Are Registration Uncertainty and Error Monotonically Associated

Alexandra Golby; Duo Wang; Jie Luo; Masashi Sugiyama; Sarah Frisken; William M. Wells III

REVIEW 4 major objections 4 minor 38 references

Are Registration Uncertainty and Error Monotonically Associated

T0 review · 4 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read Gaussian process registration uncertainty and non-rigid registration error are weakly to moderately monotonically correlated, and the association varies by how landmarks are chosen.

desk verdict A genuinely new empirical question about GP registration uncertainty and error, but the missing kernel length-scale parameter makes the headline correlations non-reproducible and the qualitative claim conditional. read the letter →

arxiv 1908.07709 v2 pith:XHZ6RNUW submitted 2019-08-21 cs.CV

Jie Luo , Sarah Frisken , Duo Wang , Alexandra Golby , Masashi Sugiyama , William M. Wells III This is my paper · ORCID

classification cs.CV

keywords registrationuncertaintyerrorGaussianprocessSpearman'srankcorrelationimage-guidedneurosurgeryprobabilisticimagebrainshiftultrasound

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper asks a question clinicians implicitly rely on: when a probabilistic registration reports high uncertainty at a point, is the registration error at that point also large? Using Gaussian process registration on intraoperative ultrasound from 23 neurosurgery cases, the authors measure the monotonic association between point-wise transformation uncertainty and landmark-based registration error. They find Spearman rank correlations of about 0.29 for manual landmarks and 0.40 for automatically detected landmarks, a weak-to-moderate positive relationship, while patch-wise correlations using appearance-based error metrics stay consistently low. The paper concludes that at this stage using GP uncertainty as a surrogate for non-rigid registration error is not yet feasible in practice, and the question deserves further study.

What carries the argument

The object carrying the argument is the posterior covariance of a Gaussian process displacement field. Registration uncertainty at a voxel is the diagonal entry of $\Sigma_* = K_{**} - K_*^T K^{-1} K_*$, computed with kernel $k(x,x') = \exp(-x^2/a)$; larger posterior variance means less confidence in the interpolated displacement. The authors compare this uncertainty to error using Spearman's rank correlation $\rho_s$, chosen because it tests monotonic rather than linear relationships and is robust to outliers. The comparison is done point-wise, using known landmark displacements as ground truth, and patch-wise, using appearance-based dissimilarity metrics (histogram intersection and SSD) as error surrogates.

What would settle it

Recompute the point-wise Spearman's rho on the same landmarks while varying the kernel length-scale $a$ over a wide range; if the correlation swings from 0.29-0.40 to near zero or negative, the weak-to-moderate association is an artifact of the unstated kernel calibration rather than a robust property of GP registration uncertainty.

Watch

Extended reading notes

Core claim

The central claim is that there is a positive monotonic association between point-wise GP registration uncertainty and non-rigid registration error, but it is weak-to-moderate, not strong enough to support clinical use as a surrogate. In the point-wise posterior predictive checking experiment, the mean Spearman's rho is 0.2899 for manual landmarks from the RESECT dataset and 0.4014 for automatically extracted landmarks from the MIBS dataset. In the patch-wise experiment, where error is approximated by histogram intersection and sum of squared differences over image patches, the correlations are consistently low for both datasets. The authors infer that transformation uncertainty and appearance-based error may be inherently weakly coupled, and that the distribution of landmarks matters for the strength of the association.

Load-bearing premise

The load-bearing premise is that the Gaussian process posterior variance used as 'registration uncertainty' is a faithful measure of confidence; the kernel length-scale $a$ is not reported, so if a different $a$ re-ranks the uncertainties, the reported correlations could be an artifact of calibration rather than a property of the method.

Editorial extensions

If this is right

Surgeons cannot yet treat GP uncertainty maps as reliable indicators of where non-rigid registration may be wrong, particularly in patch-level assessments.
Point-wise uncertainty carries some ranking information about error, but the strength is too low to set confidence thresholds for clinical decisions.
The higher correlation for automatic landmarks suggests that the spatial distribution of features used for interpolation modulates the uncertainty-error link.
Registration uncertainty should not be used as a proxy for error without further calibration or validation against dense ground-truth deformation.

