Pith. sign in

REVIEW 4 major objections 9 minor 173 references

GNM Head: A Generative aNthropometric Model of the human head

T0 review · 4 major / 9 minor · reviewed 2026-07-30 · grok-4.5

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

Pith's one-line read A public 3D head model that includes eyes, teeth, and tongue fits real face scans more tightly than the standard outer-shell models.

desk verdict Solid public full-head 3DMM with real internal anatomy and a clean FLAME win on home data; the mm gap is partly ecosystem-matched, but the release still matters. read the letter →

arxiv 2607.23687 v1 pith:37HIBKJC submitted 2026-07-26 cs.CV cs.GR

Stylianos Ploumpis , Jan Bednarik , Gaspard Zoss , Ruslan Guseinov , Luca Prasso , Prashanth Chandran , Oliver Boyne , Vasileios Choutas
show 21 more authors
Timo Bolkart Daoye Wang Menglei Chai Di Qiu Sebastian Winberg Gilles Rainer Lewis Bridgeman Delio Vicini Jérémy Riviere Yannick Boetzel Alexander Koumis Jay Busch Cynthia Herrera Jacob Still Scott Ysebert Peter Lincoln Sergio Orts Escolano Christoph Rhemann Erroll Wood Thabo Beeler Stefanos Zafeiriou
This is my paper
classification cs.CVcs.GR
keywords 3Dmorphablemodelparametricheadfacialreconstructionteethandtonguemodelingoculargeometryexpressionblendshapessemanticsamplingcomputervision
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

Most public 3D face models treat the head as a hollow skin shell. They leave out teeth, tongue, and detailed eyes, so animation, speech, and generative AI often fail when the mouth opens or the gaze matters. This paper introduces GNM, a single linear parametric model of the full head—skin, neck, eyeballs, teeth, and tongue—built from thousands of high-resolution multi-view scans plus artist-made internal anatomy. On a held-out set of 15,000 scans it reports lower reconstruction error than the widely used FLAME baselines, and the authors release the full framework for research and commercial use. A dual conditional autoencoder further maps everyday labels (gender, ethnicity, action expressions) onto the model’s coefficients so users need not dial raw principal-component knobs. The practical claim is that a controllable, anatomically complete head prior is now available as open infrastructure for reconstruction, neural rendering, and generative conditioning.

What carries the argument

GNM itself: a linear identity-and-expression 3D morphable model with part-based bases (head/eye/teeth identity; left-eye, right-eye, lower-face, tongue, pupil expression), four-joint linear blend skinning, and hybrid real-plus-artist submodels for teeth, tongue, and two-sphere eyeballs, plus a dual-CVAE semantic sampler over demographic and action labels.

What would settle it

Fit GNM and FLAME to a large held-out set of independently captured full-head scans that include true intra-oral and ocular geometry (not just outer skin), then check whether GNM’s mean scan-to-mesh advantage on the face still holds and whether teeth, tongue, and cornea errors stay low when those structures are directly measured.

Watch

Extended reading notes

Core claim

GNM shows that a linear morphable head spanning outer skin plus ocular and intra-oral structure, trained on high-fidelity registrations and specialized internal assets, reconstructs held-out 3D face scans more accurately than FLAME while remaining a standard, publicly released parametric prior for animation and vision pipelines.

Load-bearing premise

Artist-made and procedurally generated teeth, tongue, and eye geometry, registered mostly from outer-face cues and sparse landmarks under occlusion, are assumed faithful enough that the learned internal subspaces track real human variation rather than the artist prior.

Editorial extensions

If this is right

  • Speech-driven and open-mouth animation can condition on explicit teeth and tongue geometry instead of inventing mouth interiors.
  • Single- and multi-view face fitting pipelines gain denser landmark anchors inside the mouth and on the eyes, improving extreme-expression tracking.
  • Generative image and neural-rendering systems can use one shared, controllable 3D prior that keeps lips, teeth, tongue, and pupils physically coordinated.
  • Privacy-safe synthetic head datasets can sample identities and expressions with internal anatomy included rather than baked into texture.
  • A released dual-CVAE sampler lets non-experts drive the model from gender, ethnicity, and named action labels without editing raw PCA coefficients.

Reading between the lines

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

  • If the internal subspaces are only weakly constrained by real scans, downstream speech and gaze systems may still need per-subject refinement of teeth and cornea before clinical or high-end VFX use.
  • Part-based expression masks that stop mouth motion from leaking into the eyes suggest a general recipe for other body models where global PCA couples distant regions.
  • Open release of both the mesh model and dense landmarks could become the default geometric scaffold for head Gaussian-splat and NeRF avatars, shifting the bottleneck from topology to appearance.
  • Broader demographic coverage beyond the four ethnicity and binary gender labels used in the sampler will be needed before the semantic interface is treated as demographically neutral.
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.

Referee Report

4 major / 9 minor

Summary. The manuscript introduces GNM, a linear parametric model of the full human head — skin, neck, eyeballs (sclera/cornea/limbus/pupil), teeth, and tongue — built from a proprietary corpus of ~5,000 subjects / ~150,000 high-resolution multi-view scans plus artist-authored synthetic assets for the internal anatomy. Architecturally it is a standard PCA + LBS 3DMM (Eqs. 1–3) with composite, part-stacked identity and expression bases (Eqs. 4–5), regional uncentered expression PCA, a two-sphere parametric eyeball, a PCA teeth model over 5,000 procedurally generated dental shapes, and a tongue basis from rig fits to registrations and external sparse keypoints. A dual-CVAE Semantic Sampler maps demographic/expression attributes to coefficients. The headline quantitative claim (Table 2) is mean scan-to-mesh distance of 0.748 mm versus 0.968/0.971 mm for two FLAME variants on 15,000 held-out scans, with regional/demographic breakdowns, generalization and specificity curves, a synthetic single-view landmark fitting benchmark (Table 3), and qualitative multi-view and in-the-wild results. The model is publicly released, including commercial licensing.

Significance. If the results hold, this is a substantial contribution. It is, to my knowledge, the first publicly released linear head model integrating statistically modeled teeth, tongue, and ocular structures in one manifold, trained on an order of magnitude more high-resolution data than FLAME, and released under a license permitting commercial use — a genuine enabler for the community. The registration and fitting pipelines are described in implementable detail, the model ships as a runnable artifact (github.com/google/GNM), and the evaluation is unusually thorough in coverage (15k held-out scans, region/demographic/expression breakdowns, per-component generalization curves, specificity against a real database). The part-based formulation is simple, compatible with standard graphics pipelines, and the internal-anatomy submodels address a real and acknowledged gap. The principal caveat is that essentially all quantitative evidence lives inside the authors' own capture/registration ecosystem, and the paper's most novel components (teeth, tongue, eyes) are validated only qualitatively.

