Pith. sign in

REVIEW 4 major objections 6 minor 45 references

LDM-Morph: Latent diffusion model guided deformable image registration

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read LDM-Morph claims that injecting latent diffusion features into a dual-stream encoder with cross-attention and a pixel-plus-latent similarity loss outperforms existing CNN- and Transformer-based deformable registration methods on four…

desk verdict Solid LDM-based registration recipe, but the accuracy claim over Transformers is statistically shaky; the folding gains are the stronger result. read the letter →

arxiv 2411.15426 v1 pith:NV6XSNUS submitted 2024-11-23 cs.CV

Jiong Wu , Kuang Gong This is my paper · ORCID
classification cs.CV
keywords deformableimageregistrationlatentdiffusionmodelcross-attentionunsupervisedlearningcardiacimagingtopologypreservationhierarchicalsimilaritylossSwinTransformer
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 tries to establish that deformable image registration improves when the network is given high-level semantic features from a pre-trained latent diffusion model, not just pixel-level features. To do this, it proposes LDM-Morph, an unsupervised framework with three components: an LDM-based feature extractor, a latent-and-global cross-attention module that fuses LDM features with Swin-style global features, and a hierarchical similarity loss evaluated in both pixel and latent space. The claim is that this combination yields higher Dice similarity coefficients and fewer deformation folds than the compared CNN-, Transformer-, and diffusion-based registration methods on four 2D cardiac datasets. If true, the paper would show that semantic information from generative models can be a practical ingredient in registration, improving both accuracy and topology preservation.

What carries the argument

The key machinery is the LDM-based latent feature extraction module, which uses a frozen pre-trained latent diffusion model to encode moving and fixed images into latent variables via DDIM inversion and then extracts features from four convolutional layers of the denoising U-Net. These latent features are fed into the upper stream of a dual-stream encoder, while a lower stream computes Swin-style windowed self-attention; the LGCA module then performs window-based multi-head cross-attention by swapping queries between the two streams. The loss combines pixel-space MSE with latent-space MSE computed through the LDM encoder, weighted by beta = 0.6, plus a gradient regularizer.

What would settle it

A reader could retrain the LDM on only the training splits of the four datasets, or use an LDM pre-trained on a completely unrelated image collection, and then rerun the registration experiments; if the Dice advantage over TransMatch disappears or the folding rate rises, the central claim is falsified.

Watch

Extended reading notes

Core claim

The central claim is that LDM-Morph outperformed existing state-of-the-art CNN- and Transformer-based registration methods in accuracy and topology preservation on the CAMUS-2CH, CAMUS-4CH, EchoNet-Dynamic, and ACDC datasets. The paper reports average Dice values of 0.882, 0.889, 0.883, and 0.850 respectively, with lower percentages of non-positive Jacobian determinants than most learned baselines, and better cross-dataset generalization when trained on EchoNet and tested on CAMUS. The authors attribute this to the LDM features enriching semantic information, the LGCA module enhancing interaction between latent and global features, and the latent-space similarity term reducing folding.

Load-bearing premise

The load-bearing premise is that the pre-trained LDM encoder provides meaningful and generalizable semantic features for cardiac images, which is only true if the LDM was not effectively trained on the same images used for evaluation.

Editorial extensions

If this is right

  • Using semantic features from a pre-trained latent diffusion model can improve deformable registration accuracy over CNN- and Transformer-only encoders.
  • Adding a latent-space similarity term to the loss reduces deformation folding, as measured by non-positive Jacobian determinants.
  • The framework generalizes across cardiac datasets when trained on one dataset and tested on another.
  • The improvements are incremental but consistent: roughly 1–3% higher Dice than Transformer baselines and larger margins over CNN and diffusion baselines.
  • The ablation study suggests that each added component—LDM features, LGCA, and the latent loss—contributes a measurable gain.