Reading between the lines

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

The paper studies only Gaussian process registration; other probabilistic registration approaches may show a different uncertainty-error association, so the practical verdict should not be read as a general statement about registration uncertainty.
Because the kernel length-scale is not reported, varying it in a follow-up experiment would show how much of the measured correlation is intrinsic to GP uncertainty and how much is an artifact of calibration.
A synthetic phantom with a known dense deformation field could test the same question without relying on sparse landmarks or appearance-based patch error, giving a cleaner ground-truth comparison.
The weak patch-wise correlation may reflect the difference between transformation uncertainty and label uncertainty; a model that estimates label uncertainty directly might yield clinically useful error maps where this one does not.

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.

Desk Editor's Note

read the letter

The short version: this is a pilot study with a genuinely new question, and a guarded answer, but the missing GP kernel length-scale in Eq. 3 makes the headline numbers impossible to reproduce. That is the one thing I would want fixed before trusting the quantitative claim.

What is actually new: nobody has systematically asked whether registration uncertainty from GP registration tracks error pointwise in clinical data. The authors use Spearman's rho, which is appropriate since clinical usefulness only needs monotonicity, and they get mean rho around 0.29 for manual landmarks and 0.40 for automatic ones on two datasets, one public (RESECT) and one proprietary (MIBS). They are refreshingly cautious in the conclusion, explicitly saying it is too early to conclude it is safe to use uncertainty as a surrogate.

The paper is honest about its own soft spots. In Section 3.2 they state the intensity-based patch-wise error metric is sub-optimal, and they flag the transformation-vs-label uncertainty distinction. That is good scholarly behavior.

Where it gets soft: the length-scale a in k(x,x') = exp(-x^2/a) is never specified anywhere, including the experiments. This is not a nitpick. The posterior variance in Eq. 3 depends directly on a. As a grows large, the covariance becomes nearly constant, so uncertainty is flat and rank correlation degenerates; as a shrinks, uncertainty is concentrated around the interpolating landmarks. The rank ordering, and therefore every rho value in Fig. 5, depends on a. Without a value or a sensitivity analysis, the reported numbers cannot be reproduced, and one cannot rule out that the chosen a was cherry-picked. The stress-test note about this is on point.

Also missing: confidence intervals or significance tests. The point-wise correlations are computed over many landmarks, but we never learn whether rho = 0.29 is statistically distinguishable from zero. The patch-wise results are exploratory, and the authors admit the metric is sub-optimal, so those numbers carry little weight.

Proportionately: these are real but addressable flaws. The central qualitative conclusion -- that GP uncertainty has only weak-to-moderate monotonic association with error -- may survive once a is fixed, but the paper as written leaves it conditional on an unstated hyperparameter.

For whom: anyone working on uncertainty-driven error assessment in image-guided interventions. The paper is a reasonable pilot, not a definitive negative result.

Recommendation: send it to peer review, but require the authors to report a and run a sensitivity analysis, plus some measure of uncertainty on the Spearman values. That would turn a conditional pilot into a useful benchmark.

Referee Report

4 major / 4 minor

Summary. This paper asks whether Gaussian-process (GP) registration uncertainty is monotonically associated with non-rigid registration error in image-guided neurosurgery. Using two clinical ultrasound datasets (RESECT and MIBS), the authors compute Spearman rank correlations between point-wise GP posterior uncertainties and landmark displacement errors, as well as patch-wise correlations using intensity dissimilarity measures (HI and SSD). They report weak-to-moderate positive point-wise correlations (mean rho = 0.2899 for manual landmarks, 0.4014 for automatic landmarks) and consistently low patch-wise correlations, concluding that GP uncertainty is not yet a reliable surrogate for registration error.

