The DOI matches its Crossref record; 77 of the 94 references it cites were found in a registry or an index; none of the 43 sources its claims cite is checked for support yet.
Quicksilver: Fast predictive image registration – A deep learning approach
Open accessAbstract
This paper introduces Quicksilver, a fast deformable image registration method. Quicksilver registration for image-pairs works by patch-wise prediction of a deformation model based directly on image appearance. A deep encoder-decoder network is used as the prediction model. While the prediction strategy is general, we focus on predictions for the Large Deformation Diffeomorphic Metric Mapping (LDDMM) model. Specifically, we predict the momentum-parameterization of LDDMM, which facilitates a patch-wise prediction strategy while maintaining the theoretical properties of LDDMM, such as guaranteed diffeomorphic mappings for sufficiently strong regularization. We also provide a probabilistic version of our prediction network which can be sampled during the testing time to calculate uncertainties in the predicted deformations. Finally, we introduce a new correction network which greatly increases the prediction accuracy of an already existing prediction network. We show experimental results for uni-modal atlas-to-image as well as uni-/multi-modal image-to-image registrations. These experiments demonstrate that our method accurately predicts registrations obtained by numerical optimization, is very fast, achieves state-of-the-art registration results on four standard validation datasets, and can jointly learn an image similarity measure. Quicksilver is freely available as an open-source software.
Overview
AI snapshot
Written from the paper’s full text.
TL;DR
Quicksilver is a deep-learning registration method that predicts the initial momentum of LDDMM shooting patch-by-patch from moving/target image patches, with optional Bayesian uncertainty and a correction network. It achieves registration accuracy close to numerical LDDMM optimization on several brain MRI datasets while running much faster.
Key findings
- Predicting the initial momentum gave better atlas-to-image accuracy than predicting initial velocity or displacement; momentum and velocity networks produced positive Jacobian determinants in all tested cases, whereas direct displacement prediction did not guarantee diffeomorphic maps.
- On the four Klein et al. validation datasets, the prediction+correction model (LPC) approached numerical LDDMM optimization (LO): paired TOST indicated LPC was statistically equivalent to LO within a 2% relative threshold on all four datasets, and label overlap was top-tier with few outliers.
- Runtime for a 229×193×193 volume: LDDMM optimization took about 10.8 minutes, while Quicksilver prediction took 7.63 s plus 8.9 s for shooting, totaling 18.43 s, a roughly 35× speedup; adding the correction network doubled time but remained an order of magnitude faster than optimization. Patch pruning and stride 14 reduced the number of predicted patches by 99.995% for such volumes.
- Multimodal T1w–T2w prediction reduced deformation error relative to affine registration and was only slightly worse than the T1w–T1w upper bound; training with only 10 images (90 registration pairs) still worked with a small performance drop. Bayesian sampling produced uncertainty maps with high values near ventricles and large deformations.
Methods
- The method uses the LDDMM shooting formulation and trains a deep encoder-decoder network to predict vector-valued initial momentum from paired 15×15×15 patches of moving and target images. Whole-image prediction uses a sliding window with stride 14 and patch pruning, followed by LDDMM shooting to generate deformations.
- Data: atlas-to-image used 150 OASIS images (100 training, 50 testing); image-to-image used 373 OASIS images for training and tested on LPBA40, IBSR18, MGH10, and CUMC12, totaling 2,168 test registrations; multimodal used the IBIS 3D dataset with 375 T1w/T2w images (359 training, 16 testing, 250 test cases). A correction network was trained sequentially on the residual between ground-truth and predicted momentum; a probabilistic version used dropout (probability 0.2) and Monte Carlo sampling.
- Evaluation included voxel-wise deformation error percentiles, Jacobian determinant positivity, LDDMM energy, target overlap of labels, paired t-tests, paired TOST equivalence tests, and runtime/multi-GPU scaling.
Limitations
- The paper states that the 15×15×15 patch receptive field limits prediction of very large deformations; OASIS training data had only 0.2% of voxels with deformations above 10 mm, so large deformations were underrepresented. Outliers with inconsistent anatomy can cause extreme optimization deformations that prediction may not match, though the conservative prediction can also avoid implausible warps.
- The correction network doubles runtime, and applying correction multiple times can produce more drastic deformations and numerical instability; 12 LPP cases in LPBA40 failed LDDMM shooting due to integration problems. Patch pruning may not transfer to parametric models if parameter locality is not guaranteed, and multi-GPU scaling is limited by communication overhead and PyTorch’s beta-stage performance.
- Evident from the text: validation is restricted to brain MRI with a specific LDDMM regularizer and SSD similarity, so generalization to other organs, modalities, or registration models is not demonstrated.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
0 of 43 cited sources checked · 0 supported · 0 flagged
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Finally, we introduce a new correction network which greatly increases the prediction accuracy of an already existing prediction network.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > Abstract
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- The authors state this as an assumption: A set of LDDMM parameters obtained by numerical optimization already exists and yields good registration results.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Our experiments show that our approach also predicts accurate deformations for multi-modal registration.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 1. Introduction
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- The authors state this as an assumption: A set of LDDMM parameters obtained by numerical optimization already exists and yields good registration results.