Reading between the lines

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

  • The paper does not specify which images were used to pre-train the LDM; if the same cardiac datasets were used, including test pairs, the reported gains may partly reflect memorization rather than generalizable semantic features.
  • The hierarchical loss behaves like a perceptual regularizer, so a similar benefit might appear in other modalities with low contrast or high noise, such as ultrasound or low-dose CT.
  • The method could likely be made end-to-end trainable by fine-tuning the LDM encoder, which the authors list as a limitation; this might yield larger gains but would also change the unsupervised setup.
  • A direct 3D extension is nontrivial because the LDM features are extracted from a 2D architecture, but applying the same idea with a 3D autoencoder is a plausible testable direction.
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 / 6 minor

Summary. The paper proposes LDM-Morph, an unsupervised deformable registration method that combines features from a frozen latent diffusion model (LDM) with a Swin-style global feature stream through a latent-global cross-attention (LGCA) module, and trains with a hierarchical similarity loss evaluated in both pixel space and LDM latent space. The method is evaluated on four public 2D cardiac datasets (CAMUS-2CH, CAMUS-4CH, EchoNet-Dynamic, ACDC) against two classical algorithms and five deep-learning baselines, reporting Dice similarity coefficients, percentages of non-positive Jacobian determinants, and runtime. The authors claim state-of-the-art accuracy and topology preservation with comparable computational cost, and they include an ablation study plus a cross-dataset generalization experiment.

Significance. If the claims are substantiated, the paper would make a useful contribution: it shows that semantic features from a pretrained LDM can be injected into a registration network and used in a dual-space loss, with public code and experiments across multiple cardiac datasets. The strengths are the breadth of the evaluation, the inclusion of topology metrics, the cross-dataset generalization test, and the ablation isolating the three proposed components. However, the central 'outperformed' claim currently rests on mean DSC differences of about 1 percentage point over the strongest Transformer baselines, and the paper lacks significance testing; the method description also contains a concrete loss-formula inconsistency and leaves the LDM pretraining data unspecified. These issues are fixable, but they are load-bearing for the paper's headline claims.

major comments (4)
  1. [Sec. 5.1, Table 1] The abstract and Section 5.1 claim that LDM-Morph 'outperformed' the Transformer-based baselines in registration accuracy, but the reported mean DSC advantages over TransMorph/TransMatch are 0.011 on CAMUS-2CH (0.882 vs 0.871), 0.013 on CAMUS-4CH (0.889 vs 0.876), and 0.008 on ACDC (0.850 vs 0.842), with standard deviations around 0.05–0.08 and test sizes of 80, 80, and 50. No significance tests are reported, and on these three datasets the gaps are of the same order as the sampling noise; only the ECHO comparison (0.883 vs 0.874, n=1,276) is clearly beyond noise. Please provide paired significance tests, with multiple-comparison correction, for both DSC and the |Jφ|≤0 metric, or soften the accuracy claim to a 'comparable accuracy with improved topology preservation' claim. As it stands, the accuracy half of the headline claim is not established by the reported statistics.
  2. [Sec. 5.3, Fig. 7 and Sec. 4.2] The balance weight β in Eq. (13) is selected using the test sets. Figure 7 plots DSC and |Jφ|≤0 on the four testing datasets for β from 0.2 to 1.0, and Section 4.2 then fixes β=0.6 'for all the experiments.' This means the final test-set numbers in Table 1 are the product of test-set hyperparameter tuning, which biases the comparison in favor of LDM-Morph. Select β on the validation folds (or via nested cross-validation), report the chosen value separately for each dataset, and freeze it before evaluating on the test sets. The observation that CAMUS-4CH has its best DSC at β=0.4 while 0.6 is used globally further highlights the need for a principled selection procedure.
  3. [Secs. 3.2 and 4.2] The pretraining data for the LDM are never specified. Section 3.2 states that 'an LDM was pretrained based on the image pairs as described below,' and Section 4.2 only gives the architecture (VQGAN plus a U-Net with channel size 64). If the LDM was trained on the same four datasets, including their test folds, then the hierarchical loss and the cross-dataset generalization experiment in Section 5.2 are not independent, and the reported gains could partly reflect the LDM having memorized the test distribution. Please disclose the exact pretraining dataset(s), the train/validation/test split used for the LDM, and confirm that the LDM training data do not overlap the test sets used in Tables 1 and 2.
  4. [Sec. 3.4, Eqs. (13)–(15)] The hierarchical similarity loss is double-weighted as written. Eq. (14) defines L_org as (1/|Ω|) Σ_p β [M∘φ^{-1}(p)−F(p)]^2, and Eq. (15) defines L_lat as (1/|Ω|) Σ_p (1−β) [E(M∘φ^{-1}(p))−E(F(p))]^2, while Eq. (13) then forms β L_org + (1−β) L_lat. The result is β^2 times the pixel MSE plus (1−β)^2 times the latent MSE, not the stated convex combination. This changes the interpretation of β in Figure 7 and Section 4.2. Please remove the inner β and (1−β) factors (or redefine L_org and L_lat as unweighted means), and state explicitly which form was actually implemented in the experiments.
minor comments (6)
  1. [Sec. 3.2, Eq. (5)] The posterior mean formula is inconsistent with the notation: α_t is defined in Eq. (2) as the cumulative product ∏_{s=1}^t (1−β_s), whereas the standard DDPM mean uses 1/√α_t with α_t = 1−β_t. Please correct the formula or align the notation with the standard DDPM expression.
  2. [Sec. 3.4, Eq. (16)] The smoothness term is an ordinary L2 gradient penalty on the deformation field; describing it as a 'diffusion-based regularizer' is misleading because no diffusion process is involved in that term.
  3. [Sec. 3.2] The feature-set notation '{ f i M L fF i|i = 1, 2,..., n }' is malformed and should be written as {f^i_{M_L}, f^i_F}; in the same passage, 'Marhov Chain' should be 'Markov chain.'
  4. [Table 3] The ablation rows are identified only by check marks, leaving the first row's configuration (the baseline without any of the three modules) implicit. Add explicit configuration labels to each row.
  5. [Sec. 5.1] The statement that LDM-Morph achieved the 'lowest standard deviation in DSC values' is true from Table 1, but the differences in standard deviations are very small (0.001–0.003); please avoid overinterpreting these differences without a formal variance comparison.
  6. [Eq. (8) and surrounding text] In the sentence defining self-attention, 'matrics' should be 'matrices.'