Significance. If the empirical result holds, it provides a useful cautionary quantification for clinicians considering GP uncertainty maps as error surrogates, and it opens a question that is indeed overlooked in the registration literature. The study has clear strengths: the point-wise test is direct, uses ground-truth displacements, and involves no circularity because the GP posterior variance is computed from landmark positions and kernel choices independent of the error measurement. The use of two clinical datasets is also a plus. However, the quantitative conclusions are conditional on an unstated kernel hyperparameter and lack any statistical uncertainty assessment, so the current weight of the evidence is limited.

major comments (4)

[Section 2.1, Eq. (3)] The covariance kernel is written as k(x, x') = exp(-x^2/a), but no value or range for a is given anywhere in the manuscript. The posterior variance in Eq. (3) depends on a and the landmark positions, but not on the observed displacements D; consequently the rank ordering of uncertainty values and every Spearman coefficient in Fig. 5 is a function of this unstated parameter. The authors must report the exact value of a used and provide a sensitivity analysis over a range of a, since the headline means (0.2899 and 0.4014) cannot otherwise be reproduced and the influence of this hyperparameter on the conclusions cannot be assessed.
[Section 3.1, Fig. 5] The central claim of weak-to-moderate positive monotonic correlation rests only on mean rho values across image pairs. The authors do not report the number of image pairs and landmarks, the per-pair correlation values, confidence intervals, or significance tests, so the reader cannot judge whether the correlations are statistically distinguishable from zero or how stable they are across patients. Please provide this information, for instance with bootstrap confidence intervals for the mean rho and per-case results.
[Section 3.2] The patch-wise experiment is interpreted as showing consistently low correlation, but the authors themselves state that the intensity-based HI/SSD dissimilarity is 'sub-optimal' for estimating registration error under large deformation. Because the error surrogate is acknowledged to be flawed, the low patch-wise rho values cannot be cleanly attributed to the uncertainty-error relationship. The patch-wise conclusion should either be supported with a validated local error measure or explicitly downweighted and removed from the paper's conclusions.
[Section 3] The experimental setup is not fully specified: the manuscript does not state how many interpolating landmarks were used per image, how they were selected, how many test points were evaluated, or how the automatic landmark detection was performed in MIBS. These details materially affect both the GP uncertainty and the error distribution and are needed for reproducibility of the reported correlations.

minor comments (4)

[Section 2.1] The kernel notation k(x, x') = exp(-x^2/a) is ambiguous because k takes two spatial arguments; the exponent should be written as e.g. exp(-||x - x'||^2 / a).
[Section 2.2] The rank-transformation example is correct, but the authors should state how tied ranks are handled when computing Spearman's rho.
[Section 3.2] The sentence 'while intensities over patches is label uncertainty' is unclear; patch intensities are observations used to define the dissimilarity, not label uncertainty, and the connection to the transformation-versus-label uncertainty distinction should be explained more carefully.
[Section 3.1] The text says the automatic landmarks 'achieved an average rho of 0.4014, which can be categorized as a moderate-to-strong correlation,' but the correlation categories are not defined; the reader would benefit from an explicit threshold or reference for the labels 'weak-to-moderate' and 'moderate-to-strong.'

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: GP uncertainty comes from the posterior covariance while error is measured against ground-truth displacements, from independent sources; no fitted parameter or self-citation chain forces the result.

full rationale

The paper's central claim is an empirical association, not a derivation. The uncertainty u is taken from the diagonal of the GP posterior covariance matrix (Eq. 3), which depends on the kernel and landmark geometry but not on the observed displacements. The error is computed independently as the Euclidean distance between the ground-truth displacement and the GP posterior mean (Eq. 2, Sec. 2.3). No parameter was fitted to maximize the Spearman correlation, and the test points are explicitly stated not to have been used for GP interpolation (Sec. 3). The self-citations to earlier GP registration work and to the transformation-versus-label uncertainty taxonomy are contextual and are not load-bearing for the empirical correlation result; there is no imported uniqueness theorem or ansatz presented as a derived constraint. One reproducibility concern exists: the GP kernel length-scale a in k(x,x') = exp(-x^2/a) is never specified, so the numerical correlation values are conditional on an unstated hyperparameter and lack a sensitivity analysis. That is a calibration/reproducibility gap, not circularity, because the uncertainty measure is still computed from a distinct source than the error, and the paper does not claim to derive the association from that kernel choice.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The central empirical result depends on the GP model, the choice of kernel, the landmark sampling, and the error proxies. No new entities or fitted constants are introduced beyond the kernel length-scale, which is left unspecified.

free parameters (1)

