Automatic Skull Reconstruction by Deep Learnable Symmetry Enforcement

Daria Hemmerling; Marek Wodzinski; Mateusz Daniol

REVIEW 3 major objections 6 minor 39 references

Automatic Skull Reconstruction by Deep Learnable Symmetry Enforcement

T0 review · 3 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read This paper shows that a neural network trained to predict a skull's symmetry plane can be used both as a training loss and as a registration objective to substantially improve automatic cranial defect reconstruction, reaching near…

desk verdict A practical, low-compute symmetry-enforced skull reconstruction method that mostly delivers on its claims; the main open question is a missing ablation on the symmetry network's behavior with partially reconstructed volumes. read the letter →

arxiv 2411.17342 v1 pith:ISMAJH7A submitted 2024-11-26 eess.IV cs.CVcs.LG

Marek Wodzinski , Mateusz Daniol , Daria Hemmerling This is my paper · ORCID

classification eess.IVcs.CVcs.LG

keywords skullreconstructioncranialimplantdesignlearnablesymmetryplanelossdeformableregistrationdeeplearningmedicalimageanalysisvisiontransformer

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper proposes to improve automatic reconstruction of cranial defects by enforcing skull symmetry in a learnable way. The authors train a symmetry network to predict the symmetry plane of a skull, then use that plane in two ways: as an extra loss during training of the reconstruction network, and as an objective for a deformable registration refinement after reconstruction. On the SkullBreak and SkullFix benchmarks, the symmetry-preserving reconstruction reaches DSC of 0.94 and 0.95, surface Dice of 0.94 and 0.95, and HD95 of 1.31 and 1.28 mm, substantially better than the baseline without symmetry and close to the best published method, while training with under 500 GPU hours instead of over 100,000. The authors argue this makes accurate, scalable, near-real-time skull reconstruction feasible and potentially extendable to other symmetric structures.

What carries the argument

The symmetry network (SN) and the symmetry loss (SL). The SN is an encoder-regressor (volumetric ResNet or vision transformer) that maps a skull volume V to a plane equation π; the SL equals DSL(V, R(V, π)), the Dice loss between the volume and its reflection about the predicted plane, which makes the plane trainable without manual labels. The same SL is reused after reconstruction as a differentiable objective for a deformable registration with diffusive regularization that deforms the reconstructed defect to maximize the whole skull's symmetry. This dual use, as a training loss and as a refinement objective, is the mechanism that transfers healthy-skull symmetry knowledge to defect reconstruction.

What would settle it

For the SkullBreak test set, compute the symmetry network's predicted plane on the intermediate reconstructed volume VRec (defective skull plus initial reconstruction) and measure its deviation from the plane predicted on the healthy ground-truth skull. If the deviation is large on cases where registration refinement fails, or if using the ground-truth plane in the refinement does not produce better Dice than using the predicted plane, the claim that the symmetry network generalizes to reconstructed volumes is not supported.

Watch

Extended reading notes

Core claim

The central claim is that skull symmetry can be learned by a vision-transformer-based encoder-regressor trained with a fully differentiable symmetry loss, defined as SL(V, π) = DSL(V, R(V, π)) where V is the skull volume, π is the predicted plane equation, and R is the reflection operator. The same loss is then used in two integration modes: as an additional objective during reconstruction-network training (Seg-SN), and as an instance-optimization objective for a deformable registration refinement that warps only the reconstructed defect (Reg-SN). In the best configuration, Seg-SN plus Reg-SN, the method reaches DSC/SDSC/HD95 of 0.94/0.94/1.31 on SkullBreak and 0.95/0.95/1.28 on SkullFix, outperforming all previously published methods except the diffusion-augmented one, and it successfully reconstructs all 11 real clinical cases while the baseline manages only 6. A notable design finding is that a vision transformer is better for symmetry estimation while a convolutional RUNet is better for the reconstruction itself.

Load-bearing premise

The symmetry network, trained only on healthy skulls, is assumed to predict a symmetry plane accurate enough on reconstructed volumes that contain the implant and defect, and the registration refinement must deform the implant toward the correct shape rather than reinforcing a wrong plane.

