Pith. sign in

REVIEW 3 major objections 5 minor 48 references

PaNDaS: Learnable Deformation Modeling with Localized Control

T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read PaNDaS represents a mesh deformation as a per-triangle feature field with a maskable global pose code, so selected regions can be deformed, poses mixed, and localized statistics computed with no optimization at inference.

desk verdict A genuinely useful architecture for local mesh deformation control, but the flagship partial-deformation claim is plausible rather than proven: the masking identity is never enforced and never measured. read the letter →

arxiv 2412.02306 v2 pith:L2VINPUG submitted 2024-12-03 cs.CV

Thomas Besnier , Emery Pierson , Sylvain Arguillere , Maks Ovsjanikov , Mohamed Daoudi This is my paper · ORCID
classification cs.CV
keywords non-rigidshapedeformationpartial3DmeshinterpolationJacobianfieldslocalizedcontrolposemixingstatisticslearning-based
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

PaNDaS (Partial Non-rigid Deformations and interpolations of Surfaces) is a learning-based model for non-rigid deformation of triangular surface meshes that is designed to deform only selected parts of a shape. Its central idea is to represent a deformation not by one global latent vector but by a per-triangle feature field obtained by concatenating local features of the source mesh with a global code that encodes the difference between target and source poses. Because the global code is replicated across triangles, the user can multiply it by a binary mask and thereby restrict the deformation to a chosen region, interpolate between poses, mix several pose codes on different regions, and compute localized shape statistics. The paper reports that this field-based representation outperforms global-latent baselines in reconstruction accuracy on hand and body datasets, is competitive on face expressions, and produces smoother interpolations for strongly nonlinear motions, all without any optimization at inference time.

What carries the argument

The load-bearing object is a feature field over the source mesh whose per-triangle vectors combine a local geometric descriptor with a global deformation code; the field is what makes masking and mixing possible. The global code is produced by a deformation encoder that projects per-triangle target features onto the first $s=4$ eigenvectors of the cotangent Laplacian, $p_k = \frac{1}{\mathrm{Area}(Y)}\sum_t \mathrm{Area}(t)\, g_t e^k_t$, then passes $(p_1,p_2,p_3,p_4)$ through a linear layer and takes the difference $z = z_Y - z_X$, so identical source and target give $z=0$. The deformation generator is a neural Jacobian field: it predicts per-triangle $3\times 3$ Jacobians from the feature field and recovers a consistent displacement field by solving the Poisson equation $\nabla_X v = \nabla^T M J$. This pipeline lets the user manipulate the feature field directly (mask, interpolate, sum codes) while the Poisson solve couples the result into a smooth mesh.

What would settle it

Take a trained PaNDaS model, pick a non-neutral source mesh $X$ (for example a hand with bent fingers), set the global code to $z=0$, and decode; if the output differs from $X$ by more than the model's usual reconstruction error, then the zero-code identity that masking relies on fails for non-neutral sources. A complementary check is to apply a mask to a small patch on such a source and measure the displacement of vertices far outside the mask: large motion there would contradict the claimed locality.

Watch

Extended reading notes

Core claim

The paper's central claim is that localized control of non-rigid deformation can be achieved by making the deformation generator depend on a per-face feature field $\tilde f_t = (f_t, z)\in\mathbb{R}^{l+r}$, where $f_t$ is a local geometric feature on triangle $t$ of the source mesh and $z = \mathrm{Enc}_D(Y) - \mathrm{Enc}_D(X)$ is a global deformation code. The generator $G$ maps this field to a per-triangle Jacobian field $(J_t)_t$, which is integrated into vertex displacements through a Poisson solve, $\nabla_X v = \nabla^T M J$. Masking the code, $\tilde f_t^{\mathrm{partial}} = (f_t, M_t\odot z)$, yields deformations that the paper shows are concentrated near the masked region, with the displacement norm decaying rapidly away from the mask boundary. From this construction the paper derives partial interpolation, pose mixing by summing masked codes, partial motion transfer, and a statistical model of deformations in which Euclidean operations on feature fields correspond to natural shape operations. The paper supports the claim with experiments on MANO hand meshes, DFAUST bodies, and COMA faces, reporting state-of-the-art reconstruction error on MANO and DFAUST and improved interpolation on strongly nonlinear sequences.

Load-bearing premise

The masking trick works only if setting the global deformation code to zero makes the generator output the source mesh for any source shape, but this identity is not enforced by a loss term and is only implicitly learned from training pairs in which the source is a neutral pose.

Editorial extensions

If this is right

  • A user can deform only the triangles covered by a binary mask, and the resulting displacement decays quickly away from the mask boundary, so partial edits of a mesh need no skeleton, texture, or per-query optimization.
  • New poses can be generated by summing several masked pose codes on different regions, giving a mix-and-match strategy for combining, say, one pose's arm motion with another pose's leg motion.
  • Partial motion transfer is available: because the encoder is trained to encode pose, one can restrict the transferred motion to a chosen part of the target identity.
  • Euclidean statistics on feature fields translate back into shape statistics, so one can compute mean shapes and principal deformation components that respect locality and work on unregistered scans.
  • On the reported benchmarks, the method beats global-latent deformation generators in reconstruction MSE on MANO and DFAUST and stays competitive on COMA; interpolation errors on nonlinear DFAUST motions are the lowest among compared methods.

