested real-world scenario.
  • The ablation without the image encoder (0.909 versus 0.913 mean Dice) suggests the added encoder contributes only modestly, implying the diffusion refinement itself carries most of the benefit, which could simplify future architectures.
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

4 major / 5 minor

Summary. This manuscript presents SpineSegDiff, a 2D conditional diffusion model for lumbar spine MRI segmentation of vertebrae, intervertebral discs, and spinal canal, designed to work on both T1w and T2w images. The method uses a dual-encoder U-Net denoiser that directly infers masks, a presegmentation strategy in which an nnU-Net provides an initial mask, a multi-sample DDIM inference with exponentially time-weighted averaging, and entropy-based uncertainty heatmaps. Using the public SPIDER dataset with 5-fold cross-validation, the authors report Dice scores for SpineSegDiff, ablations without the image encoder, Diff-UNet 2D, IISDM, and nnU-Net across T1w, T2w, and combined contrasts. They also analyze the effect of degenerative pathologies on segmentation Dice and include an appendix on 3D baselines. The central claim is that SpineSegDiff achieves performance comparable to nnU-Net, 'particularly in improving the identification of degenerated IVDs.'

Significance. If the advertised claims are fully supported, the paper would be a useful contribution to medical-image segmentation, demonstrating that a diffusion-based 2D model can match nnU-Net on multi-contrast lumbar spine MRI while also producing uncertainty maps. The authors provide reproducible code, use a public multicenter dataset, and evaluate several pathological conditions, which are notable strengths. However, the headline differentiator of the method, the 'particularly in improving the identification of degenerated IVDs' clause, is not actually tested with a stratified model-by-pathology comparison, and the main numeric comparisons lack paired significance tests. The methodological contribution is incremental, but the study is well within the scope of the venue and the limitations appear addressable rather than fatal.

major comments (4)
  1. [Abstract; Section 3.2, Table 1] The abstract's claim that SpineSegDiff is 'particularly' better at identifying degenerated IVDs is not supported by the evidence in the manuscript. Table 1 reports IVD Dice pooled over all discs, while Section 3.2.1 and Appendix B test the effect of pathologies within each model separately and never test a model-by-degeneration interaction. In fact, Figure 2 shows disc degeneration significantly reduces SpineSegDiff's IVD Dice (p = 0.0003), so degenerated IVDs are harder for SpineSegDiff, but no comparison shows that SpineSegDiff outperforms nnU-Net specifically on degenerated IVDs. Please add stratified Dice for degenerated versus non-degenerated IVDs for SpineSegDiff and nnU-Net (ideally per disc level), and test the interaction, e.g., with a mixed-effects model or a paired test stratified by degeneration status.
  2. [Table 1 versus Section 3.2] The numerical comparisons in Table 1 are reported as means with standard deviations, but no paired significance test is given for the key differences, such as IVD Dice 0.90 for SpineSegDiff versus 0.84 for nnU-Net on T1w+T2w. Since the same 5-fold splits are used for all models, a paired per-subject or per-fold Wilcoxon signed-rank test, or a bootstrap over patients, is needed before claims of improvement over nnU-Net are made. Without this, the safe claim is that the models are comparable, not that SpineSegDiff is better.
  3. [Section 3.1; Section 3.2] The evaluation protocol is restricted to the central slice of each volume, and 18 oblique MRI series were excluded from the evaluation set but retained for training. This creates a gap between the reported Dice scores and real-world performance on full volumes or oblique acquisitions. The authors should either report quantitative results on non-central and oblique slices, or explicitly restrict the scope of the claims to central, non-oblique slices; ideally, they should compare central-slice Dice with full-volume Dice on a subset to justify the central-slice proxy.
  4. [Section 2.1.2; Table 2] SpineSegDiff with presegmentation consumes the output of a pretrained nnU-Net as conditioning input, and the same nnU-Net serves as the baseline. The comparison is not circular because Dice is computed on the diffusion output, but the architecture comparison is partly self-referential: the reported gain may reflect the additional presegmentation information rather than the diffusion model alone. Table 2 only reports T = 0 as the non-diffusion baseline for the presegmentation pipeline; the main comparison would be more informative if it also included SpineSegDiff trained from random noise without the nnU-Net prior, so that the standalone diffusion contribution can be separated from the presegmentation contribution.
minor comments (5)
  1. [Equation (2)] Equation (2) contains a mismatched closing parenthesis in the norm expression; please correct the typo.
  2. [Equation (3)] In Equation (3), the notation \(\bar{x}_t\) is first defined as the average over S samples, then the same symbol is used in the weighted sum; please clarify the time index and the relationship between the sample-averaged prediction and the final weighted output, and state whether the weights are normalized.
  3. [Figure 2 caption] The Figure 2 caption says 'Significant differences (p < 0.005)' while Section 3.2.1 states the Benjamini-Hochberg correction is applied at alpha = 0.05; please align the caption with the text.
  4. [Appendix B.1; Table 5] The appendix claims Diff-UNet 3D is 'superior' on IVD segmentation based on a mDICE difference of 0.920 versus 0.917 and 0.917; no significance test is provided, so the wording should be softened to 'comparable' or supplemented with statistical testing.
  5. [Section 3.2.1; Appendix B.1] The pathology analyses in Figure 2 and Appendix B use patient-level t-tests on Dice scores aggregated over structures, which ignores multiple discs per patient and potential clustering; consider reporting level-specific analyses or cluster-robust tests.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the results are empirical benchmark measurements, and the presegmentation dependency on nnU-Net is a system-design feature rather than a fitted input renamed as prediction.