major comments (4)
  1. [§4.1.1–4.1.2, Table 2] §4.1.1–§4.1.2, Table 2: the headline 0.748 vs ~0.97 mm comparison is measured entirely within the GNM ecosystem, and three asymmetries make the margin hard to attribute to model quality. (i) The held-out scans come from the same 22-camera rig, expression protocol, and demographic pool as training — held out in identity, not in distribution; FLAME was trained on different, older data, so part of the gap is dataset quality rather than model architecture, yet the claim 'SotA performance on fitting target 3D face scans' will be read generally. (ii) The test registrations are produced by the pipeline of §3.2, which explicitly 'relies on the GNM as a geometric prior,' and stage-1 fitting to these GNM-topology registrations is 'trivial' for GNM by the authors' own words, while FLAME is bootstrapped through hand-annotated sparse correspondences plus non-rigid template alignment. (iii) Stage 2 (I
  2. [§4.1.2, Table 1 vs Table 2] Capacity asymmetry: Table 2 compares full models, but GNM carries 253 identity and 382 expression components (Table 1) against FLAME's standard 300 identity / 100 expression. GNM's ~3.8× larger expression basis is a plausible alternative explanation for the improvement, which is concentrated in the lower face and open-mouth expressions (Figures 10–12) — precisely where extra expression capacity helps most. Figure 13's generalization curves partially address this, but no matched-component operating point is reported. Please add a capacity-controlled comparison (e.g., GNM truncated to 100 expression components, or FLAME retrained/extended on the authors' registered data) so the reader can separate the contribution of the data from the contribution of basis size.
  3. [§3.7 and §3.8] The paper's central novelty — the internal anatomy — is the least validated part of the manuscript, and the main quantitative evaluation structurally excludes it. The teeth identity basis is PCA over 5,000 procedurally generated artist shapes (§3.7.1); the tongue basis is PCA over ~2.5K rig fits to selected registrations and sparse external keypoints (§3.7.2); the eyeball identity basis is PCA over samples from a parametric two-sphere model with fixed scleral radius and hand-chosen physiological Gaussians (§3.7.3). These subspaces may encode the artist prior (plus correlation with the outer face via the §3.8 backfill heuristics) rather than real anatomical variation. Meanwhile §4.1 restricts all quantitative comparison to 'a common facial region' (Figure 12 mask), which excludes the inner mouth — so Tables 2–3 say nothing about the teeth/tongue/eye components that distinguish GNM. Some
  4. [§4.2.1, Table 3] The provenance of the 2,000 synthetic test images is not stated. If the identities/expressions used to render them were generated in GNM's own coefficient space (or rendered from GNM-topology geometry), then fitting GNM to their ground-truth landmarks is close to self-reconstruction and the 1.68 vs 2.17 mm margin is not independent evidence of prior quality. Please specify exactly how the synthetic geometry was generated, and if it is GNM-derived, add a variant rendered from non-GNM geometry (e.g., FLAME-space identities or an independent avatar system).
minor comments (9)
  1. [Figure 11] The age axis labels read '20-35, 36-55, 36-56+' — the third bin is presumably '56+'; also check the boundary overlap between bins two and three.
  2. [§3.5] §3.5 references 'the teeth dataset X_T of Section 3.7.2', but teeth modeling is §3.7.1 while §3.7.2 is the tongue. Please fix the cross-reference.
  3. [§4.3.1–4.3.2] Typo 'Architecuture'; the expression encoder layer sizes are listed as '512, 128, 256, 64' — the non-monotonic middle layers look like a typo, please verify. Also 'ehtnicity' in §4.3.2.
  4. [Table 1] In the current layout the identity totals ('80 253') and expression totals ('31 1 382') run together and are hard to parse; the pupil component count of 1 should be visually separated from the total of 382.
  5. [§4.1.3] Specificity is evaluated for the identity basis only. Since GNM's expression basis is regional and uncentered, an expression-specificity analysis (random expression samples vs. real expressive scans) would be informative, particularly to show the regional masks do not produce implausible blends at region boundaries.
  6. [§3.7.3] The physiological parameter values (r_s ≈ 14.6 mm; r_l μ=6.0/σ=0.44; r_c μ=8.5/σ=0.73) are stated without citation; please cite the ophthalmologic source. Likewise the claim that LED glints are 'primary features in gaze-target prediction' should be supported.
  7. [§4.2, Eq. (6)] The loss weights w_lan, w_prior, w_anat, w_temp (Eq. 6) are deferred to Wood et al. (2022); since the internal-anatomy collision terms are new here, their formulation and weights should be specified in this paper rather than by reference.
  8. [§2/§4] No quantitative comparison is offered against specialized part models the paper itself cites (e.g., Wu et al. 2016 / Zhang et al. 2022 for teeth, Medina et al. 2022 / Ploumpis et al. 2022 for tongue). Even a small per-part benchmark would strengthen the claim that unification does not sacrifice regional fidelity.
  9. [§1/§5] The public release is a major strength; please state the exact license name/version in the paper (not only in the repository) and clarify whether the registered dataset, the dense landmark predictor, and the Semantic Sampler weights are included, since §5 promises a future release of the landmarks.

Circularity Check

1 steps flagged · score 1.0 of 10

Standard 3DMM construction and held-out fitting; no derivation that reduces to its inputs by construction. Only mild evaluation asymmetry from coregistration priors.

  1. other [§3.2 Head Registration; §4.1.2 Stage 1 fitting]
    "Following FLAME (Li et al., 2017), GNM employs an iterative coregistration cycle (Hirshberg et al., 2012) alternating between face registration and statistical model building. ... This iterative coregistration relies on the GNM as a geometric prior to accurately deform internal components from sparse signals such as visible teeth landmarks. ... As the target registrations are provided in the GNM topology, aligning the GNM model to the registration is trivial, as the vertices of the model and the registered shape are already in correspondence."

    Test meshes used for the Table 2 comparison are registered with a pipeline that uses the evolving GNM as prior and are already in GNM vertex correspondence, so Stage-1 error for GNM is near-zero by construction while FLAME requires sparse hand correspondences and non-rigid bootstrap. This is a mild evaluation asymmetry (partially mitigated by Stage-2 ICP to the raw scan), not a scientific prediction equal to a fitted input; recorded only as the paper's nearest circularity-adjacent step.

full rationale

GNM is an engineering/statistical morphable model: identity and expression bases are ordinary (region-split) PCA on registered meshes plus LBS, with artist-synthetic teeth/tongue/eye assets explicitly injected (Secs. 3.3–3.8). The headline claim is empirical SotA scan-to-mesh error versus FLAME on 15k held-out scans (Table 2), not a first-principles prediction. Specificity samples N(0,Σ) and measures distance to a real scan database; the Semantic Sampler is a trained dual-CVAE, not a claimed closed-form forecast. The only mild circularity-adjacent point is the usual 3DMM coregistration loop and the fact that test registrations live in GNM topology and were produced with GNM as geometric prior, so Stage-1 alignment is trivial for GNM and harder for FLAME—but Stage-2 ICP is run on the raw scan for both models with the same optimizer budget, and the paper does not relabel a fitted quantity as an independent prediction. No self-definitional identity, no uniqueness theorem imported from the authors, no ansatz smuggled in as derivation. Score 1 reflects that single methodological self-reference, not a load-bearing circular claim.