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the registration pipeline is trained and evaluated against external public datasets and reimplemented baselines, and the frozen LDM feature extractor and hierarchical loss do not reduce the reported gains to the training objective by construction.

full rationale

The paper's central claim is an empirical comparison rather than a first-principles derivation, and none of the load-bearing components reduces to its own inputs. The LDM-FE module is a frozen encoder: equations (1)-(6) describe the standard latent diffusion pretraining, and Section 3.2 states that after training the parameters of the encoder and U-Net are frozen, so the extracted features are not fitted to the registration loss or to the reported Dice/Jacobian metrics. The LGCA module is a cross-attention fusion defined by equations (9)-(11) that mixes latent and global features without presupposing the outcome. The hierarchical loss in equations (13)-(15) is a weighted sum of pixel-space MSE and latent-space MSE using the same frozen encoder E, but that is a deliberate architectural choice rather than a definitional equivalence between the loss and the evaluation metrics. The ablations in Table 3 show incremental gains from LDM-FE, LGCA, and Llat, and the comparisons in Tables 1-2 are against public datasets and reimplemented external baselines, so the evaluation is self-contained. The self-citations to Wu et al. and Gong et al. appear only in related-work contexts and are not load-bearing. The only ambiguity is that Section 3.2 says 'an LDM was pretrained based on the image pairs' while Section 4.2 specifies only the LDM architecture as in Rombach et al. without stating the pretraining corpus; if the LDM had seen the test images, the latent-similarity term would be a weaker independent measure. That is an unstated data-hygiene detail, not a circularity demonstrated by the paper's own equations, so the circularity score is 0.

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

The central claim rests on the quality and transferability of the LDM features, the correctness of the hierarchical loss formulation, and the independence of the pre-training data. These are not free parameters, but they are domain assumptions the paper does not validate.

free parameters (3)
  • beta (loss weight between pixel and latent terms) = 0.6
    Chosen by sweeping beta on the test sets (Sec. 5.3, Fig. 7); this is a test-set-selected hyperparameter.
  • lambda (smoothness weight) = 0.01
    Set to 0.01 for all experiments (Sec. 4.2), not reported as tuned.
  • Diffusion timestep t for feature extraction = 1
    Features extracted at t=1, described as empirically selected (Sec. 4.2).
assumptions (3)
  • domain assumption A pre-trained LDM (VQGAN plus U-Net) provides semantically meaningful latent features and a suitable similarity space for cardiac images.
    The method relies on the LDM encoder E to compute L_lat and on U-Net features to boost registration; no evidence is given that these features are superior to other self-supervised features on this data.
  • domain assumption The LDM used for feature extraction is trained on data that do not include the test images, or if it does, the effect on evaluation is negligible.
    The paper does not specify which datasets were used for LDM pre-training, so the generalization experiment and evaluation could be influenced by pre-training on test domains.
  • domain assumption DDIM inversion with t=1 steps yields latent codes that preserve the spatial structure needed for registration.
    The method assumes that latent variables at t=1 (nearly noiseless) still provide useful features; this is plausible but not justified with analysis.

