Pith. sign in

REVIEW 4 major objections 4 minor 40 references

AutoFFS: Adversarial Deformations for Facial Feminization Surgery Planning

T0 review · 4 major / 4 minor · reviewed 2026-08-02 · deepseek-v4-flash

Pith's one-line read A targeted adversarial attack on sex classifiers yields counterfactual skull morphologies that could ground facial feminization surgery planning in quantitative anatomy.

desk verdict A clever, well-engineered application of adversarial deformations to FFS planning, but the central validation is circular: the main metric is the optimization objective, and the only independent signal (a small human study) shows perceptual shift, not anatomical correctness. read the letter →

arxiv 2603.02288 v2 pith:5P55TOYB submitted 2026-03-02 cs.CV eess.IV

classification cs.CVeess.IV
keywords facialfeminizationsurgerycounterfactualshapeeditingadversarialdeformationsfree-formdeformationsexclassificationskullmorphologysurgicalplanningensembleattack
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

AutoFFS claims that a targeted adversarial attack on an ensemble of binary sex classifiers can serve as a clinically meaningful shape-editing tool: by optimizing a regularized free-form deformation of a skull until the classifiers confidently assign the opposite sex, the method produces a counterfactual skull morphology — the skull 'as if' it belonged to the target sex. The paper argues this provides a quantitative basis for preoperative planning in facial feminization surgery, which currently relies heavily on subjective clinical judgment. Three lines of evidence support the claim: hold-out classifiers flip on 100% of transformed test samples, two new population-level distributional metrics (MFD and MKD) show the generated populations align with their real target populations, and a human perceptual study found raters perceived the intended target sex in 63% of generated skulls. The deformations concentrate in brow, chin, and zygomatic regions, matching anthropological accounts of sexual dimorphism, despite no anatomical priors being imposed.

What carries the argument

Free-form deformation (FFD): a 3D cubic B-spline control lattice that parameterizes a smooth displacement field applied to the skull image. At test time, the control-point offsets are optimized with respect to a loss that combines a smooth worst-case margin over an ensemble of sex classifiers (the targeted adversarial attack) with Jacobian-smoothness and bending-energy regularizers. The deformation field is applied to the input skull X via resampling, X' = X ∘ Φ, and the whole pipeline is differentiable end-to-end through the classifier gradients. The central object is the optimized deformation field itself — it is the counterfactual morphology.

What would settle it

Train the same pipeline on a multi-center CT dataset with broad ethnic, age, and sex diversity, and apply it to male and female CT skulls. If the generated feminized skulls no longer flip hold-out CT-trained classifiers at high rates, or if the deformation hotspots move away from the brow, chin, and zygomatic regions that the paper identifies, the 'feminization direction' learned from the MR cohort is a dataset artifact rather than a general skeletal dimorphism.

Watch

Extended reading notes

Core claim

The central discovery is that the gradient from a targeted adversarial attack on a pre-trained sex classifier ensemble, when channeled through a B-spline free-form deformation and regularized for smoothness and bending energy, identifies a meaningful direction in skull morphology space: the direction that shifts a skull across the learned boundary between male and female distributions. Applying this to a frontal-cropped skull scan produces anatomically plausible counterfactual morphologies in which the brow ridge is attenuated, the chin becomes less projecting, and the zygomatic region narrows for feminization, with the inverse for masculinization. These morphologies fool hold-out classifier

Load-bearing premise

The load-bearing premise is that the sex classifiers trained on the Swiss MR cohort capture true skeletal sexual dimorphism, so that pushing a skull across their decision boundary is a clinically meaningful feminization direction rather than a dataset-specific artifact.

Editorial extensions

If this is right

  • If correct, a surgeon could query 'what would this patient's skull look like as the opposite sex' and obtain a specific, spatially localized map of how much and where the bone should change, rather than relying on qualitative reference morphologies.
  • The method extends beyond feminization to masculinization (FMS) by flipping the target label, and the authors note the framework could apply to the broader class of gender-affirming facial surgery.
  • The ensemble strategy, compared to a single classifier, raises the hold-out flip rate from ~71% to 100%, indicating that fooling a diverse set of learned representations yields more robust and population-consistent deformations.
  • The two new metrics, MFD and MKD, provide a way to evaluate counterfactual shape editing in the absence of ground-truth paired data — a setting common to many medical shape-generation tasks.
  • The observed concentration of deformations in brow, chin, and zygomatic regions, despite no anatomical priors, suggests the learned classifier representations encode genuine skeletal sexual dimorphism that could be used to build quantitative atlases.