In the paper
“achieves state-of-the-art registration results on four standard validation datasets”
- Also stated in: § 1. Introduction · boosted
“demonstrate registration accuracies on these datasets on par with the state-of-the-art”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > Abstract
What the sources say
- Not checked yetcites [33] Kingma, 2014
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- The authors state this as an assumption: A set of LDDMM parameters obtained by numerical optimization already exists and yields good registration results.
- Also stated in: § 1. Introduction · boosted
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“These experiments demonstrate that our method accurately predicts registrations obtained by numerical optimization”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > Abstract
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- The authors state this as an assumption: A set of LDDMM parameters obtained by numerical optimization already exists and yields good registration results.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Experiments show that the correction network improves registration results to the point where optimization-based and predicted registrations achieve a similar level of registration accuracy on registration validation experiments.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 1. Introduction
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- The authors state this as an assumption: A set of LDDMM parameters obtained by numerical optimization already exists and yields good registration results.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“is very fast”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > Abstract
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- The authors state this as an assumption: A set of LDDMM parameters obtained by numerical optimization already exists and yields good registration results.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“and can jointly learn an image similarity measure”
- Also stated in: § 1. Introduction · plain
“our framework can simultaneously predict registrations and learn a multi-modal image similarity measure”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > Abstract
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- The authors state this as an assumption: A set of LDDMM parameters obtained by numerical optimization already exists and yields good registration results.
- Also stated in: § 1. Introduction · plain
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“For a typical 3D image of size 128 × 128 × 128 using a 15 × 15 × 15 patch for prediction will require more than 1.4 million patch predictions.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.2. Deep network for LDDMM prediction
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesA result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We observe that the initial momentum and the initial velocity networks indeed generate diffeomorphic deformations in all scenarios.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.1. Atlas-to-Image registration
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesThe paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Predicting the displacement directly cannot guarantee diffeomorphic deformations even for a small stride.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.1. Atlas-to-Image registration
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes
- It may state two separate claims as one. 46
- Probably not a substantial claim. 22
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We experimentally show that by using patch pruning and a large sliding window stride, our method achieves dramatic speedups compared to the optimization approach, while maintaining good registration accuracy.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 1. Introduction
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
- The authors state this as an assumption: A set of LDDMM parameters obtained by numerical optimization already exists and yields good registration results.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“the total number of 15 × 15 × 15 3D training patches to train the prediction network is 1,002,404. This amounts to approximately 3.4 billion voxels and is much larger than the total number of parameters in the network.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.2. Deep network for LDDMM prediction
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passesIt may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“The number of 3D filters used in the network is 975,360. The overall number of parameters 21,826,344.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.2. Deep network for LDDMM prediction
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Compared to the label overlapping accuracy boost from LP to LPC, LPC2 and LPC3 do not greatly improve the registration accuracy, and may even generate worse results (e.g., LPC3 on LPBA40).”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Load-bearingStrength:hedged· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Compared with the initial LDDMM energy based on affine registration to the atlas space in the initial column, both LP and LPC have drastically lower LDDMM energy values; further, these values are only slightly higher than those for LO.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“LPP, LPC2 and LPC3 generate more drastic deformations (i.e., more spread out histograms indicating areas of stronger expansions and contractions).”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“In fact, they produce more drastic deformations that can lead to worse label overlap and even numerical stability problems.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Our prediction result, however, still generates reasonable deformations (where plausibility is based on the deformations that were observed during training) while matching the brain structures as much as possible.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Using the correction network together with the prediction network doubles the computation time, but the overall runtime is still an order of magnitude faster than direct LDDMM optimization.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.4. Runtime study
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Note that, at a stride of 1, computational cost increases about 3000-fold in 3D, resulting in runtimes of about 5½ hours for 3D image registration (eleven hours when the correction network is also used).”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.4. Runtime study
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Using our prediction network with a sliding window stride of 14, the initial momentum prediction time is, on average, 7.63 seconds. Subsequent geodesic shooting to generate the deformation field takes 8.9 seconds, resulting in a total runtime of 18.43 seconds.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.4. Runtime study
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
In the paper
“For a relative threshold difference of 2% LPC can be considered statistically equivalent to LO for all four datasets and to many of the other top methods (e.g., LPC vs. SyN on MGH10 and IBSR18).”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
What the sources say
- Not checked yetcites [71] Zhang, 2015
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Among the prediction models, LPC improves the prediction accuracy compared with LP, and generally achieves the highest deformation prediction accuracy for up to 80% of the voxels.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The multi-modal networks (T1w-T2w, LP/LPC) significantly reduce deformation error compared to affine registration, and only suffer a slight loss in accuracy compared to their T1w-T1w counterparts.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.3. Multi-modal image registration
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Second, the deformations in the OASIS training images have a median of 2.609 mm, which is similar to the median observed in the four testing datasets. However, only 0.2% of the voxels in the OASIS training dataset have deformations larger than 10 mm.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“A slight performance drop with respect to LO can be observed for LPC and a slightly bigger performance drop for LP, which only uses the prediction model, but no correction model.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Furthermore, compared with LP, LPC generates LDDMM energy values that are closer to LO, which indicates that using the prediction+correction approach results in momenta which are closer to the optimal solution than the ones obtained by using the prediction network only.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“It is also on-par with other prediction models for up to 99% of the voxels, where the largest deformations are in the range between 7.317 mm-9.026 mm for the four datasets.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Compared with other methods, our prediction+correction network achieves top-tier performance for label matching accuracy at a small fraction of the computational cost.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Second, using only the prediction network results in a slight performance drop compared to the numerical optimization results (LO), but the result is still competitive with the top-performing registration methods.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“As shown in the table, the largest deformations that LDDMM optimization generates are 23.393 mm for LPBA40, 36.263 mm for IBSR18, 18.753 mm for CUMC12 and 18.727 mm for MGH10.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passesIt may state two separate claims as one.