how reviews work

0 comments
Cite this review

Pith. "Pith review of LDM-Morph: Latent diffusion model guided deformable image registration." pith.science (2026). https://pith.science/paper/NV6XSNUS

@misc{pith2026241115426,
  author       = {Pith},
  title        = {Pith review of: LDM-Morph: Latent diffusion model guided deformable image registration},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NV6XSNUS}},
  note         = {Machine review of arXiv:2411.15426}
}
read the original abstract

Deformable image registration plays an essential role in various medical image tasks. Existing deep learning-based deformable registration frameworks primarily utilize convolutional neural networks (CNNs) or Transformers to learn features to predict the deformations. However, the lack of semantic information in the learned features limits the registration performance. Furthermore, the similarity metric of the loss function is often evaluated only in the pixel space, which ignores the matching of high-level anatomical features and can lead to deformation folding. To address these issues, in this work, we proposed LDM-Morph, an unsupervised deformable registration algorithm for medical image registration. LDM-Morph integrated features extracted from the latent diffusion model (LDM) to enrich the semantic information. Additionally, a latent and global feature-based cross-attention module (LGCA) was designed to enhance the interaction of semantic information from LDM and global information from multi-head self-attention operations. Finally, a hierarchical metric was proposed to evaluate the similarity of image pairs in both the original pixel space and latent-feature space, enhancing topology preservation while improving registration accuracy. Extensive experiments on four public 2D cardiac image datasets show that the proposed LDM-Morph framework outperformed existing state-of-the-art CNNs- and Transformers-based registration methods regarding accuracy and topology preservation with comparable computational efficiency. Our code is publicly available at https://github.com/wujiong-hub/LDM-Morph.

Figures

Figures reproduced from arXiv: 2411.15426 by the authors.

Figure 1
Figure 1. Overview of the proposed deformable registration framework (LDM-Morph). It comprised three main components: a latent di [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. The architecture of the global feature extraction module. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. The architecture of the proposed latent and global feature cross [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Boxplot of DSC values for registration results on the CAMUS-2CH, CAMUS-4CH, ECHO, and ACDC datasets obtained from di [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Visual comparison of registration methods on the ECHO dataset. The first column displays: the fixed image with segmentation overlay, the moving [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Visual comparison of different registration methods on the CAMUS-2CH dataset. The first column displays: the fixed image with segmentation overlay, the moving image with segmentation overlay, and the difference between the moving and fixed images. Subsequent columns pr…
Figure 7
Figure 7. Figure 7: Average DSC values and corresponding average number of pixels with [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

45 extracted references · 36 canonical work pages

  1. [1]

    , author Rahman, S.U

    author Alam, F. , author Rahman, S.U. , author Ullah, S. , author Gulati, K. , year 2018 . title Medical image registration in image guided surgery: Issues, challenges and research opportunities . journal Biocybernetics and Biomedical Engineering volume 38 , pages 71--89

    work page 2018

  2. [2]

    , author Epstein, C.L

    author Avants, B.B. , author Epstein, C.L. , author Grossman, M. , author Gee, J.C. , year 2008 . title Symmetric diffeomorphic image registration with cross-correlation: evaluating automated labeling of elderly and neurodegenerative brain . journal Medical image analysis volume 12 , pages 26--41

    work page 2008

  3. [3]

    , author Zhao, A

    author Balakrishnan, G. , author Zhao, A. , author Sabuncu, M.R. , author Guttag, J. , author Dalca, A.V. , year 2019 . title Voxelmorph: a learning framework for deformable medical image registration . journal IEEE transactions on medical imaging volume 38 , pages 1788--1800

    work page 2019

  4. [4]

    , author Miller, M.I

    author Beg, M.F. , author Miller, M.I. , author Trouv \'e , A. , author Younes, L. , year 2005 . title Computing large deformation metric mappings via geodesic flows of diffeomorphisms . journal International journal of computer vision volume 61 , pages 139--157

    work page 2005

  5. [5]

    , author Lalande, A

    author Bernard, O. , author Lalande, A. , author Zotti, C. , author Cervenansky, F. , author Yang, X. , author Heng, P.A. , author Cetin, I. , author Lekadir, K. , author Camara, O. , author Ballester, M.A.G. , et al., year 2018 . title Deep learning techniques for automatic mri cardiac multi-structures segmentation and diagnosis: is the problem solved? j...