Reading between the lines

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

  • A natural stress test: apply AutoFFS to CT-based skulls from diverse ethnic backgrounds. If the deformation hotspots remain anatomically aligned with the known dimorphic regions and the classifiers generalize, the approach would be robust; if not, the learned dimorphism is cohort-specific.
  • Because the method produces a per-patient deformation field, the same pipeline could in principle be inverted or constrained to generate patient-specific osteotomy and cutting-guide designs — an extension the authors mention only as future work.
  • The framework effectively treats the sex classifier as a statistical atlas of dimorphism. One could use the resultant deformation directions to visualize and quantify the in-between along the male-female morphology continuum, which might be of independent interest to anthropologists studying craniofacial sex differences.
  • The reliance on MR-derived bone segmentations raises a question the paper acknowledges: whether the learned dimorphism transfers to CT (the clinical standard). If it does not, the quantitative guidance would need recalibration on CT before use in the operating room.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. The paper proposes AutoFFS, a framework that generates counterfactual skull morphologies for facial feminization/masculinization surgery planning. The method performs a targeted adversarial attack: a regularized B-spline free-form deformation is optimized to flip an ensemble of pre-trained binary sex classifiers from the source sex to the target sex, thereby producing a deformed skull that the classifiers assign to the target class. Evaluation uses hold-out classifier flip rates, two proposed distributional metrics (MFD and MKD) computed in a hold-out classifier's embedding space, a qualitative comparison with anthropological dimorphism descriptions, and a human perceptual study. The central claim is that the resulting counterfactual morphologies provide a quantitative, data-driven foundation for preoperative FFS planning.

Significance. If the central claim were validated, this would be a meaningful contribution to an underserved clinical area. The technical framing—repurposing adversarial attacks as anatomically regularized shape editing—is clever, and the ensemble-plus-FFD pipeline is clearly described. The paper also ships an open project page and proposes new evaluation metrics (MFD/MKD) that could be reused. However, the clinical significance depends entirely on an equivalence that the paper does not establish: that a deformation which flips a sex classifier trained on a single MR cohort is equivalent to feminization as understood by craniofacial surgeons. The current evidence is largely circular or subjective, so the contribution is best viewed as a proof-of-concept in need of external anatomical and clinical validation.