In the paper
“It is based on Riemannian geometry, and hence induces a distance metric on the space of images.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 1. Introduction
What the sources say
- Not checked yetcites [17] Gal, 2016
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim. It may state two separate claims as one.
In the paper
“It uses the initial momentum as the registration parameter, which does not need to be spatially smooth, and hence can be predicted patch-by-patch, and from which the whole geodesic path can be computed.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 1. Introduction
What the sources say
- Not checked yetcites [17] Gal, 2016
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“the initial momentum is generally not smooth, but is compactly supported at image edges”
- Also stated in: § 1. Introduction · hedged
“the momentum is typically compactly supported around image edges”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.1. LDDMM Shooting
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes? Based: unclearIt may state two separate claims as one.
- Also stated in: § 1. Introduction · hedged
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“the initial velocity is generated by applying a smoothing kernel K to the initial momentum. Therefore, the smoothness of the deformation does not need to be specifically considered during the parameter prediction step, but is imposed after the prediction.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.1. LDDMM Shooting
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Hence, by predicting the momentum, we retain all the convenient mathematical properties of LDDMM and, at the same time, are able to predict diffeomorphic transformations fast.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 1. Introduction
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
In the paper
“We demonstrate that the momentum-parameterization for LDDMM shooting [17] is a convenient representation for regression approaches as (i) the momentum is typically compactly supported around image edges and (ii) there are no smoothness requirements on the momentum itself.”
- Also stated in: § Abstract · plain
“Specifically, we predict the momentum-parameterization of LDDMM, which facilitates a patch-wise prediction strategy while maintaining the theoretical properties of LDDMM, such as guaranteed diffeomorphic mappings for sufficiently strong regularization.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 1. Introduction
What the sources say
- Not checked yetcites [17] Gal, 2016
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
- Also stated in: § Abstract · plain
None of the sources cited for this claim have been checked yet.
In the paper
“General approaches address multimodal image registration by either performing image synthesis [22, 23] to change the problem to a uni-modal image registration task, or by proposing complex, hand-crafted [24, 25, 26, 27] or learned [28, 29, 30, 31, 32] multi-modal image similarity measures.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 1. Introduction
What the sources say
- Not checked yetcites [22] Gutierrez-Becker, 2017, Guiding multimodal registration with learned optimization updates, MedIA
- Not checked yetcites [23] Gutiérrez-Becker, 2016
- Not checked yetcites [24] Haber, 2007, Image registration with guaranteed displacement regularity, Int. J. Comput. Vis. source
- Not checked yetcites [25] Hart, 2009, An optimal control approach for deformable registration
- Not checked yetcites [26] Hazlett, 2017, Early brain development in infants at high risk for autism spectrum disorder, Nature source
- Not checked yetcites [27] He, 2015
- Not checked yetcites [28] Hermosillo, 2002, Variational methods for multimodal image matching, IJCV source
- Not checked yetcites [29] Hill, 2001, Medical image registration, Phys. Med. Biol. source
- Not checked yetcites [30] Hong, 2012
- Not checked yetcites [31] Horn, 1981, Determining optical flow, Artif. Intell. source
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Because M and T ○ Φ are in the same coordinate system, the differences between these two images are small as long as the predicted deformation is reasonable, and more accurate predictions can be expected.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.3. Correction network
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
None of the sources cited for this claim have been checked yet.