Editorial extensions

If this is right

Skull reconstruction can run in under 15 seconds per case (under 5 seconds without refinement), fast enough for potential intraoperative use, while staying close to state-of-the-art accuracy.
The method approaches the ceiling of what symmetry alone can provide, since the best achievable reflected-implant DSC (0.956 on SkullBreak, 0.969 on SkullFix) is close to the final reconstruction DSC, indicating symmetry is the dominant recoverable signal on these benchmarks.
Training both networks costs under 500 A100 GPU hours, roughly 200 times less than the diffusion-augmented alternative, making the approach feasible for smaller groups and retraining on other symmetric anatomical structures.
The registration-based refinement acts as per-case instance optimization, improving out-of-distribution real defects, which is where segmentation-only networks typically fail.

Reading between the lines

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

The symmetry plane's accuracy on the intermediate reconstructed volumes (defective skull plus initial implant) is the linchpin of the method; directly measuring plane error on those volumes would show whether failures stem from the symmetry network or from the registration optimizer.
A natural, untested extension is a multi-resolution or adaptive regularization schedule for the refinement step, since the authors note the strong diffusive regularization leaves small holes on the largest clinical defects.
Because real skulls are only approximately symmetric, the method's ceiling is limited by intrinsic asymmetry; adding a per-case asymmetry estimate could allow the refinement to weight symmetry only where it is trustworthy.
The same symmetry-loss refinement could transfer to other bilateral structures such as the pelvis or mandible, though the paper only evaluates cranial data.

Share X Bluesky LinkedIn Reddit HN

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

Bottom line: the core result is real. The authors integrate a learned symmetry-plane regressor into cranial implant reconstruction in two ways—as an extra Dice-based loss during training and as an objective for deformable registration refinement at test time. Prior mirroring methods needed a manually chosen axis; this removes that constraint. On SkullBreak/SkullFix the method reaches DSC around 0.94/0.95, within a hair of the diffusion-augmented state of the art, with two orders of magnitude less training compute. That claim is the main contribution and it holds up as far as the paper reports.

Credit where due: the comparison is against public benchmarks, the ablation across five reconstruction architectures is informative, and the compute numbers are specific enough to check. The finding that a ViT is better for symmetry regression while a U-Net remains better for reconstruction is a useful data point. The failure mode for completely wrong initial reconstructions is acknowledged in the text.

Soft spots, in order of importance. First, the registration refinement recomputes the symmetry plane from the partially reconstructed volume at every step, but Table 1 only validates the symmetry network on healthy skulls and implants, not on those VRec volumes. That is a genuine gap. It is not evidence the method is wrong—the test-set metrics improved after refinement, so a biased plane would likely have shown up there—but the paper should add an ablation measuring plane error on VRec and showing refinement is robust to initial reconstruction quality. Second, the paper promises that every significance claim is backed by a Wilcoxon test with p<0.01, but no p-values appear anywhere. That is a broken reporting promise, not a statistical argument. Third, no code or detailed configuration is released; for a method whose value is partly practical reproducibility, that matters. Fourth, the 11 real clinical cases are called successfully reconstructed with only visual evidence; that is qualitative and should be labeled as such.

The stress-test note about circularity overstates the risk. The symmetry network is trained unsupervised on healthy skulls, and the test-set results are measured against ground truth. If refinement were chasing a biased plane, the reported DSC/HD95 would likely worsen, not improve. The concern is legitimate as a missing evaluation, not as a demonstrated flaw.

For whom: people working on cranioplasty, implant design, or symmetry-based shape completion. It deserves a serious referee; ask for the p-values, the VRec plane-error ablation, and code/configuration, then it can be accepted.

Referee Report

3 major / 6 minor

Summary. The manuscript proposes a two-network pipeline for cranial defect reconstruction. A symmetry network (SN) is trained unsupervised on healthy skulls to regress a reflection plane pi using a Dice-based symmetry loss; the plane is then used (i) as an auxiliary loss during training of a reconstruction network (RN), Eq. (2), and (ii) in an optional deformable registration refinement after inference, Eq. (3). The method is evaluated on SkullBreak and SkullFix, with ablations over RN and SN architectures, and compared with prior work. The best configuration (Seg-SN plus Reg-SN) reports DSC/SDSC/HD95 of 0.936/0.932/1.312 on SkullBreak and 0.945/0.951/1.278 on SkullFix, close to a diffusion-augmented state of the art while claiming under 500 GPU-hours of training. Qualitative results are shown for 11 real clinical defects.