    2018

  6. [6]

    , author Fan, Q

    author Chen, C.F.R. , author Fan, Q. , author Panda, R. , year 2021 . title Crossvit: Cross-attention multi-scale vision transformer for image classification , in: booktitle Proceedings of the IEEE/CVF international conference on computer vision , pp. pages 357--366

    work page 2021

  7. [7]

    , author Frey, E.C

    author Chen, J. , author Frey, E.C. , author He, Y. , author Segars, W.P. , author Li, Y. , author Du, Y. , year 2022 . title Transmorph: Transformer for unsupervised medical image registration . journal Medical image analysis volume 82 , pages 102615

    work page 2022

  8. [8]

    , author Zheng, Y

    author Chen, Z. , author Zheng, Y. , author Gee, J.C. , year 2023 . title Transmatch: A transformer-based multilevel dual-stream feature matching network for unsupervised deformable image registration . journal IEEE transactions on medical imaging volume 43 , pages 15--27

    work page 2023

Show all 45 references
  1. [9]

    , author Yang, H

    author Cui, C. , author Yang, H. , author Wang, Y. , author Zhao, S. , author Asad, Z. , author Coburn, L.A. , author Wilson, K.T. , author Landman, B.A. , author Huo, Y. , year 2023 . title Deep multimodal fusion of image and non-image data in disease diagnosis and prognosis:...

    2023

  2. [10]

    , author Berendsen, F.F

    author De Vos, B.D. , author Berendsen, F.F. , author Viergever, M.A. , author Sokooti, H. , author Staring, M. , author I s gum, I. , year 2019 . title A deep learning framework for unsupervised affine and deformable image registration . journal Medical image analysis volume ...

    2019

  3. [11]

    , author Bu, J

    author Ding, Y. , author Bu, J. , author Qin, Z. , author You, L. , author Cao, M. , author Qin, Z. , author Pang, M. , year 2024 . title C2fresmorph: A high-performance framework for unsupervised 2d medical image registration . journal Pattern Recognition volume 154 , pages 110615

    2024

  4. [12]

    , author Gong, K

    author Dong, Y. , author Gong, K. , year 2024 . title Head and neck tumor segmentation from [18f] f-fdg pet/ct images based on 3d diffusion model . journal Physics in Medicine and Biology

    2024

  5. [13]

    , author Rombach, R

    author Esser, P. , author Rombach, R. , author Ommer, B. , year 2021 . title Taming transformers for high-resolution image synthesis , in: booktitle Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pp. pages 12873--12883

    2021

  6. [14]

    , author Khateri, M

    author Ghahremani, M. , author Khateri, M. , author Jian, B. , author Wiestler, B. , author Adeli, E. , author Wachinger, C. , year 2024 . title H-vit: A hierarchical vision transformer for deformable image registration , in: booktitle Proceedings of the IEEE/CVF Conference on...

    2024

  7. [15]

    , author Johnson, K

    author Gong, K. , author Johnson, K. , author El Fakhri, G. , author Li, Q. , author Pan, T. , year 2024 . title Pet image denoising based on denoising diffusion probabilistic model . journal European Journal of Nuclear Medicine and Molecular Imaging volume 51 , pages 358--368

    2024

  8. [16]

    , author Dey, N

    author Gopalakrishnan, V. , author Dey, N. , author Golland, P. , year 2024 . title Intraoperative 2d/3d image registration via differentiable x-ray rendering , in: booktitle Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. pages 11662--11672

    2024

  9. [17]

    , author Hansen, L

    author Hering, A. , author Hansen, L. , author Mok, T.C. , author Chung, A.C. , author Siebert, H. , author H \"a ger, S. , author Lange, A. , author Kuckertz, S. , author Heldmann, S. , author Shao, W. , et al., year 2022 . title Learn2reg: comprehensive multi-task medical im...