In the paper
“Gal et al. [58] show that ELBO maximization is achieved by training with dropout [57].”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.2. Deep network for LDDMM prediction
What the sources say
- Not checked yetcites [58] Vaillant, 2004, Statistics on diffeomorphisms via tangent space representations, NeuroImage source
- Not checked yetcites [57] Srivastava, 2014, Dropout: a simple way to prevent neural networks from overfitting, JMLR
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This is reasonable for our initial momentum parameterization because of the compact support (at edges) of the initial momentum and the spatial shift invariance we obtain via the pooling/unpooling operations.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.2. Deep network for LDDMM prediction
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
None of the sources cited for this claim have been checked yet.
In the paper
“for image-based LDDMM [17, 19, 45] the momentum can be written as m(x,t) = λ(x,t)∇I(x,t), where λ is a scalar field and ∇I is the spatial gradient of the image. Hence, for homogeneous areas, ∇I = 0 and consequentially m = 0.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.1. LDDMM Shooting
What the sources say
- Not checked yetcites [17] Gal, 2016
- Not checked yetcites [19] Grabner, 2006, Symmetric Atlasing and Model Based Segmentation: an Application to the hippocampus in Older Adults
- Not checked yetcites [45] Niethammer, 2011
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Multi-modal image registration is, in general, significantly more difficult than uni-modal image registration since image appearance can change drastically between different modalities.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 1. Introduction
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“These experiments indicated that increasing the training data size further only results in marginal improvements, which are clearly outperformed by a combined prediction + correction strategy.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.4. Datasets/Setup
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
In the paper
“recent research [56] suggests that the degrees of freedom for a deep network can be significantly smaller than the number of its parameters.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.2. Deep network for LDDMM prediction
What the sources say
- Not checked yetcites [56] Springenberg, 2014
Strength:hedged· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
None of the sources cited for this claim have been checked yet.
In the paper
“using dropout [57], which can be viewed as (Bernoulli) approximate inference in Bayesian neural networks [58, 59].”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.2. Deep network for LDDMM prediction
What the sources say
- Not checked yetcites [57] Srivastava, 2014, Dropout: a simple way to prevent neural networks from overfitting, JMLR
- Not checked yetcites [58] Vaillant, 2004, Statistics on diffeomorphisms via tangent space representations, NeuroImage source
- Not checked yetcites [59] Van Essen, 2013, WU-Minn HCP Consortium, the WU-Minn human connectome project: an overview, NeuroImage source
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We chose the 1-norm instead of the 2-norm as our loss function to be able to tolerate outliers and to generate sharper momentum predictions.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.2. Deep network for LDDMM prediction
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“given the initial momentum, we are able to generate the whole geodesic path using the geodesic shooting equations.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.3. Correction network
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✗ Based: failsProbably not a substantial claim. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Assuming no flips of the entire coordinate system, a diffeomorphic deformation map should have positive Jacobian determinants everywhere, otherwise foldings occur in the deformation maps.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.1. Atlas-to-Image registration
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim. It may state two separate claims as one.
None of the sources cited for this claim have been checked yet.
In the paper
“Gutierrez et al. [9] use a regression forest and gradient boosted trees [8] based on hand-crafted features to learn update steps for a rigid and a B-spline registration model.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 1. Introduction
What the sources say
- Not checked yetcites [9] Cao, 2014, Multi-modal registration for correlative microscopy using image analogies, MedIA
- Not checked yetcites [8] Bruveris, 2012, Mixture of kernels and iterated semidirect product of diffeomorphisms groups, Multiscale Model. Simul. source
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Wang et al. [11] predict deformations by key-point matching using sparse learning followed by dense deformation field generation with radial basis function interpolation. The performance of the method heavily depends on the accuracy of the key point selection.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 1. Introduction
What the sources say
- Not checked yetcites [11] Cheng, 2015, Deep similarity learning for multimodal medical images, Comput. Methods Biomech. Biomed. Eng. Imaging Vis.
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“the correction for the initial momentum is then performed in the original coordinate space (of the moving image) which allows us to obtain an overall corrected initial momentum, m0. This is for example a useful property when the goal is to do statistics with respect to a fixed coordinate system, for example, an atlas coordinate system.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.3. Correction network
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“While using larger patches would be an option, this would require a network with substantially larger capacity (to store the information for larger image patches and all meaningful deformations) and would also likely require much larger training datasets8.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.3. Correction network
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
In the paper
“In contrast to ReLU, PReLU avoids a zero gradient for negative inputs, effectively improving the network performance.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.2. Deep network for LDDMM prediction
What the sources say
- Not checked yetcites [52] Simpson, 2011
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“For testing, we keep the dropout layers enabled to maintain the probabilistic property of the network”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.2. Deep network for LDDMM prediction
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✗ Based: failsProbably not a substantial claim. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“That this is not the case for the momentum parameterization suggests that it is easier for the network to learn to predict the momentum, as it indeed has become more specialized to the training data which was obtained with a stride of 14.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.1. Atlas-to-Image registration
Strength:hedged· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passesIt may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“To address these shortcomings, we propose a two-step prediction approach to improve overall prediction accuracy.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.3. Correction network
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✗ Based: failsProbably not a substantial claim. The paper may give no reason why it holds.