Significance. If the reported numbers hold, the paper makes a useful practical contribution: it demonstrates that a lightweight learnable symmetry prior can bring CNN-based reconstruction close to much more expensive generative augmentation methods, with a plausible clinical runtime. The paper has several strengths: it uses two public benchmarks, includes architecture ablations, separates segmentation-based and registration-based symmetry enforcement, reports inference times and training cost, and explicitly acknowledges that symmetry enforcement can hurt quantitative metrics and cannot fix completely incorrect initializations. The remaining weaknesses are centered on the validation of the symmetry network under the distribution it actually faces during refinement, and on statistical and hyperparameter reporting. I regard the contribution as incremental but potentially valuable for a clinical AI audience; the central claim is currently supported only by single-setting metrics on synthetic defects.

major comments (3)

[Section 2.4, Eq. (3), Table 3] The headline refinement result relies on SN(VRec), where VRec is the volume being optimized, but Table 1 validates SN only on healthy complete skulls and on implants. The paper states in the Discussion that 'the SN has great generalizability,' yet no experiment measures plane accuracy on VRec during the iterative registration loop. Since the plane is recomputed from the current reconstruction, a biased plane could be reinforced rather than corrected, and the improvement of Reg-SN over Seg-SN (for example DSC 0.931 vs 0.904 on SkullBreak) is exactly the claim that needs that missing evidence. Please add an experiment that either (a) compares SN(VRec) with the plane obtained from the corresponding intact skull, reporting angle and distance error, or (b) runs the refinement with a fixed plane from the intact skull and compares final metrics to the proposed feedback scheme. The Discussion acknowledges failure when the initial reconstruction is completely incorrect, but the intermediate partial-error regime is not characterized, and that regime is the load-bearing one for the accuracy figures in Table 4.
[Section 2.6 and Section 3] The paper promises that any claim of statistical significance is supported by a Wilcoxon signed-rank test with p-value below 0.01, but no p-values, test statistics, or confidence intervals appear anywhere in the Results or in Tables 2-4. Several relevant differences are small (for example Seg-SN plus Reg-SN versus Reg-SN alone differs by 0.005 DSC on SkullBreak and 0.004 on SkullFix), so significance is genuinely load-bearing. Please report the test statistics and p-values for the paired comparisons, state whether the comparisons are across cases or across architectures, and address multiple testing if several paired tests are performed.
[Section 2.3, Section 2.4, Tables 2-3] The method's two hyperparameters, alpha in Eq. (2) and lambda in Eq. (3), are fixed at alpha=1 and lambda=1e5 without sensitivity analysis or an explicit selection rule. A reader cannot tell whether the reported gains are due to the symmetry prior itself or to a favorable choice of these two constants. Please include a small sweep (at least over lambda, and preferably alpha) on the validation split, and report the dependence of DSC and HD95 on these parameters. This is also needed to support the claim that registration-based refinement improves over segmentation-based symmetry enforcement, which currently rests on a single lambda value.

minor comments (6)

[Section 2.6] The phrase 'signed-rak test' should be 'signed-rank test'.
[Abstract and Tables 2-4] The abstract uses 'bDSC' while the tables use 'SDSC' for the same quantity; please unify the terminology.
[Table 4] There are citation and formatting errors: the entry for 'Yang et al. [22]' does not appear to match reference [22], and 'Mazzocchetti et al. et al.' and 'Kesornsri et al. et al.' contain duplicated 'et al.'.
[Figure 4] The claim that the refinement leaves only a small hole is a subjective visual reading; please add a quantitative surface-distance map or specify the evaluation conditions.
[Section 2.6] The statement that networks were 'trained until convergence' lacks a convergence criterion; please report the number of epochs or iterations and the early-stopping rule.
[General] No code or data availability statement is provided; since the method is evaluated on public datasets, releasing the symmetry network weights or an implementation would materially aid reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the symmetry network is trained on healthy skulls by a self-supervised objective, the symmetry loss is an additional regularizer for the reconstruction network, and the self-referential registration refinement is explicitly acknowledged as a limitation rather than hidden as a prediction.