    2022

  10. [18]

    , author Jain, A

    author Ho, J. , author Jain, A. , author Abbeel, P. , year 2020 . title Denoising diffusion probabilistic models . journal Advances in neural information processing systems volume 33 , pages 6840--6851

    2020

  11. [19]

    , author Sun, G

    author Huo, X. , author Sun, G. , author Tian, S. , author Wang, Y. , author Yu, L. , author Long, J. , author Zhang, W. , author Li, A. , year 2024 . title Hifuse: Hierarchical multi-scale feature fusion network for medical image classification . journal Biomedical Signal Pro...

    2024

  12. [20]

    , author Lois, C

    author Jang, S.I. , author Lois, C. , author Thibault, E. , author Becker, J.A. , author Dong, Y. , author Normandin, M.D. , author Price, J.C. , author Johnson, K.A. , author Fakhri, G.E. , author Gong, K. , year 2023 . title Taupetgen: Text-conditional tau pet image synthesi...

    2023 arXiv

  13. [21]

    , author Bartlett, J

    author Jia, X. , author Bartlett, J. , author Zhang, T. , author Lu, W. , author Qiu, Z. , author Duan, J. , year 2022 . title U-net vs transformer: Is u-net outdated in medical image registration? , in: booktitle International Workshop on Machine Learning in Medical Imaging ,...

    2022

  14. [22]

    , author Imran, M

    author Jiang, H. , author Imran, M. , author Ma, L. , author Zhang, T. , author Zhou, Y. , author Liang, M. , author Gong, K. , author Shao, W. , year 2024 . title Fast denoising diffusion probabilistic models for medical image-to-image generation . journal arXiv preprint arXi...

    2024 arXiv

  15. [23]

    , author Han, I

    author Kim, B. , author Han, I. , author Ye, J.C. , year 2022 . title Diffusemorph: Unsupervised deformable image registration using diffusion model , in: booktitle European conference on computer vision , organization Springer . pp. pages 347--364

    2022

  16. [24]

    , author Kim, D.H

    author Kim, B. , author Kim, D.H. , author Park, S.H. , author Kim, J. , author Lee, J.G. , author Ye, J.C. , year 2021 . title Cyclemorph: cycle consistent unsupervised deformable image registration . journal Medical image analysis volume 71 , pages 102036

    2021

  17. [25]

    , author Banda, S

    author Kuppala, K. , author Banda, S. , author Barige, T.R. , year 2020 . title An overview of deep learning methods for image registration with focus on feature-based approaches . journal International Journal of Image and Data Fusion volume 11 , pages 113--135

    2020

  18. [26]

    , author Smistad, E

    author Leclerc, S. , author Smistad, E. , author Pedrosa, J. , author stvik, A. , author Cervenansky, F. , author Espinosa, F. , author Espeland, T. , author Berg, E.A.R. , author Jodoin, P.M. , author Grenier, T. , et al., year 2019 . title Deep learning for segmentation usin...

    2019

  19. [27]

    , author Kooi, T

    author Litjens, G. , author Kooi, T. , author Bejnordi, B.E. , author Setio, A.A.A. , author Ciompi, F. , author Ghafoorian, M. , author Van Der Laak, J.A. , author Van Ginneken, B. , author S \'a nchez, C.I. , year 2017 . title A survey on deep learning in medical image analy...

    2017

  20. [28]

    , author Lin, Y

    author Liu, Z. , author Lin, Y. , author Cao, Y. , author Hu, H. , author Wei, Y. , author Zhang, Z. , author Lin, S. , author Guo, B. , year 2021 . title Swin transformer: Hierarchical vision transformer using shifted windows , in: booktitle Proceedings of the IEEE/CVF intern...

    2021

  21. [29]

    , author Xu, Y

    author Ma, M. , author Xu, Y. , author Song, L. , author Liu, G. , year 2022 . title Symmetric transformer-based network for unsupervised image registration . journal Knowledge-Based Systems volume 257 , pages 109959

    2022

  22. [30]

    , author Feng, D

    author Meng, M. , author Feng, D. , author Bi, L. , author Kim, J. , year 2024 . title Correlation-aware coarse-to-fine mlps for deformable medical image registration , in: booktitle Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. pages ...