In the paper
“the vector-valued momentum has been found to numerically better behave for LDDMM optimizations [50].”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 2. Materials and Methods > 2.2. Deep network for LDDMM prediction
What the sources say
- Not checked yetcites [50] Shams, 2010, A survey of medical image registration on multicore and the GPU, IEEE Signal Process. Mag. source
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“One of the important properties of LDDMM shooting is its ability to generate diffeomorphic deformations.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.1. Atlas-to-Image registration
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✗ Based: failsProbably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Hence, the initial momentum parameterization is the preferred choice among our three tested parameterizations as it achieves the best prediction accuracy and guarantees diffeomorphic deformations.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.1. Atlas-to-Image registration
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This is reasonable as these methods are all non-parametric diffeomorphic or homeomorphic registration methods, allowing the modeling of large deformations between image pairs.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The proposed correction network is a step towards highly accurate deformation prediction, while only decreasing the computation speed by a factor of 2.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 4. Discussion
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passesIt may state two separate claims as one.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“Note that among these methods, only LPC is specifically trained to match the LDDMM optimization results and in particular to predict corrections to the initial momentum obtained by the prediction model (LP) in the tangent space of the moving image. In contrast, LPP, LPC2 and LPC3 lack this theoretical motivation.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Using a smaller step-size would regain numerical stability in this case.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Tests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“However, note that, in general, a low energy will only imply a good registration result if the registration model is fully appropriate for the registration task.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✗ Based: failsProbably not a substantial claim. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“However, as the number of GPUs increases, the communication overhead between GPUs becomes larger which explains why computation time does not equal to 11.23/number of GPUs seconds.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.4. Runtime study
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Hence the initial momentum parameterization, which can tolerate large sliding window strides, is essential for fast deformation prediction with high accuracy while guaranteeing diffeomorphic deformations.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.4. Runtime study
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“These large deformations created by numerical LDDMM optimization are not always desirable; and consequentially registration errors of the prediction models with respect to the numerical optimization result are in fact sometimes preferred.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Moreover, the patch-wise prediction approach of our methods enables a multi-GPU implementation, further increasing the prediction speed.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 4. Discussion
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passesIt may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This can be regarded as an advantage of our network, where the conservative nature of patch-wise momentum prediction is more likely to generate reasonable deformations.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 3. Results > 3.2. Image-to-Image registration
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The proposed approach allows taking large strides for patch-wise prediction, without a substantial decrease in registration accuracy, resulting in fast and accurate deformation prediction.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 4. Discussion
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Our framework is very general and can be directly applied to many other registration techniques.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 4. Discussion
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Our method retains all theoretical properties of LDDMM and results in diffeomorphic transformations if appropriately regularized, but computes these transformations an order of magnitude faster than a GPU-based optimization for the LDDMM model.”
- Also stated in: § 4. Discussion · plain
“In effect, our Quicksilver registration approach converts a notoriously slow and memory-hungry registration approach to a fast method, while retaining all of its appealing mathematical properties.”
§ Quicksilver: Fast Predictive Image Registration – a Deep Learning Approach > 4. Discussion
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
- Also stated in: § 4. Discussion · plain
The authors state this as an assumption.
A set of LDDMM parameters obtained by numerical optimization already exists and yields good registration results.
“We assume that we already have a set of LDDMM parameters which result in good registration results.”
The paper relies on this without support.
The local variance of deformation fields generated from multiple dropout samples provides an uncertainty estimate of the predicted deformation field.
“The local variance of these deformation fields can then be used as an uncertainty estimate of the predicted deformation field.”
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Some of these approaches report interval estimates, either by reformulating the registration objective function as a probability distribution and drawing samples (Risholm et al., 2013; Le Folgoc et al., 2016; Schultz et al., 2018), approximating the sampling process using test-time NN drop-out (Yang et al., 2017) or sampling using the test-time deformation covariance matrices embedded within a”
From The Impact of Machine Learning on 2D/3D Registration for Image-Guided Interventions: A Systematic Review and Perspective (2021) · cites this paper as Yang et al., 2017
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“Yang et al. [323] implemented stacked convolutional layers as an encoder-decoder approach to predict the morphing of the input pixel into its last formation using MRI brain scans from the OASIS dataset. They employed a registration model known as Large Deformation Diffeomorphic Metric Mapping (LDDMM) and attained remarkable enhancements in computation time.”
From Review of deep learning: concepts, CNN architectures, challenges, applications, future directions (2021) · cites this paper as [323]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Image-to-image test total does not match the dataset counts
MinorThe four listed dataset counts sum to 2,088 registrations, not 2,168.
Reported 2168 · recomputed 2088
“resulting in a total of 2168 registration (1560 from LPBA40, 306 from IBSR18, 90 from MGH10, 132 from CUMC12) test cases.”
Multimodal test-pair count is inconsistent with 16 images
MinorPairwise directed registrations among 16 images yield 16 × 15 = 240 distinct-image cases, not 250.