full rationale

The central derivation chain is self-contained. The symmetry network (SN) is trained unsupervised on healthy skulls with SL(V, pi)=DSC(V, R(V, pi)) (Eq. 1), which defines the symmetry plane as the one maximizing self-overlap; this is a well-posed self-supervised objective, not a quantity derived from the reconstruction target. The reconstruction network (RN) is trained with Eq. (2), ORec = DSC(Rec, Gt) + alpha*SL(VRec, pi), where the ground-truth term DSC(Rec, Gt) is the primary supervised signal and the symmetry term is an independent regularizer. The ablation in Table 2 shows the symmetry term substantially improves metrics over alpha=0, so it is not redundant by construction. Table 1 evaluates the SN on complete skulls and implants against ground-truth defects, which is an external check, not a fit. The registration refinement in Eq. (3) recomputes pi = SN(VRec) from the volume being optimized, making the objective self-referential; however, the paper explicitly states the failure mode: 'if the initial reconstruction is completely incorrect, the registration-based refinement will not be able to improve the reconstruction', and the final metrics (DSC, HD95) are measured against ground truth, not against the symmetry objective. Self-citations in Table 4 (Wodzinski et al. [34-37]) are prior published baselines; none is used to justify the method's assumptions or to forbid alternatives, and reference [37] is the strongest comparison method, not a supporting premise. No parameter is fitted to test data, and no uniqueness theorem or ansatz is imported from the authors' prior work. The main residual concern is a generalization assumption about SN on partially reconstructed volumes, which is a correctness risk, not a circular step.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The method relies on a small number of domain assumptions about skull symmetry and about the generalization of the symmetry network from healthy to reconstructed volumes. The only hand-set parameters are the loss weight α and regularization weight λ, both fixed without systematic tuning. No new physical entities are proposed.

free parameters (2)

α (symmetry loss weight) = 1
Hand-set to 1 because both Dice loss and symmetry loss are in [0,1]; no tuning or sensitivity analysis performed. Used in Eq. 2.
λ (registration regularization weight) = 1e5
Chosen by hand to constrain the diffusive displacement field during registration-based refinement; only one value tested (Eq. 3, Table 3).

assumptions (3)

domain assumption Human skulls are approximately symmetric, especially the neurocranium, so the intact side can inform the missing side.
Stated in the Introduction with citation [28]; this is the core biological premise of the method.
domain assumption A symmetry network trained on healthy skulls will predict an accurate symmetry plane for reconstructed skulls containing (possibly inaccurate) implants.
Assumed in Sections 2.3 and 2.4; not directly measured on reconstructed volumes, as discussed in the weakest_assumption.
domain assumption The deformation field for the implant during registration-based refinement stays small enough that the diffusive regularizer prevents unrealistic shapes, and the initial reconstruction is at least partially correct.
Implied in Section 2.4 and explicitly stated as a limitation in the Discussion: 'if the initial reconstruction is completely incorrect, the registration-based refinement will not be able to improve the reconstruction.'

how reviews work

0 comments

Cite this review

Pith. "Pith review of Automatic Skull Reconstruction by Deep Learnable Symmetry Enforcement." pith.science (2026). https://pith.science/paper/ISMAJH7A

@misc{pith2026241117342,
  author       = {Pith},
  title        = {Pith review of: Automatic Skull Reconstruction by Deep Learnable Symmetry Enforcement},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ISMAJH7A}},
  note         = {Machine review of arXiv:2411.17342}
}