full rationale

The paper makes no derivation-from-first-principles claim; its results are empirical Dice measurements on held-out folds. The only architectural dependency that could look self-referential is the presegmentation stage (Section 2.1.2): 'SpineSegDiff takes this partially noised presegmentation as input.' Since the presegmenter is the same nnU-Net used as a baseline, SpineSegDiff contains the baseline as a component. However, the reported mDICE is computed from SpineSegDiff's own output after diffusion refinement, not from the nnU-Net input, and the comparison in Table 1 is between complete systems. This is an architectural property, not a case of the predicted quantity being defined as the fit. The timestep weighting alpha = T_s/2 in Eq. (3) is hand-set, not fitted to target Dice. No self-citations of the present authors are load-bearing; references to presegmentation (Guo et al.) and DiffU-Net (Xing et al.) are external. The abstract's 'particularly in improving degenerated IVDs' is under-supported by the unstratified Table 1, but under-support is an evidence gap, not circularity. Therefore no circular step is exhibited.

Assumptions & free parameters 5 free parameters · 6 assumptions · 0 invented entities

The central claim rests on standard diffusion model mathematics and on dataset, annotation, and evaluation assumptions inherited from the SPIDER benchmark. No new physical entities are introduced. The main hand-set choices are sampling hyperparameters and noise schedules, none of which are fit to the target Dice scores.

free parameters (5)
  • Diffusion timestep weighting decay rate alpha = alpha = T_s/2 = 5
    Introduced in Eq. (3) to set exponential weights over the last 10 sampling timesteps; no ablation or justification beyond assigning more weight to later timestep predictions.
  • Number of DDIM samples S = S = 15
    Used for averaging and uncertainty maps in Eq. (3) and Algorithm 1; chosen without sensitivity analysis.
  • Final sampling timesteps T_s = T_s = 10
    Number of timesteps over which weighted averaging is applied; chosen without ablation in the paper.
  • Training diffusion timesteps T = T = 1000
    Noise schedule length for training; standard DDPM choice, not fitted to data.
  • Linear noise schedule endpoints = beta_1 = 1e-4 to 0.02
    Noise variance schedule in Section 3.1; standard-ish but hand-set, with a cosine scheduler used for presegmentation partial noising.
assumptions (6)
  • standard math DDPM forward and reverse diffusion equations, Eq. (1) and Eq. (2), are valid for generating segmentation masks conditioned on MR images.
    The method is built directly on the cited DDPM and DDIM formulations.
  • standard math Entropy computed from softmax probabilities over four classes is a meaningful uncertainty signal for segmentation.
    Eq. (4) and Eq. (5) define the heatmaps; the paper does not calibrate or validate them against clinical judgments.
  • domain assumption The central slice of each 3D MRI volume is a sufficient proxy for lumbar spine assessment.
    Section 3.1 states 2D models were trained and evaluated only on the central slice.
  • domain assumption Excluding 18 oblique MRI series from evaluation does not bias the reported performance.
    Section 3.2 says oblique scans were excluded from the evaluation set but retained for training.
  • domain assumption Pathology labels, including Pfirrmann grading and disc degeneration, are reliable enough for subgroup t-tests.
    Section 3.2.1 and Appendix A.1 use these labels to compare Dice scores between pathology-present and pathology-absent patients.
  • domain assumption nnU-Net presegmentation provides a reasonable starting mask that the diffusion model can refine.
    Section 2.1.2 describes the two-stage pipeline where SpineSegDiff takes the nnU-Net output as a partially noised input.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Enhancing Low Back Pain Assessment with Diffusion Models for Lumbar Spine MRI Segmentation." pith.science (2026). https://pith.science/paper/WRKNOFEQ