Assumptions & free parameters 6 free parameters · 7 assumptions · 2 invented entities

GNM’s claims rest on classical linear 3DMM assumptions, a large private capture corpus, and several engineering priors for unobservable internal anatomy. The ledger separates standard math (PCA, LBS, Procrustes), domain assumptions common to face modeling, and paper-specific choices (fixed scleral radius, artist dental generator, regional masks, CVAE mixup schedule) that shape the released manifold.

free parameters (6)
  • Identity/expression retained component counts = I: 253 total; E: 382 total
    Per-region PCA truncated to explain ~99% variance; final dims fixed in Table 1 (head 170, eye 3, teeth 80; left/right eye 100 each, lower face 150, tongue 31, pupil 1).
  • Scleral radius and cornea/limbus Gaussians = rs≈14.6 mm; rl~N(6.0,0.44^2); rc~N(8.5,0.73^2)
    Eyeball identity basis built from physiological distributions with sclera held fixed for eyelid compatibility (Section 3.7.3).
  • Pupil expression coefficient range = [-3,3] → point to full iris
    Hand-crafted single-basis pupil dilation mapped to coefficient range [-3,3] with specified radii at endpoints (Section 3.7.3).
  • Fitting loss weights and Adam hyperparameters = lr=1e-3, 5000 steps; w_temp=0 for single frame
    Landmark/prior/anatomy/temporal weights and 5000-step Adam lr 1e-3 govern reported reconstruction numbers (Sections 4.1–4.2).
  • CVAE architecture and KL/mixup schedule = w_KL max 0.05; latent 64-D; 12k training samples
    Dual-CVAE layer sizes, 64-D latents, w_KL warmup to 0.05 over 4000 steps, Beta(0.2,0.2) mixup on conditions (Section 4.3).
  • Number of synthetic dental shapes / tongue set size = NT=5000; NG~2500
    Artist procedural teeth NT=5000 and tongue displacement set NG~2.5k define internal PCA subspaces (Sections 3.7.1–3.7.2).
assumptions (7)
  • domain assumption Human head shape/expression variation is well approximated by linear PCA offsets on a fixed topology plus LBS with a 4-joint skeleton.
    Core model definition M(Θ;Ψ) in Section 3.3 follows standard 3DMM/FLAME-style factorization.
  • domain assumption Rigid Procrustes/stabilization can remove pose so remaining vertex deltas are pure identity or expression.
    Used for neutral identity PCA and expression deltas; expression path adds confidence-map then semi-automatic PCA cleanup (Sections 3.5–3.6).
  • ad hoc to paper Uncentered regional expression PCA with soft masks yields local control without requiring a neutral mean face that has closed eyes/open mouth.
    Section 3.6 explicitly avoids centering XE(r) for controllability; regional orthogonality is only within-region.
  • ad hoc to paper Under identity change, teeth/tongue primarily undergo rigid translation inferred from lip vertices; under expression, lower teeth/tongue follow a rigid transform from lower-lip/chin vertices.
    Initial backfill procedure in Section 3.8 bootstraps internal bases before/during coregistration.
  • domain assumption Two-sphere eyeball with surface-of-revolution transfer from 2D polyline PCA is adequate for identity-varying cornea/limbus and shared bilateral parameters.
    Section 3.7.3; common simplified ocular model chosen for eyelid compatibility and glint-relevant cornea variation.
  • domain assumption Scan-to-mesh distance on a common painted facial region is a fair primary metric between topologically different models after non-rigid template mapping.
    Evaluation protocol Section 4.1; necessary for FLAME comparison but depends on correspondence quality.
  • standard math Eigen-decomposition and linear algebra operations behave in the usual real Euclidean setting.
    PCA/eig of covariance matrices throughout Sections 3.5–3.7.
invented entities (2)
  • GNM composite linear bases (part-stacked I and region-stacked E including teeth, tongue, pupil) independent evidence
    purpose: Provide a single controllable statistical manifold spanning outer head and internal oral/ocular structures.
    Not a new physical entity; a constructed parametric asset. Independent evidence is empirical fit quality and public model release, not an external physical discovery.
  • Dual-CVAE Semantic Sampler over GNM coefficients
    purpose: Map gender/ethnicity and 20 action expression labels into smooth β/φ samples without raw PCA walks.
    Engineering interface on top of GNM; validated qualitatively by sampled meshes, not by external human perceptual benchmarks in the text.

how reviews work

0 comments
Cite this review

Pith. "Pith review of GNM Head: A Generative aNthropometric Model of the human head." pith.science (2026). https://pith.science/paper/37HIBKJC

@misc{pith2026260723687,
  author       = {Pith},
  title        = {Pith review of: GNM Head: A Generative aNthropometric Model of the human head},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/37HIBKJC}},
  note         = {Machine review of arXiv:2607.23687}
}
read the original abstract

Parametric models of the human head are essential tools traditionally used in computer vision and graphics for animation, rendering, and reconstruction. More recently, they serve as crucial conditioning signals within generative large vision models, allowing for tight spatial control of generated imagery. However, existing publicly available models are typically limited in anatomical scope, modeling only outer geometry while ignoring intra-oral and ocular structures, and frequently suffer from reduced geometric quality stemming from low-fidelity input datasets. In this report we introduce a new parametric model dubbed Generative aNthropometric Model (GNM), named as a homophone of the human genome. GNM encompasses the head, face, neck, eyeballs, teeth, and tongue, and it is built on an extensive database of high-resolution 3D scans combined with high-quality anatomy specific artist-made samples. This report details the data provenance, the model architecture including the specialized sub-models for the ocular and intra-oral structures, and shows its SotA performance on fitting target 3D face scans. To foster community innovation, the complete GNM framework is made publicly available.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

173 extracted references · 5 canonical work pages

  1. [1]

    A. S. Abdelrehim, A. A. Farag, A. M. Shalaby, and M. T. El-Melegy. 2d-pca shape models: Application to 3d reconstruction of the human teeth from a single image. In International MICCAI Workshop on Medical Computer Vision, pages 44--52. Springer, 2013

    2013

  2. [2]

    Alexander, M

    O. Alexander, M. Rogers, W. Lambeth, J.-Y. Chiang, W.-C. Ma, C.-C. Wang, and P. Debevec. The digital emily project: Achieving a photorealistic digital actor. IEEE Computer Graphics and Applications, 30 0 (4): 0 20--31, 2010

    2010

  3. [3]

    Aneja, J

    S. Aneja, J. Thies, A. Dai, and M. Nie ner. Facetalk: Audio-driven motion diffusion for neural parametric head models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 21263--21273, 2024

    2024

  4. [4]

    Bednarik, E

    J. Bednarik, E. Wood, V. Choutas, T. Bolkart, D. Wang, C. Wu, and T. Beeler. Learning to stabilize faces. In Computer Graphics Forum, volume 43, 2024

    2024

  5. [5]

    Beeler and D

    T. Beeler and D. Bradley. Rigid stabilization of facial expressions. ACM Transactions on Graphics (TOG), 33 0 (4): 0 1--9, 2014

    2014

  6. [6]

    Beeler, B

    T. Beeler, B. Bickel, P. Beardsley, B. Sumner, and M. Gross. High-quality single-shot capture of facial geometry. In ACM Transactions on Graphics. Association for Computing Machinery (ACM), 2010

    2010

  7. [7]

    Beeler, F

    T. Beeler, F. Hahn, D. Bradley, B. Bickel, P. A. Beardsley, C. Gotsman, R. W. Sumner, and M. H. Gross. High-quality passive facial performance capture using anchor frames. ACM Trans. Graph., 30 0 (4): 0 75, 2011

    2011

  8. [8]

    B \'e rard, D

    P. B \'e rard, D. Bradley, M. Nitti, T. Beeler, and M. H. Gross. High-quality capture of eyes. ACM Trans. Graph., 33: 0 223--1, 2014

    2014

Show all 173 references
  1. [9]

    B \'e rard, D

    P. B \'e rard, D. Bradley, M. Gross, and T. Beeler. Lightweight eye capture using a parametric model. ACM Transactions on Graphics (TOG), 35 0 (4): 0 1--12, 2016

    2016

  2. [10]

    B \'e rard, D

    P. B \'e rard, D. Bradley, M. Gross, and T. Beeler. Practical person-specific eye rigging. In Computer Graphics Forum, volume 38, 2019

    2019

  3. [11]

    P. J. Besl and N. D. McKay. Method for registration of 3-d shapes. In Sensor fusion IV: control paradigms and data structures, volume 1611, pages 586--606. Spie, 1992

    1992

  4. [13]

    Booth, A

    J. Booth, A. Roussos, S. Zafeiriou, A. Ponniah, and D. Dunaway. A 3d morphable model learnt from 10,000 faces. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 5543--5552, 2016

    2016

  5. [14]

    M. C. Buehler, G. Li, E. Wood, L. Helminger, X. Chen, T. Shah, D. Wang, S. Garbin, S. Orts-Escolano, O. Hilliges, D. Lagun, J. Riviere, P. Gotardo, T. Beeler, A. Meka, and K. Sarkar. Cafca: High-quality novel view synthesis of expressive faces from casual few-shot captures. In...