read the original abstract

Every year, thousands of people suffer from skull damage and require personalized implants to fill the cranial cavity. Unfortunately, the waiting time for reconstruction surgery can extend to several weeks or even months, especially in less developed countries. One factor contributing to the extended waiting period is the intricate process of personalized implant modeling. Currently, the preparation of these implants by experienced biomechanical experts is both costly and time-consuming. Recent advances in artificial intelligence, especially in deep learning, offer promising potential for automating the process. However, deep learning-based cranial reconstruction faces several challenges: (i) the limited size of training datasets, (ii) the high resolution of the volumetric data, and (iii) significant data heterogeneity. In this work, we propose a novel approach to address these challenges by enhancing the reconstruction through learnable symmetry enforcement. We demonstrate that it is possible to train a neural network dedicated to calculating skull symmetry, which can be utilized either as an additional objective function during training or as a post-reconstruction objective during the refinement step. We quantitatively evaluate the proposed method using open SkullBreak and SkullFix datasets, and qualitatively using real clinical cases. The results indicate that the symmetry-preserving reconstruction network achieves considerably better outcomes compared to the baseline (0.94/0.94/1.31 vs 0.84/0.76/2.43 in terms of DSC, bDSC, and HD95). Moreover, the results are comparable to the best-performing methods while requiring significantly fewer computational resources (< 500 vs > 100,000 GPU hours). The proposed method is a considerable contribution to the field of applied artificial intelligence in medicine and is a step toward automatic cranial defect reconstruction in clinical practice.

Figures

Figures reproduced from arXiv: 2411.17342 by the authors.

**Figure 1.** Example of a damage in the frontal bone that cannot be reconstructed directly by reflecting about the symmetry axis because the defect breaks the skull symmetry resulting in missing data. dedicated to efficient and fast personalized reconstruction of cranial defects by enforcing the symmetry of the reconstructed skull. We propose a method based on learnable symmetry using deep vision transformers that can be inte… view at source ↗

**Figure 2.** An overview of the proposed method. The method starts with the segmentation-based [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** Exemplary cases from the three datasets: (i) the SkullFix containing real skulls with synthetic defects in parietal bones, (ii) the SkullBreak consisting of real skulls with synthetic defects in all areas of skull, (iii) the real skulls with real, clinical defects. 2.6 Experimental Setup We performed several ablation studies to verify the impact of the proposed method. First, we calculated the maximum quantitativ… view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: The influence of the symmetry enforcement on a real defective case. Note that the baseline [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 39 canonical work pages

[1]

Ameen, A

W. Ameen, A. Al-Ahmari, M. Mohammed, O. Abdul- hameed, U. Umer, and K. Moiduddin. Design, finite element analysis (fea), and fabrication of custom tita- nium alloy cranial implant using electron beam melt- ing additive manufacturing. Advances in Production Engineering & Management, 13(3):267–278, 2018

work page 2018
[2]

Bayat, S

A. Bayat, S. Shit, A. Kilian, J. T. Liechtenstein, J. S. Kirschke, and B. H. Menze. Cranial implant pre- diction using low-resolution 3d shape completion and high-resolution 2d refinement. In Towards the Autom- atization of Cranial Implant Design in Cranioplasty: First Challenge, AutoImplant 2020, Held in Conjunc- tion with MICCAI 2020, Lima, Peru, Octobe...

work page 2020
[3]

D. J. Bonda, S. Manjila, W. R. Selman, and D. Dean. The recent revolution in the design and manufacture of cranial implants: modern advancements and future directions. Neurosurgery, 77(5):814, 2015

work page 2015
[4]

Chauhan, P

D. Chauhan, P. K. Chattopadhyay, and A. Thakur. Re- construction of cranial defect with patient-specific im- plants: four different cost-effective techniques. Na- tional Journal of Maxillofacial Surgery , 13(1):136– 142, 2022

work page 2022
[5]

De Bonis, P

P. De Bonis, P. Frassanito, A. Mangiola, C. G. Nucci, C. Anile, and A. Pompucci. Cranial repair: how com- plicated is filling a “hole”? Journal of neurotrauma, 29(6):1071–1076, 2012

work page 2012
[6]

D. G. Ellis and M. R. Aizenberg. Deep learning using augmentation via registration: 1st place solution to the autoimplant 2020 challenge. In Towards the Autom- atization of Cranial Implant Design in Cranioplasty: First Challenge, AutoImplant 2020, Held in Conjunc- tion with MICCAI 2020, Lima, Peru, October 8, 2020, Proceedings 1, pages 47–55. Springer, 2020

work page 2020
[7]

Fishman, J

Z. Fishman, J. G. Mainprize, G. Edwards, O. Antonyshyn, M. Hardisty, and C. Whyne. Thickness and design features of clinical cranial implants—what should automated methods strive to replicate? International Journal of Computer Assisted Radiology and Surgery, pages 1–10, 2024

work page 2024
[8]