Reported 250 · recomputed 240
“For testing, we perform pair-wise T1w-T2w registrations for all 16 test images, resulting in 250 test cases.”
Multimodal training-set size changes from 359 to 365
MinorThe dataset split explicitly assigns 359 images to training, but the later description calls the training set 365 images.
“using only 10 of the 365 training images which are randomly chosen for training.”
“We select 359 of the images for training and use the remaining 16 images for testing.”
Reported prediction runtime exceeds its stated components
MinorThe stated prediction and shooting times add to 16.53 seconds, rather than the reported total of 18.43 seconds.
Reported 18.43 · recomputed 16.53
“the initial momentum prediction time is, on average, 7.63 seconds. Subsequent geodesic shooting to generate the deformation field takes 8.9 seconds, resulting in a total runtime of 18.43 seconds.”
Citations
Cites
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946 · cited 2×
“The network is implemented in PyTorch, and optimized using Adam [36], where we set the learning rate to 0.”
MethodFully convolutional networks for semantic segmentation
Long, J., Shelhamer, E., & Darrell, T. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298965 · Cited by 38,834 · cited 1×
“layer [24] to perform “unpooling” (Dec1(F ) in Fig.”
Background(2014) · Cited by 33,419
Braun, W. J., & Murdoch, D. J. (2021) · Cambridge University Press eBooks · doi:10.1017/9781108993456.011 · Cited by 14,099
(2010) · International Conference on Machine Learning · Cited by 13,995
Fully convolutional networks for semantic segmentation
Shelhamer, E., Long, J., & Darrell, T. (2017) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2572683 · Cited by 11,898
Horn, B. K. P., & Schunck, B. G. (1981) · Artificial Intelligence · doi:10.1016/0004-3702(81)90024-2 · Cited by 10,312
Nocedal, J. M., & Wright, S. J. (2006) · Springer series in operations research/Springer series in operations research and financial engineering · doi:10.1007/978-0-387-40065-5 · Cited by 10,206
A fast diffeomorphic image registration algorithm
Ashburner, J. (2007) · NeuroImage · doi:10.1016/j.neuroimage.2007.07.007 · Cited by 8,122 · cited 3×
“Several points should be noted: first, the LDDMM optimization performance is on par with SyN [68], ART [69] and the SPM5 DARTEL Toolbox ( SPM5D ) [70].”
MethodHorn, B. K. P., & Schunck, B. G. (1981) · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · doi:10.1117/12.965761 · Cited by 7,443 · cited 1×
“Optical flow approaches, such as the classical Horn and Schunck optical flow [41], the more recent total variation approaches [42], or methods based on linear elasticity theory [1] are examples for displacementbased registration formulations.”
Method
Flagged first, then by citations
17 of the 94 references couldn’t be matched to a known paper.
Cited by
Sorted by citations
Items 601–654 of 654
Boosting 2D brain image registration via priors from large model
Lin, H., & Song, Y. (2025) · Medical Physics · doi:10.1002/mp.17696
Ramakoteswararao, S., Kumari, P. R., Veenanand, K., & Lakshmi, G. (2025) · AIP conference proceedings · doi:10.1063/5.0241878
Nonrigid registration method for longitudinal chest CT images in COVID-19
Iwao, Y., Kawata, N., Sekiguchi, Y., & Haneishi, H. (2024) · Heliyon · doi:10.1016/j.heliyon.2024.e37272
VoteNet++: Registration Refinement for Multi-Atlas Segmentation
Ding, Z., & Niethammer, M. (2020) · arXiv · doi:10.48550/arxiv.2010.13484
Self-citationIyer, D., & Bhan, A. (2025) · International Conference on Computing Methodologies and Communication (ICCMC) · doi:10.1109/iccmc65190.2025.11140734
Multi-spectral image shift-estimation error calculations using simulated phenomenology
Demars, C. D., Tyler, D. W., Allen, D., et al. (2018) · Applied Optics · doi:10.1364/ao.57.008989
Discriminative Representation Learning for Rigid Registration of Diffusion-Weighted MR Images
Wang, J., Zhang, C., & Zhao, J. (2022) · doi:10.2139/ssrn.4094742
Cross-spectral registration of natural images with SIPCFE
Farzaneh, A. H., & Qi, X. (2020) · Machine Vision and Applications · doi:10.1007/s00138-020-01057-6
Classification of Deformable Smooth Shapes Through Geodesic Flows of Diffeomorphisms
Dabirian, H., Sultamuratov, R., Herring, J., Tallawi, C. E., Zoghbi, W. A., Mang, A., & Azencott, R. (2024) · Journal of Mathematical Imaging and Vision · doi:10.1007/s10851-024-01211-z