major comments (4)
  1. [§3.1, Classifier-Based Evaluation, Fig. 3] The headline 100% flip rate is effectively the optimization objective reproduced on a different architecture. The hold-out classifiers are trained on the same 444-scan SMSC cohort and same train/test split as the optimization classifiers; adversarial examples are known to transfer across architectures trained on the same data. Consequently, the flip rate demonstrates that the attack transfers, not that the deformation represents a clinically valid female morphology. To support the central claim, the deformations should be compared against independent anatomical ground truth, e.g., quantitative sex-dimorphism measurements from a separate CT dataset or expert craniofacial assessment of whether the induced changes match FFS targets.
  2. [§3.1, Distributional Alignment, Eqs. (8)–(9)] MFD and MKD are computed in the penultimate-layer feature space of a hold-out classifier trained on the same SMSC cohort. Any cohort-specific or scanner-specific bias in that classifier is inherited by these metrics, so they do not constitute an independent test of morphological alignment. Moreover, the intra-class noise floor is MFD≈0.2–0.6, while the generated-to-real MFD is ≈10–11 after reduction from ≈57. The 82–86% reduction is relative to a very large inter-sex baseline; the generated distribution remains an order of magnitude farther from the real target than two random halves of the same real population. Reporting absolute distances and, ideally, metrics on independent landmark or surface measurements would give a more honest picture.
  3. [§3.1, Human Perceptual Evaluation, Table 3] The perceptual study (N=11) shows a statistically suggestive drop in perceived-male accuracy for generated skulls (37%), but this only tests whether raters perceive the target sex; it does not establish anatomical correctness, surgical feasibility, or consistency with actual female skull morphology. The low intra-rater agreement for generated skulls (κ_intra=0.43) indicates the morphologies are ambiguous even to the raters. The claim that these morphologies provide a 'quantitative foundation for preoperative planning' needs additional evidence, such as a structured evaluation by FFS surgeons on the plausibility and operability of the proposed bone changes, rather than only a sex-perception task on renders.
  4. [§4, Limitations] The paper explicitly states that no paired ground-truth or post-operative data exist and declines post-operative comparison. This is the core load-bearing gap. Given that the optimization target and the main evaluation metrics are all derived from classifiers trained on the same single-cohort MR dataset, the current evidence cannot distinguish a genuinely feminizing deformation from a dataset-specific artifact. The paper should either temper the central claim to 'classifier-consistent counterfactuals' or add an external validation loop—for example, testing on an independent CT dataset, comparing the deformation field to published quantitative FFS planning norms, or having surgeons rate whether the deformation matches a feasible operative plan.
minor comments (4)
  1. [Abstract, §2, §3] Typos and formatting: 'we proposeAutoFFS' and 'sexandgender' lack spaces; 'Morphological Fréchet Distance' appears without consistent spacing in the abstract; Figure 2 labels 'Classifier 1Classifier 2...' run together without separators.
  2. [§3.1, Implementation Details] The term 'hold-out classifiers' is slightly misleading: the two evaluation classifiers are trained on the same dataset split as the optimization classifiers, just with different architectures. Please clarify that they are architecture-hold-out, not data-hold-out.
  3. [§3.1, Footnote 4] Hyperparameters (λ_smooth, λ_bend, γ, τ, lattice) were selected based on qualitative assessment of anatomical plausibility. A sensitivity analysis, or at least a statement of how much the qualitative pattern depends on these values, would strengthen the reproducibility.
  4. [§3, Dataset] The dataset is an MS patient cohort from Switzerland; the paper notes this limitation but could also state explicitly that the classifiers therefore learn dimorphism as present in this specific population, and that transfer to other populations is unknown without additional data.

Circularity Check

2 steps flagged · score 5.0 of 10

The headline classifier-flip evaluation is defined in the same sex-classifier space used to generate the deformations; the human perceptual study and qualitative literature check provide partial, but not anatomical, external support.

  1. fitted input called prediction [Sec. 2 'Deformation-Based Targeted Adversarial Attack' / Algorithm 1 and Sec. 3.1 'Classifier-Based Evaluation']
    "we perform a targeted adversarial attack on the pre-trained classifier ensemble ... optimizing our deformation field Φytarget to produce a deformed scan X′ that fools these classifiers into predicting the target class label ytarget. ... For the classifier-based evaluation, we first transform the entire test set and evaluate the resulting class probabilities using the hold-out classifiers ... achieving a flip rate of 100%."

    The deformation is obtained by maximizing target-class logits of an ensemble of sex classifiers (Algorithm 1). 'Feminized skull' is therefore operationally defined as 'classified as female by such networks'. The main evaluation then re-measures classifier output on hold-out architectures trained on the same split and with the same labels from the same 444-scan cohort. Because adversarial examples are known to transfer across architectures, a deformation that pushes one set of sex classifiers beyond a logit margin is expected to also flip similarly trained hold-outs. Consequently, the 100% flip rate is partly the optimization objective renamed as validation, not an independent test of anatomical feminization.

  2. other [Sec. 3.1 'Distributional Alignment', Table 2]
    "Both operate on penultimate-layer feature embeddings extracted from a hold-out classification network. ... After feminization (m→f), MFD drops to 10.32 (CI:4.36–18.11) and MKD to 96.68 (CI:32.98–117.93), corresponding to reductions of 83% and 86% from the inter-sex baseline."