Friedrich, J

P. Friedrich, J. Wolleb, F. Bieder, F. M. Thieringer, and P. C. Cattin. Point cloud diffusion models for auto- matic implant generation. In International Conference on Medical Image Computing and Computer-Assisted Intervention, pages 112–122. Springer, 2023

work page 2023

Show all 39 references

[9]

Gao, L.-X

L. Gao, L.-X. Zhang, H.-Y . Meng, Y .-H. Ren, Y .-K. Lai, and L. Kobbelt. PRS-Net: Planar reflective sym- metry detection net for 3D models. IEEE transactions on visualization and computer graphics , 27(6):3007– 3018, 2020

2020
[10]

Grammer and R

K. Grammer and R. Thornhill. Human (Homo sapi- ens) facial attractiveness and sexual selection: the role of symmetry and averageness. Journal of comparative psychology, 108(3):233, 1994

1994
[11]

Y . Jin, J. Li, and J. Egger. High-resolution cra- nial implant prediction via patch-wise training. In Towards the Automatization of Cranial Implant De- sign in Cranioplasty: First Challenge, AutoImplant 2020, Held in Conjunction with MICCAI 2020, Lima, Peru, October 8, 2020, ...

2020
[12]

Kesornsri, N

T. Kesornsri, N. Asawalertsak, N. Tantisereepatana, P. Manowongpichate, B. Lohwongwatana, C. Puncre- obutr, T. Achakulvisut, and P. Vateekul. CraNeXt: Au- tomatic Reconstruction of Skull Implants With Skull Categorization Technique. IEEE Access, 12:84907– 84922, 2024

2024
[13]

Kodym, J

O. Kodym, J. Li, A. Pepe, C. Gsaxner, S. Chilamkurthy, J. Egger, and M. ˇSpanˇel. Skullbreak/skullfix–dataset for automatic cranial implant design and a benchmark for volumetric shape learning tasks. Data in Brief, 35:106902, 2021. 9

2021
[14]

Kodym, M

O. Kodym, M. ˇSpanˇel, and A. Herout. Cranial defect reconstruction using cascaded cnn with alignment. In Towards the Automatization of Cranial Implant De- sign in Cranioplasty: First Challenge, AutoImplant 2020, Held in Conjunction with MICCAI 2020, Lima, Peru, October 8, 202...

2020
[15]

Kodym, M

O. Kodym, M. ˇSpanˇel, and A. Herout. Deep learning for cranioplasty in clinical practice: Going from syn- thetic to real patient data. Computers in Biology and Medicine, 137:104766, 2021

2021
[16]

Kroviakov, J

A. Kroviakov, J. Li, and J. Egger. Sparse convolu- tional neural network for skull reconstruction. In To- wards the Automatization of Cranial Implant Design in Cranioplasty II: Second Challenge, AutoImplant 2021, Held in Conjunction with MICCAI 2021, Stras- bourg, France, Octo...

2021
[17]

J. Li, D. G. Ellis, O. Kodym, L. Rauschenbach, C. Rieß, U. Sure, K. H. Wrede, C. M. Alvarez, M. Wodzinski, M. Daniol, et al. Towards clinical ap- plicability and computational efficiency in automatic cranial implant design: An overview of the autoim- plant 2021 cranial implant...

2021
[18]

J. Li, C. Gsaxner, A. Pepe, D. Schmalstieg, J. Kleesiek, and J. Egger. Sparse convolutional neu- ral network for high-resolution skull shape comple- tion and shape super-resolution. Scientific Reports , 13(1):20229, 2023

2023
[19]

J. Li, A. Pepe, C. Gsaxner, G. v. Campe, and J. Eg- ger. A baseline approach for autoimplant: the mic- cai 2020 cranial implant design challenge. In Work- shop on Clinical Image-Based Procedures, pages 75–

2020
[20]

J. Li, A. Pepe, C. Gsaxner, Y . Jin, and J. Egger. Learning to rearrange voxels in binary segmentation masks for smooth manifold triangulation. In To- wards the Automatization of Cranial Implant Design in Cranioplasty II: Second Challenge, AutoImplant 2021, Held in Conjunction...