@misc{pith2026260804906,
  author       = {Pith},
  title        = {Pith review of: Enhancing Low Back Pain Assessment with Diffusion Models for Lumbar Spine MRI Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WRKNOFEQ}},
  note         = {Machine review of arXiv:2608.04906}
}
read the original abstract

This study introduces a diffusion-based framework for robust and accurate semantic segmentation of lumbar spine MRI scans from patients with low back pain (LBP), regardless of whether the scans are T1- or T2-weighted. We compared with advanced models for segmenting vertebrae, intervertebral discs (IVDs), and spinal canal using the SPIDER dataset. The results showed that SpineSegDiff achieved a segmentation performance comparable to that of the state-of-the-art non-diffusion nnUnet, particularly in improving the identification of degenerated IVDs. In addition, the uncertainty maps generated by our model provide valuable insights for clinical review, enhancing the robustness and reliability of the segmentation results. The potential of diffusion models to enhance the diagnosis and management of LBP through more precise analysis of pathological spine MRI is underscored by our findings.

Figures

Figures reproduced from arXiv: 2608.04906 by the authors.

Figure 1
Figure 1. SpineSegDiff architecture overview (top): the 2D MRI scan (y) is concatenated with the partially noised mask to generate the segmentation image xT . The framework combines a multi-scale image encoder (IE) with a UNet-based denoising model. Segmentation inference (bottom) employs a multi-sample step-weighted sum, simultaneously generating uncertainty-based heatmaps at the time step t during the DDIM sampling: hˆ t = … view at source ↗
Figure 2
Figure 2. Statistical analysis of segmentation performance in the presence of specific spinal pathologies in each subplot, including modic changes, spondylolisthesis, disc her￾niation, disc narrowing, disc bulging, and disc degeneration. Significant differ￾ences (p < 0.005) identified via T-tests with Benjamini-Hochberg correction. 3.3. presegmentation Time Diffusion Steps To evaluate the effectiveness of the presegmentation … view at source ↗
Figure 3
Figure 3. The visual comparisons on segmentation results on the central slice produced by the evaluated baseline and diffusion models for the three anatomical structures: spinal canal (blue), vertebrae (green), and intervertebral discs (red), along with the uncertainty maps for SpineSegDiff, where regions of higher uncertainty are denoted by darker red hues. The examples highlight challenging pathological cases with advanced … view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: (a) Training pipeline with presegmentation where nnU-Net generates initial mask xpre from MRI input y, followed by partial noising to obtain xT for diffusion training. A.4. Spinsegdiff Sampling and Uncertainty Maps The computation of uncertainty maps in SpineSegDiff in…
Figure 5
Figure 5. Figure 5: Dice scores boxplot for IISDM (left) and nnUnet (right) [PITH_FULL_IMAGE:figures/full_fig_p016_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 46 canonical work pages

  1. [1]

    1986 , isbn =

    Lamport, Leslie , title =. 1986 , isbn =

    1986

  2. [2]

    Distilling the knowledge in a neural network , author=

  3. [3]

    Analysing Diffusion Segmentation for Medical Images , author =

    work page

  4. [4]

    Ho, Jonathan and Jain, Ajay and Abbeel, Pieter , year = 2020, month = 6, url =

    work page 2020

  5. [5]