Reading between the lines

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

  • Because the zero-code identity $G((f_t,0))=X$ is only learned implicitly from neutral source meshes, the method's partial-deformation guarantees are likely to degrade when the source mesh is already far from neutral; a direct test would be to decode $z=0$ from a strongly posed source and measure the residual displacement.
  • The hard binary mask creates boundary artifacts, as the paper itself notes; a weighted or learned mask would likely smooth the transition region and could be trained with a small auxiliary loss on the deformation outside the mask.
  • The global code discards all Laplacian frequency components above $s=4$, so very fine high-frequency pose details are filtered out; keeping per-part or per-frequency codes could improve fidelity for detailed expressions while preserving locality.
  • Since the final displacements come from a Poisson solve, the deformed region is never strictly compactly supported; measuring the exact support size on a dense sequence of masks would quantify how 'local' the control really is.
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

3 major / 5 minor

Summary. The paper proposes PaNDaS, a learning-based method for non-rigid deformation of triangle meshes. For a source mesh X (typically the neutral pose) and a target mesh Y, the method extracts per-triangle local features f_t with DiffusionNet and a global deformation code z = EncD(Y) - EncD(X), where EncD aggregates Laplacian-frequency projections of per-triangle features. The deformation generator G predicts per-triangle Jacobians from the concatenated feature field (f_t, z), and a Poisson solve yields vertex displacements. At inference, masking the global-code part of the feature field (f_t, M_t ⊙ z) is proposed as a mechanism for partial/localized deformation, and the paper also demonstrates feature-field interpolation, pose mixing, motion transfer, and local statistics. Experiments on MANO, DFAUST, and COMA report reconstruction errors in Table 1, interpolation errors in Tables 2 and 3, and qualitative partial-deformation results in Figures 5, 6, and 9, with an ablation study in Table 5.

Significance. If the localized-control mechanism works as claimed, PaNDaS would provide a useful and flexible tool for partial non-rigid mesh deformation without per-instance optimization, on top of a clearly described architecture that builds on Neural Jacobian Fields and DiffusionNet but introduces a novel combination of local feature fields, global deformation codes, and Laplacian-based aggregation. The main reconstruction and interpolation evaluations are not circular, and the paper includes ablations and comparisons on three benchmark datasets. However, the central localized-control claim rests on an identity property that is neither enforced during training nor quantitatively validated, and the abstract's unqualified 'state-of-the-art' statement is contradicted by the COMA row of Table 1. These issues are load-bearing for the paper's main contribution, so the manuscript requires substantial additional evidence before it can be accepted.

major comments (3)
  1. [Sec. 3.5 and Sec. 3.4, Eqs. (2)-(5)] The masked decoding operation (f_t, M_t ⊙ z) is never trained or quantitatively validated. When M_t = 0 the generator receives (f_t, 0), but no term in L in Eqs. (3)-(5) enforces G((f_t,0)) = X, and the training pairs use only the neutral pose as X, so the identity behavior is not guaranteed for non-neutral source meshes, which are exactly the inputs used for motion transfer and partial statistics. The Poisson solve couples all vertices, so even per-triangle identity Jacobians would not confine the deformation to the mask, and Fig. 6 reports only the norm of the deformation, not whether the outside-mask region equals X. The ablation in Table 5 does not test masked decoding. Please add an explicit identity constraint (e.g., a loss term on G((F(X),0)) = X) or provide a quantitative evaluation of leakage, such as outside-mask error against the source on unseen identities and non-neutral sources.
  2. [Abstract and Sec. 1 vs. Table 1] The abstract claims 'state-of-the-art accuracy' and Sec. 1 states that PaNDaS 'set a new state-of-the-art,' but Table 1 shows that ARAPReg outperforms PaNDaS on COMA on all three reported metrics (MSE 0.128 vs 0.160, HD 0.0070 vs 0.0086, CD 0.474 vs 0.839). The authors themselves describe the COMA result as 'competitive' in Sec. 4.2. This unqualified SOTA claim should be corrected to reflect that PaNDaS is best or tied on MANO and DFAUST but not on COMA, or the authors should explain why the COMA setting is deprioritized.
  3. [Sec. 4.4 and Fig. 5] Partial deformations are evaluated only qualitatively; there is no metric measuring how well the masked region matches the target or how well the anchored region is preserved. Since localized control is the paper's central novelty, the authors should report region-wise errors (e.g., MSE inside and outside the mask) and compare leakage against a baseline that enforces fixed vertices. This is closely related to the first major comment but concerns the experimental evaluation rather than the training objective.
minor comments (5)
  1. [Sec. 3.2] The number of Laplacian frequencies s=4 is stated as sufficient, but no sensitivity analysis is provided; please report how reconstruction and interpolation quality vary with s.
  2. [Fig. 3 caption] The caption refers to a 'per-triangle displacement field (v_i)' although v_i is defined as a per-vertex displacement in Sec. 3.1; please reword for consistency.
  3. [Tables 2 and 3] Some individual sequences are better with ARAPReg (e.g., DFAUST 'one leg jumping' and 'one leg landing', COMA 'high smile' and 'lips up'); the text should acknowledge these cases explicitly rather than reporting only the mean values.
  4. [Sec. 5] The limitations section mentions boundary artifacts from masking but does not discuss how these artifacts affect the mixing and statistics applications; a brief note would help readers understand the practical impact.
  5. [Fig. 1] The name 'PaNDAS' is written in all caps in Fig. 1 while the rest of the paper uses 'PaNDaS'; please unify the capitalization.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the main reconstruction and interpolation results are supervised against external targets on held-out test data, while the masking-based local-control operation, although relying on an unverified identity assumption, is not defined in terms of the outputs it is used to claim.