GP kernel length-scale a = not reported
Appears in k(x,x') = exp(-x^2/a) in Section 2.1; value not stated in the paper. Uncertainties from Eq. (3) depend on it, and all correlations are computed using those uncertainties.

assumptions (4)

domain assumption The displacement field d(x) is well modeled as a zero-mean Gaussian process with kernel exp(-x^2/a).
Section 2.1, Eq. (1). This is the probabilistic model whose uncertainty is being tested; if the model is misspecified, the uncertainty may not correlate with error.
standard math Spearman's rank correlation measures monotonic association.
Section 2.2, standard definition.
domain assumption Intensity-based dissimilarity (SSD, HI) is a valid proxy for patch-wise registration error.
Section 2.4; authors acknowledge it is suboptimal under large deformation (Section 3.2), so the patch-wise conclusions are conditional on this assumption.
domain assumption Tested landmarks were not used for GP interpolation, so the error and uncertainty are measured at independent points.
Section 3, stated explicitly; necessary for the point-wise correlation to be meaningful.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Are Registration Uncertainty and Error Monotonically Associated." pith.science (2026). https://pith.science/paper/XHZ6RNUW

@misc{pith2026190807709,
  author       = {Pith},
  title        = {Pith review of: Are Registration Uncertainty and Error Monotonically Associated},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XHZ6RNUW}},
  note         = {Machine review of arXiv:1908.07709}
}

read the original abstract

In image-guided neurosurgery, current commercial systems usually provide only rigid registration, partly because it is harder to predict, validate and understand non-rigid registration error. For instance, when surgeons see a discrepancy in aligned image features, they may not be able to distinguish between registration error and actual tissue deformation caused by tumor resection. In this case, the spatial distribution of registration error could help them make more informed decisions, e.g., ignoring the registration where the estimated error is high. However, error estimates are difficult to acquire. Probabilistic image registration (PIR) methods provide measures of registration uncertainty, which could be a surrogate for assessing the registration error. It is intuitive and believed by many clinicians that high uncertainty indicates a large error. However, the monotonic association between uncertainty and error has not been examined in image registration literature. In this pilot study, we attempt to address this fundamental problem by looking at one PIR method, the Gaussian process (GP) registration. We systematically investigate the relation between GP uncertainty and error based on clinical data and show empirically that there is a weak-to-moderate positive monotonic correlation between point-wise GP registration uncertainty and non-rigid registration error.

Figures

Figures reproduced from arXiv: 1908.07709 by the authors.

**Figure 1.** An example to illustrate the usefulness of registration uncertainty in IGNs. Even though the brain clearly undergoes non-linear deformation during surgery, rigid registration is still the standard for clinical practice [6]. Although non-rigid registration has long been a goal for IGNs, this goal is hampered because non-rigid registration error is less predictable and harder to validate than rigid registration error… view at source ↗

**Figure 2.** (a) 3 displacement vectors; (b) A 10 × 10 interpolated dense deformation field; (b) A visualization of registration uncertainty. Diagonal entries of Σ∗ are the marginal transformation variances, and they can be used as the GP registration uncertainty. In this study, we choose the same kernel k(x, x ′ ) = exp(− x 2 a ) for all three displacement components. Fig.2(b) shows a 10 × 10 dense deformation field interpolate… view at source ↗

**Figure 3.** (a) An illustrative example for the point-wise posterior predictive checking experiment; (b) An illustration for how to compute ǫ and u in the context of IGNs 2.3 Point-wise posterior predictive checking When a surgeon is removing a tumor mass near a critical structure, it is vital that s/he knows how close the predicted instrument location is from the structure and how confident the prediction is. With GP registra… view at source ↗

Figures from the paper (3 more)

**Figure 4.** Figure 4: An illustration for using the HI metric to compute Spearman’s rank correlation coefficient for patches. 2.4 Patch-wise correlation test We also investigate the correlation between u and ǫ over image-patches. Because in IGNs, surgeons may be more interested in registrat…

**Figure 5.** Figure 5: The estimated ρs for the point-wise experiment. We can see a moderate positive monotonic relationship between u and ǫ. [37], while MIBS is a proprietary dataset from a local hospital. Both datasets in total contain 23 sets of p-US and i-US scans that were acquired from…