    MFD and MKD are computed in the hidden space of a sex classifier trained on the same cohort and labels. The generated morphologies were not directly optimized for these metrics, so this is a consistency check rather than a mathematical identity. However, because the deformation is constructed to cross the decision boundary of the same kind of sex classifier, shifting embeddings toward the target side in that learned space is the expected effect of the adversarial objective. The metrics therefore report alignment with the classifier's sex representation, not with an independent anatomical ground truth.

full rationale

The central derivation chain defines the target 'female morphology' through binary sex classifiers trained on the SMSC cohort, generates a deformation by adversarially maximizing those classifiers' target logits, and then validates the result primarily with (a) flip rates of similar hold-out classifiers and (b) distributional distances in the embedding space of a similar sex classifier. The paper itself concedes in Sec. 3.1: 'Evaluating our method poses an inherent challenge, as no paired data exists that would permit a direct comparison between generated and ground-truth morphologies.' This makes the classifier-based and feature-space evaluations partially self-referential: the goal, the optimization, and the success metric all live in the same learned sex-classifier space. There is no self-citation chain or imported uniqueness theorem here. The human perceptual study (N=11) is an external signal, and the qualitative comparison to anthropological literature is also independent, which prevents a higher circularity score. But the perceptual study only shows a perceived-sex shift, not that the deformations match actual female skeletal morphology or constitute a clinically valid FFS plan. Overall, the paper's quantitative 'prediction' of target-sex morphology is substantially, though not entirely, an artifact of its own classifier-defined objective.

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

The method leans on several domain assumptions about the dataset (MS cohort as surrogate) and about the semantic meaning of the classifiers' decision boundary. No new physical entities are introduced. The main free parameters are hand-tuned optimization hyperparameters, with no sensitivity analysis.

free parameters (6)
  • regularization weights λ_smooth, λ_bend = 1e8 each
    Chosen based on qualitative assessment of anatomical plausibility (footnote 4); no quantitative tuning criterion; directly control deformation smoothness.
  • logit margin γ = 4.5
    Hand-set target margin for the adversarial attack; affects how far into target distribution the deformation is pushed.
  • temperature τ = 1
    Smoothness of log-sum-exp approximation; hand-set.
  • optimization hyperparameters (lr, steps, lattice) = α=5e-3, S=100, 32^3 lattice
    Chosen ad hoc; no ablation or sensitivity analysis reported.
  • classifier training hyperparameters = batch 16, epochs 100, lr 1e-4, augmentation ranges, mask ratio 50%
    Standard but not swept; the classifier ensemble is the normative reference for the deformation direction.
  • PCA components for MFD = 32
    Retained to explain 99.8% variance; the number itself is selected.
assumptions (5)
  • domain assumption MS pathology does not alter cranial bone morphology, so SMSC MR scans are a valid surrogate for skull shape variation
    Section 3 Dataset; load-bearing for dataset representativeness; if false, the learned sex classifier reflects disease-related shape differences.
  • domain assumption MR-derived bone segmentations from T1-weighted imaging approximate CT-derived skull models
    Section 3 Dataset, citing [38]; the entire pipeline operates on MR segmentations rather than CT, the clinical reference.
  • domain assumption Training a binary sex classifier on these skull images yields representations of true sexual dimorphism
    Section 2 Classifier Training; the optimization target is defined entirely by these classifiers; if they exploit non-dimorphic artifacts, the generated shapes are not anatomically feminine/masculine.
  • ad hoc to paper Regularized adversarial deformations toward classifier-defined target are clinically meaningful feminization/masculinization
    Section 3.1 qualitative assessment; the paper provides no ground truth linking classifier-flipping deformations to actual surgical outcomes.
  • domain assumption Anterior facial features dominate sex perception, justifying cropping to the frontal half of the skull
    Section 3 Dataset, citing [18,19]; affects both the method and the human perceptual study.

how reviews work

0 comments
Cite this review

Pith. "Pith review of AutoFFS: Adversarial Deformations for Facial Feminization Surgery Planning." pith.science (2026). https://pith.science/paper/5P55TOYB

@misc{pith2026260302288,
  author       = {Pith},
  title        = {Pith review of: AutoFFS: Adversarial Deformations for Facial Feminization Surgery Planning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5P55TOYB}},
  note         = {Machine review of arXiv:2603.02288}
}
read the original abstract