Cross-Modal Image Registration via Rasterized Parameter Prediction for Object Tracking
Zhang, Q., & Xiang, W. (2023) · Applied Sciences · doi:10.3390/app13095359
EPDiff-JF-Net: Adjoint Jacobi Fields for Diffeomorphic Registration Networks
Ramon-Julvez, U., Hernández, M., & Mayordomo, E. (2024) · IEEE International Symposium on Biomedical Imaging (ISBI) · doi:10.1109/isbi56570.2024.10635118
Region-Adaptive Attention and Edge-Guided Alignment for Medical Image RegistrationPreprint
Liu, L., Zeng, Y., & Wang, W. (2025) · Research Square · doi:10.21203/rs.3.rs-7059275/v1
Voxelmorph Learns to Pay Attention
Tousi, S. M. A., Khoramdel, J., Borhani, Y., Shojaei, S., & Nikoofard, A. (2023) · RSI International Conference on Robotics and Mechatronics (ICRoM) · doi:10.1109/icrom60803.2023.10412433
Artificial Intelligence in Radiation Oncology: A Rapidly Evolving Picture
Veeraraghavan, H., & Deasy, J. O. (2022) · Springer eBooks · doi:10.1007/978-3-031-08601-4_11
Discriminative Representation Learning for Rigid Registration of Diffusion-Weighted Mr Images
Wang, J., Zhang, C., & Zhao, J. (2022) · doi:10.2139/ssrn.4071358
Lddmm-Gan: Diffeomorphic Image Registration with Unsupervised Generative Adversarial NetworksPreprint
Ramon-Julvez, U., Hernandez, M., & Mayordomo, E. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.5063931
Bridging the Celestial Divide: Cross-Modal Domain Adaptation
Zhang, Z. (2025) · Scientific computation · doi:10.1007/978-981-95-0023-9_4
Li, T., Xiao, R., Xie, T., Huang, W., Hou, J., Qiao, Y., & Xiao, C. (2026) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2026.3666545
AI in Proton Therapy Treatment Planning: A Narrative Review
Ding, Y., Feng, H., Bues, M., Fatyga, M., Liu, T., Zhang, W., Whitaker, T. J., Lin, H., Lee, N. Y., Simone, C. B., et al. (2026) · International Journal of Particle Therapy · doi:10.1016/j.ijpt.2026.101965
Development of an Image Binarization Software Tool for Net Occlusion EstimationsPreprint
Cappaert, R., Yang, W., Ross, D. J., et al. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4864733
Distributional stochastic Planner‐Actor‐Critic for deformable image registration
Du, J., Chang, Q., & Lian, L. (2024) · International Journal of Imaging Systems and Technology · doi:10.1002/ima.23109
A Survey on Deep Learning-Based Diffeomorphic Mapping
Yang, H., Lyu, J., Tam, R. C., & Tang, X. (2022) · Springer eBooks · doi:10.1007/978-3-030-03009-4_108-1
Cascade Registration and Fusion for Unaligned Infrared and Visible Images in Autonomous Driving
Xiao, L., Xie, Y., & Yao, C. (2026) · Electronics · doi:10.3390/electronics15071427
Acute Lymphoblastic Leukemia (ALL) Classification from Microscopic Images Using CNN Model ‘ALLNet’
Shiblee, M. F. H., Saki, S. M., Begum, N., et al. (2024) · 2024 IEEE International Conference on Power, Electrical, Electronics and Industrial Applications (PEEIACON) · doi:10.1109/peeiacon63629.2024.10800537
Guiding Registration with Emergent Similarity from Pre-trained Diffusion Models
Tursynbek, N., Greer, H., Demir, B., & Niethammer, M. (2025) · Lecture notes in computer science · doi:10.1007/978-3-032-04965-0_23 · Cited by 2
Self-citationDong, M., Xie, L., Das, S. R., Wang, J., Wisse, L. E. M., de Florès, R., Wolk, D. A., & Yushkevich, P. A. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2010.12948
Temporal Periodic Image Registration with Implicit Neural Representations
Lowes, M. M., Sørensen, K. A., Maxime Sermesant, B., & Paulsen, R. R. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-09513-8_43
Khawaled, S., & Freiman, M. (2021) · arXiv (Cornell University)
Constraining Volume Change in Learned Image Registration for Lung CTs.Preprint
Hering, A., Häger, S., Moltz, J. H., Leßmann, N., Heldmann, S., & van Ginneken, B. (2020) · arXiv (Cornell University)
Liu, J., Tang, C., Ye, Y., Jiang, J., & Liu, J. (2026) · Analytica Chimica Acta · doi:10.1016/j.aca.2026.345640 · Cited by 1
Registration: 2D/3D rigid registration
Liao, H., Zhou, S., & Luo, J. (2023) · Elsevier eBooks · doi:10.1016/b978-0-12-824383-1.00014-9
Ren, X., Song, H. S., Zhang, Z., & Yang, T. (2024) · Physics in Medicine and Biology · doi:10.1088/1361-6560/ad6741
A review on pulmonary carcinoma using machine learning
Elavarasu, M., & Govindaraju, K. (2023) · AIP conference proceedings · doi:10.1063/5.0178108