**Figure 6.** Figure 6: The estimated ρs(Ω) for the patch-wise experiment. Values of ρs(Ω)’s are consistently low for both datasets. values of k include 3 and 5. We calculated ǫ using SSD/HI for all patches and computed ρs(Ω) afterward. The estimated patch-wise ρs’s are shown in Fig.6. It ca…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 38 canonical work pages

[1]

Maintz, J.B.A., Viergever, M.A.: A Survey of Medical Imag e Registration. Med. Image Anal. 2(1): 1-36 (1998)

work page 1998
[2]

IEEE Trans

Sotiras, A., Davatzikos, C.: Deformable Medical Image Re gistration: A Survey. IEEE Trans. Med. Imaging 32(7), 1153-1190 (2013)

work page 2013
[3]

Gerard, I.J., et al: Brain shift in neuronavigation of bra in tumors: A review. Med. Image Anal. 35: 403-420 (2017)

work page 2017
[4]

Morin, F., et al.: Brain-shift compensation using intrao perative ultrasound and constraint-based biomechanical simulation. Med. Image An al. 40: 133-153 (2017)

work page 2017
[5]

In: Shen D

Luo M., Larson P.S., Martin A.J., Konrad P.E., Miga M.I.: A n Integrated Multi- physics Finite Element Modeling Framework for Deep Brain St imulation: Prelimi- nary Study on Impact of Brain Shift on Neuronal Pathways. In: Shen D. et al. (eds). MICCAI 2019. LNCS, vol 11768. pp. 682-690. Springer, Cham (2 019)

work page 2019
[6]

L.: Deformable registration of pre operative MR, pre-resection ultrasound, and post-resection ultrasound images of neuro surgery

Rivaz, H., Collins, D. L.: Deformable registration of pre operative MR, pre-resection ultrasound, and post-resection ultrasound images of neuro surgery. IJCARS (2015)

work page 2015
[7]

IEEE Trans

Kybic L.: Bootstrap Resampling for Image Registration Un certainty Estimation Without Ground Truth. IEEE Trans. Image Processing 19(1), 6 4-73 (2010)

work page 2010
[8]

et al.: Assessment of Rigid Registration Quality Measures in Ultrasound- Guided Radiotherapy

Shams R. et al.: Assessment of Rigid Registration Quality Measures in Ultrasound- Guided Radiotherapy. IEEE Trans. Med. Imaging 37(2), 428-4 37 (2018)

work page 2018

Show all 38 references

[9]

In: LNCS, vol

Datteri R.D., Dawant B.M.: Automatic detection of the mag nitude and spatial location of error in non-rigid registration. In: LNCS, vol. 7359, pp. 2130. (2012)

2012
[10]

IEEE Trans

Hub, M., et al.: A stochastic approach to estimate the unc ertainty involved in B- spline image registration. IEEE Trans. Med. Imaging 28(11) , pp. 17081716 (2009)

2009
[11]

Hub M., Karger C.P.: Estimation of the uncertainty of ela stic image registration with the demons algorithm. Phys. Med. & Bio, 58(9): (2013)

2013
[12]

IEEE Trans

Saygili G., et al.: Conﬁdence Estimation for Medical Ima ge Registration Based On Stereo Conﬁdences. IEEE Trans. Med. Imaging 35(2), 539-549 (2016)

2016
[13]

In: MICCAI’16

Sokooti H., et al.: Accuracy Estimation for Medical Imag e Registration Using Re- gression Forests. In: MICCAI’16. LNCS, vol 9902, pp. 107-11 5 (2016)

2016
[14]

Sokooti, H., Saygili, G., Glocker, B., Lelieveldt, B.P. F. and Staring, M.: Quantita- tive error prediction of medical image registration using r egression forests. Medical Image Analysis, 56:110121, 2019

2019
[15]

Saygili G.: Local-search based prediction of medical im age registration error. Proc. SPIE 10577, Medical Imaging’18. doi: 10.1117/12.2293740 ( 2018)