    2022 , journal =

    Yeh, Lee Ren and Zhang, Yang and others , number =. 2022 , journal =

    work page 2022

  6. [6]

    and others , number =

    Isensee, Fabian and Jaeger, Paul F. and others , number =. 2020 , journal =. doi:10.1038/s41592-020-01008-z , issn =

    doi:10.1038/s41592-020-01008-z 2020

  7. [7]

    2023 , journal =

    Matheve, Thomas and others , number =. 2023 , journal =

    work page 2023

  8. [8]

    Accelerating diffusion models via pre-segmentation diffusion sampling for medical image segmentation , year =

    Guo, Xutao and Yang, Yanwu and others , arxivId =. Accelerating diffusion models via pre-segmentation diffusion sampling for medical image segmentation , year =

    work page

Show all 52 references
  1. [9]

    and others , number =

    Fourney, Daryl R. and others , number =. 2011 , journal =

    2011

  2. [10]

    Sohl-Dickstein, Jascha and others , arxivId =

  3. [11]

    Ho, Jonathan and Jain, Ajay and Abbeel, Pieter , url =

  4. [12]

    Dhariwal, Prafulla and Nichol, Alex , arxivId =

  5. [13]

    2311.01797 , archivePrefix=

    Puheng Li and Zhong Li and Huishuai Zhang and Jiang Bian , year=. 2311.01797 , archivePrefix=

    arXiv

  6. [14]

    Spine X-ray Image Segmentation Based on Transformer and Adaptive Optimized Postprocessing , year=

    Zhang, Lingrong and others , booktitle=. Spine X-ray Image Segmentation Based on Transformer and Adaptive Optimized Postprocessing , year=

  7. [15]

    Zheng, Hua-Dong and Sun, Yue-Li and others , title=. Nat. Comm. , year=

  8. [16]

    EG-Trans3DUNet: A Single-Staged Transformer-Based Model for Accurate Vertebrae Segmentation from Spinal Ct Images , year=

    You, Xin and others , booktitle=. EG-Trans3DUNet: A Single-Staged Transformer-Based Model for Accurate Vertebrae Segmentation from Spinal Ct Images , year=

  9. [17]

    The Eleventh International Conference on Learning Representations , year=

    Diffusion Adversarial Representation Learning for Self-supervised Vessel Segmentation , author=. The Eleventh International Conference on Learning Representations , year=

  10. [18]

    Applied Sciences , VOLUME =

    Liu, Zengxin and Ma, Caiwen and She, Wenji and Xie, Meilin , TITLE =. Applied Sciences , VOLUME =. 2024 , NUMBER =

    2024

  11. [19]

    MICCAI , year=

    Zhao, Jianfeng and Li, Shuo , title=. MICCAI , year=

  12. [20]

    IEEE Transactions on Medical Imaging , volume=

    MR image denoising and super-resolution using regularized reverse diffusion , author=. IEEE Transactions on Medical Imaging , volume=. 2022 , publisher=

    2022

  13. [21]

    Medical image analysis , year=

    Diffusion models in medical imaging: A comprehensive survey , author=. Medical image analysis , year=

  14. [22]

    Spine Segmentation with Multi-view GCN and Boundary Constraint , year=

    Wang, Dexu and others , booktitle=. Spine Segmentation with Multi-view GCN and Boundary Constraint , year=

  15. [23]

    2024 , archivePrefix=

    Analysing Diffusion Segmentation for Medical Images , author=. 2024 , archivePrefix=

    2024

  16. [24]

    2023 , eprint=

    From CNN to Transformer: A Review of Medical Image Segmentation Models , author=. 2023 , eprint=

    2023

  17. [25]

    2020 , journal =

    Windsor, Rhydian and Jamaludin, Amir and others , month =. 2020 , journal =. doi:10.1007/978-3-030-59725-2_69 , issn =

    2020 doi

  18. [26]

    2019 , author =

    A gentle introduction to deep learning in medical image processing , journal =. 2019 , author =

    2019

  19. [27]

    2021 , author =

    Artificial intelligence and computer vision in low back pain: A systematic review , journal =. 2021 , author =

    2021

  20. [28]

    2019 , author =

    Artificial intelligence and machine learning in spine research , journal =. 2019 , author =

    2019

  21. [29]

    Li and others , title =

    H. Li and others , title =. Neural Computing and Applications , volume =

  22. [30]