    2024

  6. [15]

    C. Cao, Y. Weng, S. Zhou, Y. Tong, and K. Zhou. Facewarehouse: A 3d facial expression database for visual computing. IEEE Transactions on Visualization and Computer Graphics, 20 0 (3): 0 413--425, 2014. doi:10.1109/TVCG.2013.249

    2014 doi

  7. [16]

    Chandran and G

    P. Chandran and G. Zoss. Anatomically constrained implicit face models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2220--2229, 2024

    2024

  8. [17]

    Chandran, G

    P. Chandran, G. Zoss, M. Gross, P. Gotardo, and D. Bradley. Shape transformers: Topology-independent 3d shape models using transformers. In Computer Graphics Forum, volume 41, pages 195--207. Wiley Online Library, 2022

    2022

  9. [18]

    Chandran, G

    P. Chandran, G. Zoss, P. Gotardo, and D. Bradley. Continuous landmark detection with 3d queries. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 16858--16867, 2023

    2023

  10. [19]

    Chandran, G

    P. Chandran, G. Zoss, P. Gotardo, and D. Bradley. Infinite 3d landmarks: Improving continuous 2d facial landmark detection. In Computer Graphics Forum, volume 43, page e15126. Wiley Online Library, 2024

    2024

  11. [20]

    X. Chen, M. Mihajlovic, S. Wang, S. Prokudin, and S. Tang. Morphable diffusion: 3d-consistent diffusion for single-image avatar creation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 10359--10370, 2024

    2024

  12. [21]

    Chu and T

    X. Chu and T. Harada. Generalizable and animatable gaussian head avatar. Advances in Neural Information Processing Systems, 37: 0 57642--57670, 2024

    2024

  13. [22]

    Cudeiro, T

    D. Cudeiro, T. Bolkart, C. Laidlaw, A. Ranjan, and M. J. Black. Capture, learning, and synthesis of 3d speaking styles. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10101--10111, 2019

    2019

  14. [23]

    H. Dai, N. Pears, and W. Smith. A data-augmented 3d morphable model of the ear. In 2018 13th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2018), pages 404--408. IEEE, 2018

    2018

  15. [24]

    Dan e c ek, M

    R. Dan e c ek, M. J. Black, and T. Bolkart. Emoca: Emotion driven monocular face capture and animation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 20311--20322, 2022

    2022

  16. [25]

    Danecek, C

    R. Danecek, C. Schmitt, S. Polikovsky, and M. J. Black. Supervising 3d talking head avatars with analysis-by-audio-synthesis. In Thirteenth International Conference on 3D Vision, 2025

    2025

  17. [26]

    Edwards, C

    P. Edwards, C. Landreth, M. Pop awski, R. Malinowski, S. Watling, E. Fiume, and K. Singh. Jali-driven expressive facial animation and multilingual speech in cyberpunk 2077. In Special Interest Group on Computer Graphics and Interactive Techniques Conference Talks, pages 1--2, 2020

  18. [27]

    Egger, W

    B. Egger, W. Smith, A. Tewari, S. Wuhrer, M. Zollhöfer, T. Beeler, F. Bernard, T. Bolkart, A. Kortylewski, S. Romdhani, C. Theobalt, V. Blanz, and T. Vetter. 3D morphable face models--past, present and future. ACM Transactions on Graphics, 39 0 (5): 0 1--38, 2020. doi:10.1145/3395208

    2020 doi

  19. [28]

    Unreal Engine MetaHuman

    Epic Games . Unreal Engine MetaHuman . https://unrealengine.com, 2026

    2026

  20. [29]

    Y. Fan, Z. Lin, J. Saito, W. Wang, and T. Komura. Faceformer: Speech-driven 3d facial animation with transformers. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 18770--18780, 2022

    2022

  21. [30]

    Y. Feng, H. Feng, M. J. Black, and T. Bolkart. Learning an animatable detailed 3d face model from in-the-wild images. ACM Transactions on Graphics (ToG), 40 0 (4): 0 1--13, 2021

    2021

  22. [31]

    H. Fu, C. Li, X. Liu, J. Gao, A. Celikyilmaz, and L. Carin. Cyclical annealing schedule: A simple approach to mitigating kl vanishing. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technolog...

    2019

  23. [32]

    Giebenhain, T

    S. Giebenhain, T. Kirschstein, M. Georgopoulos, M. Rünz, L. Agapito, and M. Nießner. Learning neural parametric head models. In Computer Vision and Pattern Recognition, pages 21003--21012, 2023. doi:10.1109/CVPR52729.2023.02012

    2023

  24. [33]

    Giebenhain, T

    S. Giebenhain, T. Kirschstein, M. R \"u nz, L. Agapito, and M. Nie ner. Npga: Neural parametric gaussian avatars. In SIGGRAPH Asia 2024 Conference Papers, pages 1--11, 2024

    2024

  25. [34]

    E. D. Guestrin and M. Eizenman. General theory of remote gaze estimation using the pupil center and corneal reflections. IEEE Transactions on Biomedical Engineering, 53 0 (6): 0 1124--1133, 2006. doi:10.1109/TBME.2005.863952

    2006

  26. [35]

    K. Guo, P. Lincoln, P. Davidson, J. Busch, X. Yu, M. Whalen, G. Harvey, S. Orts-Escolano, R. Pandey, J. Dourgarian, T. Danhang, A. Tkach, A. Kowdle, E. Cooper, M. Dou, S. Fanello, G. Fyffe, C. Rhemann, J. Taylor, P. Debevec, and S. Izadi. The relightables: Volumetric performan...

    2019

  27. [36]

    G. Harling. General data protection regulation (gdpr). Official Journal of the European Union, 2018

    2018

  28. [37]

    Hewitt, F

    C. Hewitt, F. Saleh, S. Aliakbarian, L. Petikam, S. Rezaeifar, L. Florentin, Z. Hosenie, T. J. Cashman, J. Valentin, D. Cosker, and T. Baltru s aitis. Look ma, no markers: holistic performance capture without the hassle. ACM Transactions on Graphics (TOG), 43 0 (6), 2024

    2024

  29. [38]

    D. A. Hirshberg, M. Loper, E. Rachlin, and M. J. Black. Coregistration: Simultaneous alignment and modeling of articulated 3D shape. In Computer Vision - ECCV 2012 - 12th European Conference on Computer Vision, Florence, Italy, October 7-13, 2012, Proceedings, Part VI , volume...