2018 doi
[16]

a nd Navab, N.: Optical ﬂow estimation with uncertainties through dynamic mrfs

Glocker, B., Paragios, N., Komodakis, N., Tziritas, G. a nd Navab, N.: Optical ﬂow estimation with uncertainties through dynamic mrfs. In Computer Vision and Pattern Recgonition (CVPR), pages 18, 2008

2008
[17]

In: MIC- CAI’13

Popuri, K., et al.: A Variational Formulation for Discre te Registration. In: MIC- CAI’13. LNCS, vol. 8151, pp. 187-194. Springer, Berlin, Hei delberg (2013)

2013
[18]

Risholm, et al.: Bayesian Characterization of Uncertai nty in Intra-subject Non- rigid Registration. Med. Image Anal. 17(5), 538-555 (2013)

2013
[19]

In: LNCS, vol

Lotﬁ, T., et al.: Improving Probabilistic Image Registr ation via Reinforcement Learning and Uncertainty Evaluation. In: LNCS, vol. 8184, p p. 187-194. (2013)

2013
[20]

In: WBIR’14

Wassermann D., et al.: Probabilistic Diﬀeomorphic Regi stration: Representing Un- certainty. In: WBIR’14. LNCS, vol. 8545. pp. 72-82. Springe r, Cham (2014)

2014
[21]

In: MICCAI’15

Yang, X., Niethammer, M.: Uncertainty Quantiﬁcation fo r LDDMM Using a Low- rank Hessian Approximation. In: MICCAI’15. LNCS, vol. 9350 , pp. 289-296. (2015)

2015
[22]

Simpson, I.J.A., et al.: Probabilistic non-linear regi stration with spatially adaptive regularisation. Med. Image Anal. 26(1): 203-216 (2015) 10 j. Luo et al

2015
[23]

Heinrich, M.P., et al.: Deformable Image Registration b y Combining Uncertainty Estimates From Supervoxel Belief Propagation. Med. Image A nal. 27, 57-71 (2016)

2016
[24]

IEEE Trans

Folgoc, et al.: Quantifying Registration Uncertainty W ith Sparse Bayesian Mod- elling. IEEE Trans. Medical Imaging, 36(2): 607-617 (2017)

2017
[25]

In: MICCAI’18

Dalca, A.V., et al.: Unsupervised Learning for Fast Prob abilistic Diﬀeomorphic Registration. In: MICCAI’18. LNCS, vol. 11070, pp. 729-738. Springer, Cham (2018)

2018
[26]

In: MICCAI’18

Wang, J., et al.: Eﬃcient Laplace Approximation for Baye sian Registration Uncer- tainty Quantiﬁcation. In: MICCAI’18. LNCS, vol. 11073, pp. 880-888. (2018)

2018
[27]

In: MICCAI’18

Luo, J., et al.: A feature-driven active framework for ul trasound-based brain shift compensation. In: MICCAI’18. LNCS, vol. 11073, pp. 3038. Sp ringer, Cham (2018)

2018
[28]

and Wells, W.M

Sedghi,T.Kapur, T., Luo, J., Mousavi, P. and Wells, W.M. : Probabilistic image registration via deep multi-class clas- siﬁcation: Charac terizing uncertainty. MIC- CAI’19 UNSURE workshop, LNCS 11840:1222. (2019)

2019
[29]

et al.: Investigation of Feature-Based Nonrigi d Image Registration Using Gaussian Process

Bayer S. et al.: Investigation of Feature-Based Nonrigi d Image Registration Using Gaussian Process. In: Tolxdorﬀ T., Deserno T., Handels H., M aier A., Maier-Hein K., Palm C. (eds) Bildverarbeitung fr die Medizin 2020. Info rmatik aktuell. Springer Vieweg, Wiesbaden (2020)

2020
[30]

and Leemput, K.V.: Fast Nonparametric Mutual-In formation-based Reg- istration and Uncertainty Estimation

Agn, M. and Leemput, K.V.: Fast Nonparametric Mutual-In formation-based Reg- istration and Uncertainty Estimation. In MICCAI’19 UNSURE workshop, LNCS 11840: 42-51. (2019)