Facial feminization surgery (FFS) is a key component of gender affirmation for transgender and gender diverse patients, aiming to reshape craniofacial structures toward a female morphology. Current surgical planning procedures largely rely on subjective clinical assessment, lacking quantitative and reproducible anatomical guidance. We therefore propose AutoFFS, a novel data-driven framework that generates counterfactual skull morphologies through adversarial free-form deformations. Our method performs a deformation-based targeted adversarial attack on an ensemble of pre-trained binary sex classifiers that learned sexual dimorphism, effectively transforming individual skull shapes toward the target sex. The generated counterfactual skull morphologies provide a quantitative foundation for preoperative planning in FFS, driving advances in this largely overlooked patient group. We validate our approach through classifier-based evaluation, propose Morphological Fr\'{e}chet Distance (MFD) and Morphological Kernel Distance (MKD) to evaluate distributional alignment of generated and real populations, and perform a human perceptual study, confirming that the generated morphologies exhibit target sex characteristics.

Figures

Figures reproduced from arXiv: 2603.02288 by the authors.

Figure 1
Figure 1. Overview of the proposed pipeline. We perform a [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Illustration of the targeted adversarial attack as a trajectory through a [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Class probabilities from the hold-out classifier ensemble after applying our [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Example transformations. (Left, top) FFS example from a lateral view. (Left, bottom) FMS example from a frontal view. (Right) Result without B-spline formulation and regularization. the method to collapse to noise-like perturbation patterns, confirming the ne￾cessity o…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 4 linked inside Pith

  1. [1]

    Quality of Life Research 19(7), 1019–1024 (2010)

    Ainsworth, T.A., Spiegel, J.H.: Quality of life of individuals with and without facial feminization surgery or gender reassignment surgery. Quality of Life Research 19(7), 1019–1024 (2010)

  2. [2]

    Surgery170(1), 336–340 (2021)

    Akhavan, A.A., Sandhu, S., Ndem, I., Ogunleye, A.A.: A review of gender affirma- tion surgery: What we know, and what we need to know. Surgery170(1), 336–340 (2021)

  3. [3]

    IEEE transactions on medical imaging38(8), 1788–1800 (2019)

    Balakrishnan, G., Zhao, A., Sabuncu, M.R., Guttag, J., Dalca, A.V.: Voxelmorph: a learning framework for deformable medical image registration. IEEE transactions on medical imaging38(8), 1788–1800 (2019)

  4. [4]

    Facial Plastic Surgery & Aesthetic Medicine24(S2), S–24 (2022)

    Bannister, J.J., et al.: Sex differences in adult facial three-dimensional morphology: application to gender-affirming facial surgery. Facial Plastic Surgery & Aesthetic Medicine24(S2), S–24 (2022)

  5. [5]

    Clinics in plastic surgery34(3), 557–564 (2007)

    Becking, A.G., Tuinzing, D.B., Hage, J.J., Gooren, L.J.: Transgender feminization of the facial skeleton. Clinics in plastic surgery34(3), 557–564 (2007)

  6. [6]

    Clinical Oral Investigations29(8), 1–11 (2025)

    Beyer, M., et al.: Innovations in gender affirmation: Ai-enhanced surgical guides for mandibular facial feminization surgery. Clinical Oral Investigations29(8), 1–11 (2025)

  7. [7]

    arXiv preprint arXiv:1801.01401 (2018)

    Bińkowski, M., Sutherland, D.J., Arbel, M., Gretton, A.: Demystifying mmd gans. arXiv preprint arXiv:1801.01401 (2018)

  8. [8]

    Cambridge University Press (2004)

    Boyd, S., Vandenberghe, L.: Convex Optimization. Cambridge University Press (2004)

Show all 40 references
  1. [9]

    In: 2017 ieee symposium on security and privacy (sp)

    Carlini, N., Wagner, D.: Towards evaluating the robustness of neural networks. In: 2017 ieee symposium on security and privacy (sp). pp. 39–57. Ieee (2017)

  2. [10]

    Scientific Reports14(1), 16879 (2024)

    Çiftçi, R., et al.: Human gender estimation from ct images of skull using deep feature selection and feature fusion. Scientific Reports14(1), 16879 (2024)

  3. [11]

    Oral and Maxillofacial Surgery Clinics36(2), 171–182 (2024)