    2012

  30. [39]

    Jakob, S

    W. Jakob, S. Speierer, N. Roussel, M. Nimier-David, D. Vicini, T. Zeltner, B. Nicolet, M. Crespo, V. Leroy, and Z. Zhang. Mitsuba 3 renderer. https://mitsuba-renderer.org, 2022. Version 3.8.0

    2022

  31. [41]

    D. P. Kingma and J. Ba. Adam: A method for stochastic optimization. In International Conference on Learning Representations (ICLR), 2015. URL https://arxiv.org/abs/1412.6980

    2015 arXiv

  32. [43]

    G. Li, A. Meka, F. Mueller, M. C. Buehler, O. Hilliges, and T. Beeler. Eyenerf: a hybrid representation for photorealistic synthesis, animation and relighting of human eyes. ACM Transactions on Graphics (ToG), 41 0 (4): 0 1--16, 2022

    2022

  33. [44]

    T. Li, T. Bolkart, M. J. Black, H. Li, and J. Romero. Learning a model of facial shape and expression from 4d scans. ACM Transactions on Graphics, 36 0 (6): 0 1--17, 2017. doi:10.1145/3130800.3130813

    2017

  34. [45]

    T.-M. Li, M. Aittala, F. Durand, and J. Lehtinen. Differentiable monte carlo ray tracing through edge sampling. ACM Trans. Graph. (Proc. SIGGRAPH Asia), 37 0 (6): 0 222:1--222:11, 2018

    2018

  35. [46]

    Luo and E

    B. Luo and E. R. Hancock. Iterative procrustes alignment with the em algorithm. Image and Vision Computing, 20 0 (5-6): 0 377--396, 2002

    2002

  36. [48]

    Mildenhall, P

    B. Mildenhall, P. Srinivasan, M. Tancik, J. T. Barron, R. Ramamoorthi, and R. Ng. Nerf: Representing scenes as neural radiance fields for view synthesis. Commun. ACM, 65 0 (1): 0 405--421, 2021. doi:10.1145/3503250

    2021 doi

  37. [49]

    Nicolet, A

    B. Nicolet, A. Jacobson, and W. Jakob. Large steps in inverse rendering of geometry. ACM Transactions on Graphics (Proceedings of SIGGRAPH Asia), 40 0 (6), Dec. 2021. doi:10.1145/3478513.3480501

    2021

  38. [50]

    O'Sullivan, L

    E. O'Sullivan, L. S. van de Lande, A.-J. C. Oosting, A. Papaioannou, N. O. Jeelani, M. J. Koudstaal, R. H. Khonsari, D. J. Dunaway, S. Zafeiriou, and S. Schievano. The 3d skull 0--4 years: a validated, generative, statistical shape model. Bone reports, 15: 0 101154, 2021

    2021

  39. [51]

    O'Sullivan, L

    E. O'Sullivan, L. S. van de Lande, K. El Ghoul, M. J. Koudstaal, S. Schievano, R. H. Khonsari, D. J. Dunaway, and S. Zafeiriou. Growth patterns and shape development of the paediatric mandible--a 3d statistical model. Bone reports, 16: 0 101528, 2022

    2022

  40. [52]

    Paysan, R

    P. Paysan, R. Knothe, B. Amberg, S. Romdhani, and T. Vetter. A 3d face model for pose and illumination invariant face recognition. In 2009 sixth IEEE international conference on advanced video and signal based surveillance, pages 296--301. IEEE, IEEE, 2009. doi:10.1109/AVSS.2009.58

    2009 doi

  41. [53]

    C. Peng, T. Xu, D. Liu, N. Wang, and X. Gao. Within 3dmm space: Exploring inherent 3d artifact for video forgery detection. IEEE Transactions on Information Forensics and Security, 2025

    2025

  42. [54]

    C. Peng, J. Sun, Y. Chen, Z. Su, Z. Su, and Y. Liu. Parametric gaussian human model: Generalizable prior for efficient and realistic human avatar modeling. In 2026 International Conference on 3D Vision (3DV), pages 771--782. IEEE, 2026

    2026

  43. [55]

    Petmezas, V

    G. Petmezas, V. Vanian, K. Konstantoudakis, E. E. Almaloglou, and D. Zarpalas. Video deepfake detection using a hybrid cnn-lstm-transformer model for identity verification. Multimedia Tools and Applications, 84 0 (33): 0 40617--40636, 2025

    2025

  44. [56]

    Ploumpis, E

    S. Ploumpis, E. Ververas, E. O. Sullivan, S. Moschoglou, H. Wang, N. E. Pears, W. Smith, B. Gecer, and S. Zafeiriou. Towards a complete 3d morphable model of the human head. IEEE transactions on pattern analysis and machine intelligence, 43 0 (11): 0 4142--4160, 2019 a . doi:1...

    2019

  45. [60]

    Prinzler, E

    M. Prinzler, E. Zakharov, V. Sklyarova, B. Kabadayi, and J. Thies. Joker: Conditional 3d head synthesis with extreme facial expressions. In International Conference on 3D Vision, 2025. doi:10.1109/3DV66043.2025.00148

    2025

  46. [61]

    S. Qian, T. Kirschstein, L. Schoneveld, D. Davoli, S. Giebenhain, and M. Nie ner. Gaussianavatars: Photorealistic head avatars with rigged 3d gaussians. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 20299--20309, 2024

    2024

  47. [62]

    D. Qiu, Y. Zhang, T. Beeler, V. Tankovich, C. Hane, S. Fanello, C. Rhemann, and S. Escolano. Chosen: Contrastive hypothesis selection for multi-view depth refinement. Proceedings of the Conference on Robots and Vision, 2025. doi:10.48550/arXiv.2404.02225

    2025 doi arXiv

  48. [63]

    Rombach, A

    R. Rombach, A. Blattmann, D. Lorenz, P. Esser, and B. Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684--10695, 2022

    2022

  49. [64]

    K. Sohn, H. Lee, and X. Yan. Learning structured output representation using deep conditional generative models. Advances in neural information processing systems, 28, 2015

    2015

  50. [65]

    Sorkine, D

    O. Sorkine, D. Cohen - Or, Y. Lipman, M. Alexa, C. R \" o ssl, and H. Seidel. Laplacian surface editing. In J. Boissonnat and P. Alliez, editors, Second Eurographics Symposium on Geometry Processing, Nice, France, July 8-10, 2004, volume 71 of ACM International Conference Proc...

    2004 doi

  51. [66]

    S. G. Srinivasan, Q. Wang, J. Rojas, G. Kl \'a r, L. Kavan, and E. Sifakis. Learning active quasistatic physics-based models from data. ACM Transactions on Graphics (ToG), 40 0 (4): 0 1--14, 2021

    2021

  52. [67]

    Z. Sun, T. Lv, S. Ye, M. Lin, J. Sheng, Y.-H. Wen, M. Yu, and Y.-J. Liu. Diffposetalk: Speech-driven stylistic 3d facial animation and head pose generation via diffusion models. ACM TOG, 2024

    2024

  53. [68]

    G. Taubin. A signal processing approach to fair surface design. In SIGGRAPH, pages 351--358. ACM , 1995

    1995

  54. [69]

    Varol, J

    G. Varol, J. Romero, X. Martin, N. Mahmood, M. J. Black, I. Laptev, and C. Schmid. Learning from synthetic humans. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 109--117, 2017

    2017

  55. [70]

    E. Wood, T. Baltrušaitis, L.-p. Morency, P. Robinson, and A. Bulling. A 3d morphable eye region model for gaze estimation. In European conference on computer vision, pages 297--313. Springer, Springer International Publishing, 2016. doi:10.1007/978-3-319-46448-0_18

    2016 doi

  56. [72]

    E. Wood, T. Baltrušaitis, C. Hewitt, M. Johnson, J. Shen, N. Milosavljević, D. Wilde, S. J. Garbin, T. Sharp, I. Stojiljković, T. Cashman, and V. Julien. 3d face reconstruction with dense landmarks. In European Conference on Computer Vision, pages 160--177. Springer, Springer ...