2019
[31]

and Frisken, S.: O n the applicability of registration uncer- tainty

Luo, J., Sedghi, A., Popuri, K., Cobzas, D., Zhang, M., Pr eiswerk, F., Toews, M., Golby, A.J., Sugiyama, M., Wells, W.M. and Frisken, S.: O n the applicability of registration uncer- tainty. Medical Imaging Computing a nd Computer Assisted Intervention (MICCAI), 11765:410419, 2019

2019
[32]

and Foreman, D.I.: Nonparametric Statisti cs: A Step-by-Step Ap- proach, Wiley

Corder, G.W. and Foreman, D.I.: Nonparametric Statisti cs: A Step-by-Step Ap- proach, Wiley. (2014)

2014
[33]

Gelman, A., et al.: Bayesian data analysis, p653, Chapma n&Hall, USA (2004)

2004
[34]

Phys.4 4(7): 3875-3882 (2017)

Xiao, Y., et al.: RESECT: A clinical database of preopera tive MRI and intraoper- ative ultrasound in lowgrade glioma surgeries.Med. Phys.4 4(7): 3875-3882 (2017)

2017
[35]

Int J CARS, 13 (10): pp.1525-1538

Machado, I., et al.: Non-rigid registration of 3D ultras ound for neurosurgery us- ing automatic feature detection and matching. Int J CARS, 13 (10): pp.1525-1538. (2018)

2018
[36]

et al.: Visualization of deformable ima ge registration quality using local image dissimilarity

Schlachter, M,. et al.: Visualization of deformable ima ge registration quality using local image dissimilarity. IEEE TMI 35(10): 2319-2328 (201 6)
[37]

Y., et al.: Evaluation of MRI to ultrasound registr ation methods for brain shift correction: The CuRIOUS2018 Challenge, in IEEE Trans actions on Medical Imaging

Xiao. Y., et al.: Evaluation of MRI to ultrasound registr ation methods for brain shift correction: The CuRIOUS2018 Challenge, in IEEE Trans actions on Medical Imaging. (2020)

2020
[38]

Toews M., Wells W.M.: (2013) Eﬃcient and robust model-to -image alignment using 3D scale-invariant features, Med Image Anal, Vol 17, 271-28 2

2013

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Maintz, J.B.A., Viergever, M.A.: A Survey of Medical Imag e Registration. Med. Image Anal. 2(1): 1-36 (1998)

work page 1998

[2] [2]

IEEE Trans

Sotiras, A., Davatzikos, C.: Deformable Medical Image Re gistration: A Survey. IEEE Trans. Med. Imaging 32(7), 1153-1190 (2013)

work page 2013

[3] [3]

Gerard, I.J., et al: Brain shift in neuronavigation of bra in tumors: A review. Med. Image Anal. 35: 403-420 (2017)

work page 2017

[4] [4]

Morin, F., et al.: Brain-shift compensation using intrao perative ultrasound and constraint-based biomechanical simulation. Med. Image An al. 40: 133-153 (2017)

work page 2017

[5] [5]

In: Shen D

Luo M., Larson P.S., Martin A.J., Konrad P.E., Miga M.I.: A n Integrated Multi- physics Finite Element Modeling Framework for Deep Brain St imulation: Prelimi- nary Study on Impact of Brain Shift on Neuronal Pathways. In: Shen D. et al. (eds). MICCAI 2019. LNCS, vol 11768. pp. 682-690. Springer, Cham (2 019)

work page 2019

[6] [6]

L.: Deformable registration of pre operative MR, pre-resection ultrasound, and post-resection ultrasound images of neuro surgery

Rivaz, H., Collins, D. L.: Deformable registration of pre operative MR, pre-resection ultrasound, and post-resection ultrasound images of neuro surgery. IJCARS (2015)

work page 2015

[7] [7]

IEEE Trans

Kybic L.: Bootstrap Resampling for Image Registration Un certainty Estimation Without Ground Truth. IEEE Trans. Image Processing 19(1), 6 4-73 (2010)

work page 2010

[8] [8]

et al.: Assessment of Rigid Registration Quality Measures in Ultrasound- Guided Radiotherapy

Shams R. et al.: Assessment of Rigid Registration Quality Measures in Ultrasound- Guided Radiotherapy. IEEE Trans. Med. Imaging 37(2), 428-4 37 (2018)

work page 2018