    VerSe: A Vertebrae labelling and segmentation benchmark for multi-detector CT images , journal =

    Anjany Sekuboyina and others , keywords =. VerSe: A Vertebrae labelling and segmentation benchmark for multi-detector CT images , journal =. 2021 , issn =

    2021

  23. [31]

    2023 , author =

    Deep Learning for Multi-Tissue Segmentation and Fully Automatic Personalized Biomechanical Models from BACPAC Clinical Lumbar Spine MRI , journal =. 2023 , author =

    2023

  24. [32]

    2018 , author =

    DeepSPINE: Automated Lumbar Vertebral Segmentation, Disc-level Designation, and Spinal Stenosis Grading Using Deep Learning , journal =. 2018 , author =

    2018

  25. [33]

    Ho and A

    J. Ho and A. Jain and P. Abbeel , url =. Denoising Diffusion Probabilistic Models , year =

  26. [34]

    Xing and L

    Z. Xing and L. Wan and others , url =. Diff-UNet: A Diffusion Embedded Network for Volumetric Segmentation , year =

  27. [35]

    2000 , journal =

    Manchikanti, L , number =. 2000 , journal =

    2000

  28. [36]

    2006 , author =

    Does back pain prevalence really decrease with increasing age? A systematic review , journal =. 2006 , author =

    2006

  29. [37]

    and others , number =

    Ferreira, Manuela L. and others , number =. 2023 , journal =

    2023

  30. [38]

    2017 , author =

    Automation of reading of radiological features from magnetic resonance images (MRIs) of the lumbar spine without human intervention is comparable with an expert radiologist , journal =. 2017 , author =

    2017

  31. [39]

    2019 , author =

    Iterative fully convolutional neural networks for automatic vertebra segmentation and identification , journal =. 2019 , author =

    2019

  32. [40]

    2022 , journal =

    Mushtaq, Malaika and Akram, Muhammad Usman and others , number =. 2022 , journal =

    2022

  33. [41]

    Wu and R

    J. Wu and R. Fu and others , arxivId =. MedSegDiff: Medical Image Segmentation with Diffusion Probabilistic Model , year =

  34. [42]

    2023 , author =

    MRI image features with an evident relation to low back pain: a narrative review , journal =. 2023 , author =

    2023

  35. [43]

    2022 , author =

    Review of Deep Learning Models for Spine Segmentation , booktitle =. 2022 , author =

    2022

  36. [44]

    J. W. van der Graaf and M. L. van Hooff and others , doi =. Lumbar spine segmentation in MR images: a dataset and a public benchmark , journal =. 2024 , publisher =

    2024

  37. [45]

    2018 , author =

    Spine-GAN: Semantic segmentation of multiple spinal structures , journal =. 2018 , author =

    2018

  38. [46]

    2021 , eprint=

    SpineOne: A One-Stage Detection Framework for Degenerative Discs and Vertebrae , author=. 2021 , eprint=

    2021

  39. [47]

    2021 , author =

    SpineParseNet: Spine Parsing for Volumetric MR Image by a Two-Stage Segmentation Framework With Semantic Image Representation , journal =. 2021 , author =

    2021

  40. [48]

    Vertebrae detection and labelling in lumbar MR images , volume =

    Lootus, M and Kadir, T and Zisserman, A , year =. Vertebrae detection and labelling in lumbar MR images , volume =. Lecture Notes in Computational Vision and Biomechanics: Computational Methods and Clinical Applications for Spine Imaging , pages =

  41. [49]

    2018 , journal =

    Hartvigsen, Jan and Hancock, Mark J and others , number =. 2018 , journal =

    2018

  42. [50]

    2019 , url =

    Jeroen Bertels and others , title =. 2019 , url =

    2019

  43. [51]

    doi:10.1016/j.media.2023.102846 , issn =

    Kazerouni, Amirhossein and Aghdam, Ehsan Khodapanah and Heidari, Moein and Azad, Reza and Fayyaz, Mohsen and Hacihaliloglu, Ilker and others , year = 2023, month = 8, journal =. doi:10.1016/j.media.2023.102846 , issn =

    2023

  44. [52]

    2023 , eprint=

    On the Importance of Noise Scheduling for Diffusion Models , author=. 2023 , eprint=

    2023

Pith tools

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