2021
[21]

J. Li, P. Pimentel, A. Szengel, M. Ehlke, H. Lamecker, S. Zachow, L. Estacio, C. Doenitz, H. Ramm, H. Shi, et al. Autoimplant 2020-first miccai challenge on au- tomatic cranial implant design. IEEE transactions on medical imaging, 40(9):2329–2342, 2021

2020
[22]

Mahdi, A

H. Mahdi, A. Clement, E. Kim, Z. Fishman, C. M. Whyne, J. G. Mainprize, and M. R. Hardisty. A u-net based system for cranial implant design with pre-processing and learned implant filtering. In To- wards the Automatization of Cranial Implant Design in Cranioplasty II: Second C...

2021
[23]

F. M. Marreiros, Y . Heuz´e, M. Verius, C. Unterhofer, W. Freysinger, and W. Recheis. Custom implant de- sign for large cranial defects. International journal of computer assisted radiology and surgery, 11:2217– 2230, 2016

2016
[24]

Matzkin, V

F. Matzkin, V . Newcombe, B. Glocker, and E. Fer- rante. Cranial implant design via virtual craniectomy with shape priors. In Towards the Automatization of Cranial Implant Design in Cranioplasty: First Chal- lenge, AutoImplant 2020, Held in Conjunction with MICCAI 2020, Lima, ...

2020
[25]

Mazzocchetti, M

S. Mazzocchetti, M. Bevini, G. Badiali, G. Lisanti, L. Di Stefano, and S. Salti. Automatic Implant Gener- ation for Cranioplasty via Occupancy Networks.IEEE Access, 2024

2024
[26]

Pathak, C

S. Pathak, C. Sindhura, R. K. S. S. Gorthi, D. V . Ki- ran, and S. Gorthi. Cranial implant design using v- net based region of interest reconstruction. In To- wards the Automatization of Cranial Implant Design in Cranioplasty II: Second Challenge, AutoImplant 2021, Held in Con...

2021
[27]

I. S. Penton-V oak, B. C. Jones, A. C. Little, S. Baker, B. Tiddeman, D. Burt, and D. I. Perrett. Sym- metry, sexual dimorphism in facial proportions and male facial attractiveness. Proceedings of the Royal Society of London. Series B: Biological Sciences , 268(1476):1617–1623, 2001

2001
[28]

Ratajczak, M

M. Ratajczak, M. Ptak, A. Kwiatkowski, K. Kubicki, F. A. Fernandes, J. Wilhelm, M. Dymek, M. Saw- icki, and S. ˙Z´ołkiewski. Symmetry of the human head—are symmetrical models more applicable in nu- merical analysis? Symmetry, 13(7):1252, 2021

2021
[29]

G. E. Rhodes and L. A. Zebrowitz. Facial attractive- ness: Evolutionary, cognitive, and social perspectives. Ablex Publishing, 2002. 10

2002
[30]

Samii and M

M. Samii and M. Tatagiba. Skull base trauma: di- agnosis and management. Neurological research , 24(2):147–156, 2002

2002
[31]

Senck, M

S. Senck, M. Coquerelle, G. W. Weber, and S. Be- nazzi. Virtual reconstruction of very large skull de- fects featuring partly and completely missing mid- sagittal planes. The Anatomical Record, 296(5):745– 758, 2013

2013
[32]

Sulakhe, J

H. Sulakhe, J. Li, J. Egger, and P. Goyal. Crangan: Adversarial point cloud reconstruction for patient- specific cranial implant design. In 2022 44th Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), pages 603–608. IEEE, 2022

2022
[33]

H. Weyl. Symmetry. Journal of the Washington Academy of Sciences, 28(6):253–271, 1938

1938
[34]

Wodzinski, M

M. Wodzinski, M. Daniol, and D. Hemmerling. Im- proving the automatic cranial implant design in cran- ioplasty by linking different datasets. In Towards the Automatization of Cranial Implant Design in Cran- ioplasty II: Second Challenge, AutoImplant 2021, Held in Conjunction w...