    2022 doi arXiv

  57. [73]

    C. Wu, D. Bradley, P. Garrido, M. Zollhöfer, C. Theobalt, M. Gross, and T. Beeler. Model-based teeth reconstruction. ACM Transactions on Graphics, 35 0 (6): 0 1--13, 2016. doi:10.1145/2980179.2980233

    2016

  58. [74]

    C. Wu, T. Shiratori, and Y. Sheikh. Deep incremental learning for efficient high-fidelity face tracking. ACM TOG, 2018

    2018

  59. [76]

    Y. Xu, Z. Su, Q. Wu, and Y. Liu. Gphm: Gaussian parametric head model for monocular head avatar reconstruction. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025. doi:10.1109/TPAMI.2025.3596331

    2025

  60. [78]

    L. Yang, B. Kim, G. Zoss, B. G \"o zc \"u , M. Gross, and B. Solenthaler. Implicit neural representation for physics-driven actuated soft bodies. ACM Transactions on Graphics (ToG), 41 0 (4): 0 1--10, 2022

    2022

  61. [79]

    L. Yang, G. Zoss, P. Chandran, P. Gotardo, M. Gross, B. Solenthaler, E. Sifakis, and D. Bradley. An implicit physical face model driven by expression and style. In SIGGRAPH Asia 2023 conference papers, pages 1--12, 2023

    2023

  62. [80]

    L. Yang, G. Zoss, P. Chandran, M. Gross, B. Solenthaler, E. Sifakis, and D. Bradley. Learning a generalized physical face model from data. ACM Transactions on Graphics (TOG), 43 0 (4): 0 1--14, 2024

    2024

  63. [81]

    W. Yang, N. Marshak, D. S \`y kora, S. Ramalingam, and L. Kavan. Building anatomically realistic jaw kinematics model from data. The Visual Computer, 35 0 (6): 0 1105--1118, 2019

    2019

  64. [82]

    Zhang, M

    C. Zhang, M. Elgharib, G. Fox, M. Gu, C. Theobalt, and W. Wang. An implicit parametric morphable dental model. ACM Transactions on Graphics, 41 0 (6): 0 1--13, 2022. doi:10.1145/3550454.3555469

    2022

  65. [83]

    Zhang, M

    H. Zhang, M. Cisse, Y. N. Dauphin, and D. Lopez-Paz. mixup: Beyond empirical risk minimization. In International Conference on Learning Representations, 2018

    2018

  66. [85]

    Zhou and S

    Y. Zhou and S. Zafeiriou. Deformable models of ears in-the-wild for alignment and recognition. In IEEE International Conference on Automatic Face and Gesture Recognition, pages 626--633. IEEE, IEEE, 2017. doi:10.1109/FG.2017.79

    2017 doi

  67. [87]

    G. Zoss, D. Bradley, P. Bérard, and T. Beeler. An empirical rig for jaw animation. In ACM Transactions on Graphics, pages 1--12. Association for Computing Machinery (ACM), 2018

    2018

  68. [88]

    G. Zoss, T. Beeler, M. Gross, and D. Bradley. Accurate markerless jaw tracking for facial performance capture. ACM Transactions on Graphics (TOG), 38 0 (4): 0 1--8, 2019

    2019

  69. [89]

    International Conference on Computer Graphics and Interactive Techniques , pages =

    A morphable model for the synthesis of 3D faces , author =. International Conference on Computer Graphics and Interactive Techniques , pages =. 1999 , journal =. doi:10.1145/3596711.3596730 , publisher =

    1999

  70. [90]

    2009 Sixth IEEE International Conference on Advanced Video and Signal Based Surveillance , doi =

    A 3D face model for pose and illumination invariant face recognition , author =. 2009 Sixth IEEE International Conference on Advanced Video and Signal Based Surveillance , doi =. 2009 , organization =

    2009

  71. [91]

    IEEE Transactions on Visualization and Computer Graphics , volume =

    FaceWarehouse: A 3D facial expression database for visual computing , author =. IEEE Transactions on Visualization and Computer Graphics , volume =. 2014 , publisher =

    2014

  72. [92]

    ACM Transactions on Graphics , volume =

    Learning a model of facial shape and expression from 4D scans , author =. ACM Transactions on Graphics , volume =. 2017 , doi =

    2017

  73. [93]

    ACM Transactions on Graphics (ToG) , volume=

    Learning an animatable detailed 3D face model from in-the-wild images , author=. ACM Transactions on Graphics (ToG) , volume=. 2021 , publisher=

    2021

  74. [94]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Emoca: Emotion driven monocular face capture and animation , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  75. [95]

    Computer Vision and Pattern Recognition , pages =

    Learning neural parametric head models , author =. Computer Vision and Pattern Recognition , pages =. 2023 , journal =

    2023

  76. [96]

    IEEE International Conference on Computer Vision , pages =

    ImHead: A Large-scale Implicit Morphable Model for Localized Head Modeling , author =. IEEE International Conference on Computer Vision , pages =. 2025 , journal =. doi:10.1109/ICCV51701.2025.00950 , publisher =

    2025

  77. [97]

    arXiv preprint arXiv:2503.11792 , year=

    StyleMorpheus: A Style-Based 3D-Aware Morphable Face Model , author=. arXiv preprint arXiv:2503.11792 , year=

    arXiv

  78. [98]

    Computer Vision and Pattern Recognition , pages =

    Gaussian Head Avatar: Ultra High-Fidelity Head Avatar via Dynamic Gaussians , author =. Computer Vision and Pattern Recognition , pages =. 2024 , journal =. doi:10.1109/CVPR52733.2024.00189 , publisher =

    2024

  79. [99]

    Advances in Neural Information Processing Systems , volume=

    Generalizable and animatable gaussian head avatar , author=. Advances in Neural Information Processing Systems , volume=

  80. [100]

    IEEE transactions on pattern analysis and machine intelligence , volume =

    Towards a complete 3D morphable model of the human head , author =. IEEE transactions on pattern analysis and machine intelligence , volume =. 2019 , publisher =

    2019

  81. [101]

    Computer Vision and Pattern Recognition , pages =

    Combining 3d morphable models: A large scale face-and-head model , author =. Computer Vision and Pattern Recognition , pages =. 2019 , journal =. doi:10.1109/CVPR.2019.01119 , publisher =

    2019

  82. [102]

    Nerf: Representing scenes as neural radiance fields for view synthesis , author =. Commun. ACM , volume =. 2021 , publisher =

    2021

  83. [103]

    IEEE International Conference on Computer Vision , pages =

    Adding conditional control to text-to-image diffusion models , author =. IEEE International Conference on Computer Vision , pages =. 2023 , journal =. doi:10.1109/ICCV51070.2023.00355 , publisher =

    2023

  84. [104]

    Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

    Learning from synthetic humans , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

  85. [105]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    High-resolution image synthesis with latent diffusion models , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  86. [106]

    Advances in Neural Information Processing Systems , volume=

    Stable-pose: Leveraging transformers for pose-guided text-to-image generation , author=. Advances in Neural Information Processing Systems , volume=

  87. [107]