    2024

  23. [31]

    , author Chung, A.C

    author Mok, T.C. , author Chung, A.C. , year 2020 . title Large deformation diffeomorphic image registration with laplacian pyramid networks , in: booktitle Medical Image Computing and Computer Assisted Intervention--MICCAI 2020: 23rd International Conference, Lima, Peru, Octo...

    2020

  24. [32]

    , author He, B

    author Ouyang, D. , author He, B. , author Ghorbani, A. , author Yuan, N. , author Ebinger, J. , author Langlotz, C.P. , author Heidenreich, P.A. , author Harrington, R.A. , author Liang, D.H. , author Ashley, E.A. , et al., year 2020 . title Video-based ai for beat-to-beat as...

    2020

  25. [33]

    , author Singh, B

    author Pnvr, K. , author Singh, B. , author Ghosh, P. , author Siddiquie, B. , author Jacobs, D. , year 2023 . title Ld-znet: A latent diffusion approach for text-based image segmentation , in: booktitle Proceedings of the IEEE/CVF International Conference on Computer Vision ,...

    2023

  26. [34]

    , author Li, X

    author Qin, Y. , author Li, X. , year 2023 . title Fsdiffreg: Feature-wise and score-wise diffusion-guided unsupervised deformable image registration for cardiac images , in: booktitle International Conference on Medical Image Computing and Computer-Assisted Intervention , org...

    2023

  27. [35]

    , author Wang, Z

    author Rawat, W. , author Wang, Z. , year 2017 . title Deep convolutional neural networks for image classification: A comprehensive review . journal Neural computation volume 29 , pages 2352--2449

    2017

  28. [36]

    , author Blattmann, A

    author Rombach, R. , author Blattmann, A. , author Lorenz, D. , author Esser, P. , author Ommer, B. , year 2022 . title High-resolution image synthesis with latent diffusion models , in: booktitle Proceedings of the IEEE/CVF conference on computer vision and pattern recognitio...

    2022

  29. [37]

    , author Yang, D

    author Tang, Y. , author Yang, D. , author Li, W. , author Roth, H.R. , author Landman, B. , author Xu, D. , author Nath, V. , author Hatamizadeh, A. , year 2022 . title Self-supervised pre-training of swin transformers for 3d medical image analysis , in: booktitle Proceedings...

    2022

  30. [38]

    , author Velec, M

    author Torchia, J. , author Velec, M. , year 2024 . title Deformable image registration for composite planned doses during adaptive radiation therapy . journal Journal of Medical Imaging and Radiation Sciences volume 55 , pages 82--90

    2024

  31. [39]

    , author Pennec, X

    author Vercauteren, T. , author Pennec, X. , author Perchant, A. , author Ayache, N. , year 2009 . title Diffeomorphic demons: Efficient non-parametric image registration . journal NeuroImage volume 45 , pages S61--S72

    2009

  32. [40]

    , author Li, H

    author Wu, J. , author Li, H. , author Fan, Y. , year 2024 . title Diffeomorphic image registration with bijective consistency , in: booktitle Medical Imaging 2024: Image Processing , organization SPIE . pp. pages 700--706

    2024

  33. [41]

    , author Tang, X

    author Wu, J. , author Tang, X. , year 2020 . title A large deformation diffeomorphic framework for fast brain image registration via parallel computing and optimization . journal Neuroinformatics volume 18 , pages 251--266

    2020

  34. [42]

    , author Jiang, P

    author Xu, L. , author Jiang, P. , author Tsui, T. , author Liu, J. , author Zhang, X. , author Yu, L. , author Niu, T. , year 2023 . title 4d-ct deformable image registration using unsupervised recursive cascaded full-resolution residual networks . journal Bioengineering & tr...

    2023

  35. [43]

    , author Fletcher, P.T

    author Zhang, M. , author Fletcher, P.T. , year 2015 . title Finite-dimensional lie algebras for fast diffeomorphic image registration , in: booktitle Information Processing in Medical Imaging: 24th International Conference, IPMI 2015, Sabhal Mor Ostaig, Isle of Skye, UK, June...

    2015

  36. [44]

    , author Stendahl, J.C

    author Zhang, X. , author Stendahl, J.C. , author Staib, L. , author Sinusas, A.J. , author Wong, A. , author Duncan, J.S. , year 2023 . title An adaptive correspondence scoring framework for unsupervised image registration of medical images . journal arXiv preprint arXiv:2312.00837

    2023 arXiv

  37. [45]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.