2021
[35]

Wodzinski, M

M. Wodzinski, M. Daniol, D. Hemmerling, and M. Socha. High-resolution cranial defect reconstruc- tion by iterative, low-resolution, point cloud comple- tion transformers. In International Conference on Medical Image Computing and Computer-Assisted In- tervention, pages 333–343...

2023
[36]

Wodzinski, M

M. Wodzinski, M. Daniol, M. Socha, D. Hemmer- ling, M. Stanuch, and A. Skalski. Deep learning-based framework for automatic cranial defect reconstruction and implant modeling. Computer Methods and Pro- grams in Biomedicine, 226:107173, 2022

2022
[37]

Wodzinski, K

M. Wodzinski, K. Kwarciak, M. Daniol, and D. Hem- merling. Improving Deep Learning-based Automatic Cranial Defect Reconstruction by Heavy Data Aug- mentation: From Image Registration to Latent Dif- fusion Models. Computers in Biology and Medicine , 2024

2024
[38]

L. Yu, J. Li, and J. Egger. PCA-Skull: 3D skull shape modelling using principal component analysis. In To- wards the Automatization of Cranial Implant Design in Cranioplasty II: Second Challenge, AutoImplant 2021, Held in Conjunction with MICCAI 2021, Stras- bourg, France, Oct...

2021
[39]

Y . Zhou, S. Liu, and Y . Ma. Nerd: Neural 3d reflection symmetry detector. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion, pages 15940–15949, 2021. 11

2021

Pith tools

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

[1] [1]

Ameen, A

W. Ameen, A. Al-Ahmari, M. Mohammed, O. Abdul- hameed, U. Umer, and K. Moiduddin. Design, finite element analysis (fea), and fabrication of custom tita- nium alloy cranial implant using electron beam melt- ing additive manufacturing. Advances in Production Engineering & Management, 13(3):267–278, 2018

work page 2018

[2] [2]

Bayat, S

A. Bayat, S. Shit, A. Kilian, J. T. Liechtenstein, J. S. Kirschke, and B. H. Menze. Cranial implant pre- diction using low-resolution 3d shape completion and high-resolution 2d refinement. In Towards the Autom- atization of Cranial Implant Design in Cranioplasty: First Challenge, AutoImplant 2020, Held in Conjunc- tion with MICCAI 2020, Lima, Peru, Octobe...

work page 2020

[3] [3]

D. J. Bonda, S. Manjila, W. R. Selman, and D. Dean. The recent revolution in the design and manufacture of cranial implants: modern advancements and future directions. Neurosurgery, 77(5):814, 2015

work page 2015

[4] [4]

Chauhan, P

D. Chauhan, P. K. Chattopadhyay, and A. Thakur. Re- construction of cranial defect with patient-specific im- plants: four different cost-effective techniques. Na- tional Journal of Maxillofacial Surgery , 13(1):136– 142, 2022

work page 2022

[5] [5]

De Bonis, P

P. De Bonis, P. Frassanito, A. Mangiola, C. G. Nucci, C. Anile, and A. Pompucci. Cranial repair: how com- plicated is filling a “hole”? Journal of neurotrauma, 29(6):1071–1076, 2012

work page 2012

[6] [6]

D. G. Ellis and M. R. Aizenberg. Deep learning using augmentation via registration: 1st place solution to the autoimplant 2020 challenge. In Towards the Autom- atization of Cranial Implant Design in Cranioplasty: First Challenge, AutoImplant 2020, Held in Conjunc- tion with MICCAI 2020, Lima, Peru, October 8, 2020, Proceedings 1, pages 47–55. Springer, 2020

work page 2020

[7] [7]

Fishman, J

Z. Fishman, J. G. Mainprize, G. Edwards, O. Antonyshyn, M. Hardisty, and C. Whyne. Thickness and design features of clinical cranial implants—what should automated methods strive to replicate? International Journal of Computer Assisted Radiology and Surgery, pages 1–10, 2024

work page 2024

[8] [8]

Friedrich, J

P. Friedrich, J. Wolleb, F. Bieder, F. M. Thieringer, and P. C. Cattin. Point cloud diffusion models for auto- matic implant generation. In International Conference on Medical Image Computing and Computer-Assisted Intervention, pages 112–122. Springer, 2023

work page 2023