    IEEE International Conference on Computer Vision , pages =

    Fake it till you make it: face analysis in the wild using synthetic data alone , author =. IEEE International Conference on Computer Vision , pages =. 2021 , journal =. doi:10.1109/ICCV48922.2021.00366 , publisher =

    2021

  88. [108]

    Official Journal of the European Union , year =

    General Data Protection Regulation (GDPR) , author =. Official Journal of the European Union , year =

  89. [109]

    ACM Transactions on Graphics , volume =

    3D Gaussian Splatting for Real-Time Radiance Field Rendering , author =. ACM Transactions on Graphics , volume =. 2023 , pages =. doi:10.1145/3592433 , publisher =

    2023 doi

  90. [110]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Morphable diffusion: 3d-consistent diffusion for single-image avatar creation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  91. [111]

    2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) , pages =

    Refa: real-time egocentric facial animations for virtual reality , author =. 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) , pages =. 2024 , journal =. doi:10.1109/CVPRW63382.2024.00482 , publisher =

    2024

  92. [112]

    in-the-wild

    3D human tongue reconstruction from single" in-the-wild" images , author =. Computer Vision and Pattern Recognition , pages =. 2022 , journal =

    2022

  93. [113]

    Computer Vision and Pattern Recognition , pages =

    Speech driven tongue animation , author =. Computer Vision and Pattern Recognition , pages =. 2022 , journal =. doi:10.1109/CVPR52688.2022.01976 , publisher =

    2022

  94. [114]

    European Conference on Computer Vision , doi =

    3d face reconstruction with dense landmarks , author =. European Conference on Computer Vision , doi =. 2022 , organization =

    2022

  95. [115]

    Computer Vision and Pattern Recognition , pages =

    Synthetic prior for few-shot drivable head avatar inversion , author =. Computer Vision and Pattern Recognition , pages =. 2025 , journal =. doi:10.1109/CVPR52734.2025.01003 , publisher =

    2025

  96. [116]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , year =

    Gphm: Gaussian parametric head model for monocular head avatar reconstruction , author =. IEEE Transactions on Pattern Analysis and Machine Intelligence , year =

  97. [117]

    Eurographics Symposium on Geometry Processing , publisher =

    Sparse iterative closest point , author =. Eurographics Symposium on Geometry Processing , publisher =

  98. [118]

    Proceedings Shape Modeling Applications, 2004

    Least-squares meshes , author =. Proceedings Shape Modeling Applications, 2004. , doi =. 2004 , organization =

    2004

  99. [119]

    ACM Transactions on Graphics , publisher =

    High-quality single-shot capture of facial geometry , author =. ACM Transactions on Graphics , publisher =

  100. [120]

    Computer Graphics Forum , volume=

    Shape Transformers: Topology-Independent 3D Shape Models Using Transformers , author=. Computer Graphics Forum , volume=. 2022 , organization=

    2022

  101. [121]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Anatomically constrained implicit face models , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  102. [122]

    Proceedings of the Conference on Robots and Vision , year =

    CHOSEN: Contrastive Hypothesis Selection for Multi-View Depth Refinement , author =. Proceedings of the Conference on Robots and Vision , year =

  103. [123]

    ACM Transactions on Graphics (ToG) , volume =

    The relightables: Volumetric performance capture of humans with realistic relighting , author =. ACM Transactions on Graphics (ToG) , volume =. 2019 , publisher =

    2019

  104. [124]

    Computer Graphics Forum , year =

    Yoshiyasu, Yusuke and Ma, Wan‐Chun and Yoshida, Eiichi and Kanehiro, Fumio , title =. Computer Graphics Forum , year =. doi:10.1111/cgf.12451 , publisher =

    doi

  105. [125]

    European Conference on Computer Vision , doi =

    Raft: Recurrent all-pairs field transforms for optical flow , author =. European Conference on Computer Vision , doi =. 2020 , organization =

    2020

  106. [126]

    Proceedings of the 4th international conference on Computer graphics and interactive techniques in Australasia and Southeast Asia , pages =

    Laplacian mesh optimization , author =. Proceedings of the 4th international conference on Computer graphics and interactive techniques in Australasia and Southeast Asia , pages =. 2006 , journal =. doi:10.1145/1174429.1174494 , publisher =

    2006

  107. [127]

    and Smith, W

    Egger, B. and Smith, W. and Tewari, A. and Wuhrer, Stefanie and Zollhöfer, M. and Beeler, T. and Bernard, Florian and Bolkart, Timo and Kortylewski, A. and Romdhani, S. and Theobalt, C. and Blanz, V. and Vetter, T. , journal =. 2020 , doi =

    2020

  108. [128]

    ACM Transactions on Graphics (ToG) , volume=

    Learning active quasistatic physics-based models from data , author=. ACM Transactions on Graphics (ToG) , volume=. 2021 , publisher=

    2021

  109. [129]

    ACM Transactions on Graphics (ToG) , volume=

    Implicit neural representation for physics-driven actuated soft bodies , author=. ACM Transactions on Graphics (ToG) , volume=. 2022 , publisher=

    2022

  110. [130]

    SIGGRAPH Asia 2023 conference papers , pages=

    An implicit physical face model driven by expression and style , author=. SIGGRAPH Asia 2023 conference papers , pages=

    2023

  111. [131]

    ACM Transactions on Graphics (TOG) , volume=

    Learning a Generalized Physical Face Model From Data , author=. ACM Transactions on Graphics (TOG) , volume=. 2024 , publisher=

    2024

  112. [132]

    , author=

    High-quality capture of eyes. , author=. ACM Trans. Graph. , volume=

  113. [133]

    ACM Transactions on Graphics (TOG) , volume=

    Lightweight eye capture using a parametric model , author=. ACM Transactions on Graphics (TOG) , volume=. 2016 , publisher=

    2016

  114. [134]

    , author=

    Practical Person-Specific Eye Rigging. , author=. Computer Graphics Forum , volume=

  115. [135]

    ACM Transactions on Graphics (ToG) , volume=

    Eyenerf: a hybrid representation for photorealistic synthesis, animation and relighting of human eyes , author=. ACM Transactions on Graphics (ToG) , volume=. 2022 , publisher=

    2022

  116. [136]

    ACM transactions on graphics (TOG) , volume=

    An anatomically-constrained local deformation model for monocular face capture , author=. ACM transactions on graphics (TOG) , volume=. 2016 , publisher=

    2016

  117. [137]

    ACM Transactions on Graphics , pages =

    An empirical rig for jaw animation , author =. ACM Transactions on Graphics , pages =. 2018 , journal =

    2018

  118. [138]

    ACM Transactions on Graphics (TOG) , volume=

    Accurate markerless jaw tracking for facial performance capture , author=. ACM Transactions on Graphics (TOG) , volume=. 2019 , publisher=

    2019

  119. [139]

    The Visual Computer , volume=

    Building anatomically realistic jaw kinematics model from data , author=. The Visual Computer , volume=. 2019 , publisher=

    2019

  120. [140]

    2018 13th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2018) , pages=

    A data-augmented 3D morphable model of the ear , author=. 2018 13th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2018) , pages=. 2018 , organization=

    2018

  121. [141]

    European Conference on Computer Vision , doi =

    A 3d morphable eye region model for gaze estimation , author =. European Conference on Computer Vision , doi =. 2016 , organization =

    2016

  122. [142]

    IEEE International Conference on Automatic Face and Gesture Recognition , doi =

    Deformable models of ears in-the-wild for alignment and recognition , author =. IEEE International Conference on Automatic Face and Gesture Recognition , doi =. 2017 , organization =

    2017

  123. [143]