A Multitemporal Images Registration Approach Based on Point-Line Constraints
Wang, H., & Hu, Q. (2018) · 2018 26th International Conference on Geoinformatics · doi:10.1109/geoinformatics.2018.8557137
Li, J., Li, Y., Yu, C., Lin, H.-H., Liu, T., Zhang, F., Wang, D., Raza, W., Zhang, L., Kear, P. J., et al. (2026) · Scientific Reports · doi:10.1038/s41598-026-37163-0
A deep neural network-based distortion registration technique for astronomical images
Zhao, G., Li, J., Luo, L., & Cheng, H. (2023) · Tenth International Conference on Information Optics and Photonics · doi:10.1117/12.3000016
Structure-guided registration in learning based image analysis
Lee, M. C. H. (2019) · Spiral (Imperial College London) · doi:10.25560/79296
UDA2L: Unsupervised deformation-aware adaptive learning for medical image registration
Li, D., Zhu, X., Han, Y., et al. (2027) · Pattern Recognition · doi:10.1016/j.patcog.2026.114737
Method and application on the localization of feature-point pairs in multi-modal medical images
Gao, J., Shi, W., Miao, Y., Zhao, J., Zhang, K., Qin, J., Li, Y., He, W., He, F., Liu, J., et al. (2020) · AIP conference proceedings · doi:10.1063/5.0000142
Multi-scale Neural ODEs for 3D Medical Image RegistrationPreprint
Xu, J., Chen, E. Z., Chen, X., et al. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.08493
Circularly Deformable Medical Image Registration Based on Transformer-CNN with PromptPreprint
Li, L., Li, L., zhang, Y., et al. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4700960
MedSAM-guided geometry-aware 2D-3D feature fusion for medical image registration
Gao, F., He, Y., Jiang, H., Yu, P., Du, S., Liu, C., Li, S., & Hao, A. (2026) · Neural Networks · doi:10.1016/j.neunet.2026.109066
A Lightweight Residual Network for Unsupervised Deformable Image Registration
Siyal, A. R., Grams, A. E., & Haltmeier, M. (2024) · IEEE Access · doi:10.1109/access.2024.3513441
Gao, X., & Zheng, G. (2023) · doi:10.36227/techrxiv.23994501 · Cited by 13
Andresen, J., Kepp, T., Ehrhardt, J., et al. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-11203-4_1
Hernández, M. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4866560
Wen, Z., Cheng, J., Zhu, Y., et al. (2025) · 2025 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm66473.2025.11356115
Metadata
- Authors
- Yang, X., Kwitt, R., Styner, M., & Niethammer, M.
- Affiliations
- University of North Carolina at Chapel Hill (US)
- University of Salzburg (AT)
- Corresponding author
- Yang, Xiao
- Venue
- NeuroImage
- Publisher
- Elsevier BV
- Published
- 2017-09
- Volume
- 158
- Pages
- 378-396
- ISSN
- 1053-8119 (ISSN-L)
- PMC
- PMC6036629
- OpenAlex
- W2604920239
- Cited by
- 630
- References
- 94
- Citations by year
- 2026: 17 · 2025: 53 · 2024: 74 · 2023: 65 · 2022: 101 · 2021: 85 · 2020: 95 · 2019: 93 · 2018: 40 · 2017: 7
- FWCI
- 26.68
- Type
- journal article
- Language
- en
- Open access
- bronze
- doi.org
- Licence
- www.elsevier.com/tdm/userlicense/1.0/ (tdm)
- www.elsevier.com/legal/tdmrep-license (tdm)
- www.elsevier.com/open-access/userlicense/1.0/ (am, 320 d)
- doi.org/10.15223/policy-017 (stm-asf)
- doi.org/10.15223/policy-037 (stm-asf)
- doi.org/10.15223/policy-012 (stm-asf)
- doi.org/10.15223/policy-029 (stm-asf)
- doi.org/10.15223/policy-004 (stm-asf)
- Locations
- NeuroImage (acceptedVersion)
- PubMed (publishedVersion)
- Europe PMC (PubMed Central) (acceptedVersion)
- PubMed Central (acceptedVersion)
- Funders
- NIH (1 R41 NS091792-01, HDO55741)
- NSF (ECCS-1148870, EECS-1711776)
- National Science Foundation Major Research Instrumentation program (ACI-1429830)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Medical Image Segmentation Techniques
- Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
- Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Retinal Imaging and Analysis
- Keywords
- deformable image registration, LDDMM, encoder-decoder architecture, multimodal image registration, image similarity measures, diffeomorphic registration
- MeSH terms
- Machine Learning, Algorithms, Brain, Brain Mapping, Humans, Image Processing, Computer-Assisted, Pattern Recognition, Automated, Imaging, Three-Dimensional
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-10-05, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-05 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