    Cronin, B.J., Lee, J.C.: Preoperative radiology and virtual surgical planning. Oral and Maxillofacial Surgery Clinics36(2), 171–182 (2024)

  4. [12]

    lips, midface, mandible, chin, and laryngeal prominence

    Dang, B.N., et al.: Evaluation and treatment of facial feminization surgery: part ii. lips, midface, mandible, chin, and laryngeal prominence. Archives of Plastic Surgery49(01), 5–11 (2022)

  5. [13]

    Scientific reports13(1), 16847 (2023)

    Del Bove, A., et al.: Mapping sexual dimorphism signal in the human cranium. Scientific reports13(1), 16847 (2023)

  6. [14]

    PloS one11(3), e0152347 (2016)

    Disanto, G., et al.: The swiss multiple sclerosis cohort-study (smsc): a prospective swiss wide investigation of key phases in disease evolution and new treatment options. PloS one11(3), e0152347 (2016)

  7. [15]

    The American Journal of Bioethics18(12), 3–9 (2018) 14 P

    Dubov, A., Fraenkel, L.: Facial feminization surgery: the ethics of gatekeeping in transgender health. The American Journal of Bioethics18(12), 3–9 (2018) 14 P. Friedrich et al

  8. [16]

    Journal of Plastic, Reconstructive & Aesthetic Surgery75(7), e1–e14 (2022)

    Escandón, J.M., et al.: Applications of three-dimensional surgical planning in facial feminization surgery: a systematic review. Journal of Plastic, Reconstructive & Aesthetic Surgery75(7), e1–e14 (2022)

  9. [17]

    Journal of Craniofacial Surgery33(1), e84–e87 (2022)

    Ganry, L., Cömert, M.: Low-cost and simple frontal sinus surgical cutting guide modeling for anterior cranioplasty in facial feminization surgery: how to do it. Journal of Craniofacial Surgery33(1), e84–e87 (2022)

  10. [18]

    American journal of phys- ical anthropology147(4), 661–670 (2012)

    Garvin, H.M., Ruff, C.B.: Sexual dimorphism in skeletal browridge and chin mor- phologies determined using a new quantitative method. American journal of phys- ical anthropology147(4), 661–670 (2012)

  11. [19]

    Vision Research201, 108127 (2022)

    González-Álvarez, J., Sos-Peña, R.: Sex perception from facial structure: Catego- rization with and without skin texture and color. Vision Research201, 108127 (2022)

  12. [20]

    The journal of machine learning research13(1), 723–773 (2012)

    Gretton, A., Borgwardt, K.M., Rasch, M.J., Schölkopf, B., Smola, A.: A kernel two-sample test. The journal of machine learning research13(1), 723–773 (2012)

  13. [21]

    Facial Plastic Surgery & Aesthetic Medicine26(6), e818–e824 (2024)

    Gutiérrez-Santamaría, J., et al.: Shaping the lower jaw border with customized cutting guides: development, validation, and application in facial gender-affirming surgery. Facial Plastic Surgery & Aesthetic Medicine26(6), e818–e824 (2024)

  14. [22]

    He,K.,Zhang,X.,Ren,S.,Sun,J.:Deepresiduallearningforimagerecognition.In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 770–778 (2016)

  15. [23]

    Advances in neural information processing systems30(2017)

    Heusel,M.,Ramsauer,H.,Unterthiner,T.,Nessler,B.,Hochreiter,S.:Ganstrained by a two time-scale update rule converge to a local nash equilibrium. Advances in neural information processing systems30(2017)

  16. [24]

    In: Proceedings of the IEEE conference on computer vision and pattern recognition

    Hu, J., Shen, L., Sun, G.: Squeeze-and-excitation networks. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 7132–7141 (2018)

  17. [25]

    arXiv preprint arXiv:1412.6980 (2014)

    Kingma, D.P., Ba, J.L.: Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 (2014)

  18. [26]

    In: The NIPS’17 Competition: Building Intelligent Systems, pp

    Kurakin, A., et al.: Adversarial attacks and defences competition. In: The NIPS’17 Competition: Building Intelligent Systems, pp. 195–231. Springer (2018)

  19. [27]

    Annals of Translational Medicine 9(7), 614 (2021)