    ACM Transactions on Graphics , volume =

    An implicit parametric morphable dental model , author =. ACM Transactions on Graphics , volume =. 2022 , publisher =

    2022

  124. [144]

    International MICCAI Workshop on Medical Computer Vision , pages=

    2D-PCA shape models: Application to 3D reconstruction of the human teeth from a single image , author=. International MICCAI Workshop on Medical Computer Vision , pages=. 2013 , organization=

    2013

  125. [145]

    ACM Transactions on Graphics , volume =

    Model-based teeth reconstruction , author =. ACM Transactions on Graphics , volume =. 2016 , doi =

    2016

  126. [146]

    International Conference on 3D Vision , doi =

    Joker: Conditional 3D Head Synthesis With Extreme Facial Expressions , author =. International Conference on 3D Vision , doi =

  127. [147]

    IEEE Transactions on Biomedical Engineering , volume =

    General theory of remote gaze estimation using the pupil center and corneal reflections , author =. IEEE Transactions on Biomedical Engineering , volume =. 2006 , publisher =

    2006

  128. [148]

    , author=

    Learning to Stabilize Faces. , author=. Computer Graphics Forum , volume=

  129. [149]

    Deep Incremental Learning for Efficient High-Fidelity Face Tracking , author =

  130. [150]

    ACM Transactions on Graphics (TOG) , volume=

    Rigid stabilization of facial expressions , author=. ACM Transactions on Graphics (TOG) , volume=. 2014 , publisher=

    2014

  131. [151]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Gaussianavatars: Photorealistic head avatars with rigged 3d gaussians , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  132. [152]

    Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

    Gaussian eigen models for human heads , author=. Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

  133. [153]

    2026 International Conference on 3D Vision (3DV) , pages=

    Parametric gaussian human model: Generalizable prior for efficient and realistic human avatar modeling , author=. 2026 International Conference on 3D Vision (3DV) , pages=. 2026 , organization=

    2026

  134. [154]

    SIGGRAPH Asia 2024 Conference Papers , pages=

    Npga: Neural parametric gaussian avatars , author=. SIGGRAPH Asia 2024 Conference Papers , pages=

    2024

  135. [155]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Capture, learning, and synthesis of 3D speaking styles , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  136. [156]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Faceformer: Speech-driven 3d facial animation with transformers , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  137. [157]

    IEEE Transactions on Information Forensics and Security , year=

    Within 3DMM space: Exploring inherent 3D artifact for video forgery detection , author=. IEEE Transactions on Information Forensics and Security , year=

  138. [158]

    Multimedia Tools and Applications , volume=

    Video deepfake detection using a hybrid CNN-LSTM-Transformer model for identity verification , author=. Multimedia Tools and Applications , volume=. 2025 , publisher=

    2025

  139. [159]

    Bone reports , volume=

    Growth patterns and shape development of the paediatric mandible--A 3D statistical model , author=. Bone reports , volume=. 2022 , publisher=

    2022

  140. [160]

    Bone reports , volume=

    The 3D skull 0--4 years: a validated, generative, statistical shape model , author=. Bone reports , volume=. 2021 , publisher=

    2021

  141. [161]

    IEEE Computer Graphics and Applications , volume=

    The digital emily project: Achieving a photorealistic digital actor , author=. IEEE Computer Graphics and Applications , volume=. 2010 , publisher=

    2010

  142. [162]

    , author=

    High-quality passive facial performance capture using anchor frames. , author=. ACM Trans. Graph. , volume=

  143. [163]

    Image and Vision Computing , volume=

    Iterative procrustes alignment with the em algorithm , author=. Image and Vision Computing , volume=. 2002 , publisher=

    2002

  144. [164]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Facetalk: Audio-driven motion diffusion for neural parametric head models , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  145. [165]

    2024 , publisher=

    DiffPoseTalk: Speech-Driven Stylistic 3D Facial Animation and Head Pose Generation via Diffusion Models , author=. 2024 , publisher=

    2024

  146. [166]

    Thirteenth International Conference on 3D Vision , year=

    Supervising 3D Talking Head Avatars with Analysis-by-Audio-Synthesis , author=. Thirteenth International Conference on 3D Vision , year=

  147. [167]

    Special Interest Group on Computer Graphics and Interactive Techniques Conference Talks , pages=

    Jali-driven expressive facial animation and multilingual speech in cyberpunk 2077 , author=. Special Interest Group on Computer Graphics and Interactive Techniques Conference Talks , pages=

  148. [168]

    2026 , howpublished =

    2026

  149. [169]

    2024 , booktitle=SIGGRAPH, publisher =

    Cafca: High-quality Novel View Synthesis of Expressive Faces from Casual Few-shot Captures , author=. 2024 , booktitle=SIGGRAPH, publisher =

    2024

  150. [170]

    Advances in neural information processing systems , volume=

    Learning structured output representation using deep conditional generative models , author=. Advances in neural information processing systems , volume=

  151. [171]

    arXiv preprint arXiv:1312.6114 , year=

    Auto-encoding variational bayes , author=. arXiv preprint arXiv:1312.6114 , year=

    arXiv

  152. [172]

    International Conference on Learning Representations , year=

    mixup: Beyond empirical risk minimization , author=. International Conference on Learning Representations , year=

  153. [173]

    Cyclical annealing schedule: A simple approach to mitigating kl vanishing , author=. Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) , pages=

    2019

  154. [174]

    Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

    A 3d morphable model learnt from 10,000 faces , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

  155. [175]

    Laplacian Surface Editing , booktitle =

    Olga Sorkine and Daniel Cohen. Laplacian Surface Editing , booktitle =. 2004 , doi =

    2004

  156. [176]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Continuous landmark detection with 3d queries , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  157. [177]

    Computer Graphics Forum , volume=

    Infinite 3d landmarks: Improving continuous 2d facial landmark detection , author=. Computer Graphics Forum , volume=. 2024 , organization=

    2024

  158. [178]

    ACM Transactions on Graphics (TOG) , volume=

    Look Ma, no markers: holistic performance capture without the hassle , author=. ACM Transactions on Graphics (TOG) , volume=. 2024 , publisher=

    2024

  159. [179]

    and Ba, Jimmy , title =

    Kingma, Diederik P. and Ba, Jimmy , title =. International Conference on Learning Representations (ICLR) , year =

  160. [180]

    Sensor fusion IV: control paradigms and data structures , volume=

    Method for registration of 3-D shapes , author=. Sensor fusion IV: control paradigms and data structures , volume=. 1992 , organization=

    1992

  161. [181]

    Hirshberg and Matthew Loper and Eric Rachlin and Michael J

    David A. Hirshberg and Matthew Loper and Eric Rachlin and Michael J. Black , title =. Computer Vision -

  162. [182]

    Mitsuba 3 renderer , author =

  163. [183]

    ACM Transactions on Graphics (Proceedings of SIGGRAPH Asia) , volume =

    Baptiste Nicolet and Alec Jacobson and Wenzel Jakob , title =. ACM Transactions on Graphics (Proceedings of SIGGRAPH Asia) , volume =. 2021 , month = dec, doi =

    2021

  164. [184]

    ACM Trans

    Differentiable Monte Carlo Ray Tracing through Edge Sampling , author =. ACM Trans. Graph. (Proc. SIGGRAPH Asia) , volume =

  165. [185]

    SIGGRAPH , pages =

    Gabriel Taubin , title =. SIGGRAPH , pages =

Pith tools

Reviewed July 30, 2026 · model on record in the stance chip above.