    Kuruoglu, D., et al.: Point of care virtual surgical planning and 3d printing in facial gender confirmation surgery: a narrative review. Annals of Translational Medicine 9(7), 614 (2021)

  20. [28]

    arXiv preprint arXiv:1608.03983 (2016)

    Loshchilov, I., Hutter, F.: Sgdr: Stochastic gradient descent with warm restarts. arXiv preprint arXiv:1608.03983 (2016)

  21. [29]

    arXiv preprint arXiv:1711.05101 (2017)

    Loshchilov, I., Hutter, F.: Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101 (2017)

  22. [30]

    Annals of 3D Printed Medicine 15, 100164 (2024)

    Marin-Montealegre, V., et al.: 3d surgical planning method for lower jaw os- teotomies applied to facial feminization surgery. Annals of 3D Printed Medicine 15, 100164 (2024)

  23. [31]

    Scientific Reports14(1), 13707 (2024)

    Noel, L., et al.: Sex classification of 3d skull images using deep neural networks. Scientific Reports14(1), 13707 (2024)

  24. [32]

    In: International Conference on Machine Learning

    Pang, T., Xu, K., Du, C., Chen, N., Zhu, J.: Improving adversarial robustness via promoting ensemble diversity. In: International Conference on Machine Learning. pp. 4970–4979. PMLR (2019)

  25. [33]

    Plastic and Reconstructive Surgery pp

    Rancu, A.L., et al.: Virtual surgical planning for facial feminization surgery: Corre- lation between three-dimensional plans and results varies across procedures. Plastic and Reconstructive Surgery pp. 10–1097 (2021)

  26. [34]

    IEEE Transactions on Medical Imaging18(8), 712–721 (1999) Adversarial Deformations for Facial Feminization Surgery Planning 15

    Rueckert, D., et al.: Nonrigid registration using free-form deformations: application to breast mr images. IEEE Transactions on Medical Imaging18(8), 712–721 (1999) Adversarial Deformations for Facial Feminization Surgery Planning 15

  27. [35]

    Aesthetic Plastic Surgery48(10), 1899–1905 (2024)

    Sanz, D., Ângelo, D.F., Marques, R., Maffia, F., Cardoso, H.J.: F-chin/feminizing the chin: A genioplasty technique with virtual planning for male-to-female trans- gender patients. Aesthetic Plastic Surgery48(10), 1899–1905 (2024)

  28. [36]

    In: Seminars in Plastic Surgery

    Sharaf, B., Kuruoglu, D., Bite, U., Morris, J.M.: Point of care virtual surgical planning and 3d printing in facial feminization surgery. In: Seminars in Plastic Surgery. vol. 36, pp. 164–168. Thieme Medical Publishers, Inc. (2022)

  29. [37]

    Imaging Neuroscience2, 1–21 (2024)

    Stolte, S.E., et al.: Precise and rapid whole-head segmentation from magnetic res- onance images of older adults using deep learning. Imaging Neuroscience2, 1–21 (2024)

  30. [38]

    Craniomaxillofacial trauma & reconstruction18(3), 40 (2025)

    Ten Brink, R.S., Merema, B.J., den Otter, M.E., van Veldhuizen, W.A., Witjes, M.J., Kraeima, J.: Towards mri-only mandibular resection planning: Ct-like bone segmentation from routine t1 mri images using deep learning. Craniomaxillofacial trauma & reconstruction18(3), 40 (2025)

  31. [39]

    Plastic and Reconstructive Surgery–Global Open10(3), e4210 (2022)

    Tirrell, A.R., El Hawa, A.A.A., Bekeny, J.C., Chang, B.L., Del Corral, G.: Facial feminization surgery: A systematic review of perioperative surgical planning and outcomes. Plastic and Reconstructive Surgery–Global Open10(3), e4210 (2022)

  32. [40]

    American Journal of Physical Anthropology: The Official Publication of the American Association of Physical Anthropologists136(1), 39–50 (2008)

    Walker, P.L.: Sexing skulls using discriminant function analysis of visually assessed traits. American Journal of Physical Anthropology: The Official Publication of the American Association of Physical Anthropologists136(1), 39–50 (2008)

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.