full rationale

PaNDaS's derivation chain is self-contained rather than circular. The deformation generator G maps a per-triangle feature field (ft, z) to a vertex displacement field, and the training objective in Sec. 3.4 (Eqs. 3-5) minimizes reconstruction error against target meshes Y on held-out identities. Reconstruction (Tab. 1) and interpolation (Tabs. 2-3) are evaluated against external baselines (LIMP, VCMC, ARAPReg, NJF) on public datasets, so those central claims rest on independent benchmarks rather than on the paper's own fitted values. The localized-control operation in Sec. 3.5 decodes (ft, Mt ⊙ z); when Mt = 0, this implicitly assumes G((ft,0)) = X for the source mesh, but this identity is an unenforced inference-time assumption, not a circularity: no training loss or network definition is expressed in terms of the masked output, and the paper explicitly concedes that the Poisson solve provides no strict confinement guarantee (Sec. 3.5) and that masking can create boundary artifacts (Sec. 5). The partial-deformation experiments in Sec. 4.4 are qualitative rather than quantitatively benchmarked, which is a validation gap, not a circularity. Self-citations (e.g., DiffusionNet [34], which shares an author) are architectural building blocks with independent public implementations and are not used to justify the paper's specific deformation claims. Hence no load-bearing step reduces by construction to its own inputs.

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

The central claim rests on the learned feature field and the masking operation. The tunable hyperparameters (feature dimensions, number of Laplacian frequencies, loss weight, learning rate, epochs) are free parameters that affect the reported numbers. No new physical entities are introduced. The main axioms are the representational assumptions borrowed from Neural Jacobian Fields and DiffusionNet, plus the trained linearity of the latent space under masking and interpolation.

free parameters (6)
  • Global encoding dimension r = 64
    Dimension of the global deformation vector z; chosen by the authors.
  • Local feature dimension l = 64
    Dimension of per-triangle DiffusionNet features; chosen by the authors.
  • Number of Laplacian frequencies s = 4
    Number of eigenvector projections used in the global aggregation; stated as 's = 4 proved to be sufficient'.
  • Normal loss weight lambda_n = 10^-5
    Weight of the normal-map cosine similarity loss in the total loss.
  • Learning rate = 10^-4
    ADAM optimizer learning rate for all models.
  • Training epochs = 1000
    Number of epochs used to train all models.
assumptions (5)
  • domain assumption A mesh deformation can be represented by a per-face Jacobian field, and the closest vertex displacement is obtained by solving the Poisson equation.
    Adopted from Neural Jacobian Fields [1], used in Sec 3.3 to generate the deformed mesh.
  • domain assumption The cotangent Laplacian eigenvectors of a mesh provide a useful low-frequency basis for aggregating per-face features into a global code.
    Used in Eq (1); the authors set s=4 and rely on the spectral projection to summarize the target shape.
  • domain assumption DiffusionNet features are robust to remeshing and informative enough to drive deformation.
    The feature extractor is built on DiffusionNet [34]; the paper assumes its local features combine well with the global code.
  • domain assumption Training and evaluation assume registered, aligned meshes with a known neutral pose per identity.
    Stated in Sec 4.1 and Limitations; the partial-deformation applications are demonstrated on this setting.
  • ad hoc to paper The learned deformation space behaves linearly under interpolation of the global code and under masking.
    Sec 3.5 relies on linear interpolation (1-alpha)z and mask products Mt*z to produce plausible partial deformations; this linearity is not guaranteed and is only qualitatively validated.

how reviews work

0 comments
Cite this review

Pith. "Pith review of PaNDaS: Learnable Deformation Modeling with Localized Control." pith.science (2026). https://pith.science/paper/L2VINPUG

@misc{pith2026241202306,
  author       = {Pith},
  title        = {Pith review of: PaNDaS: Learnable Deformation Modeling with Localized Control},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/L2VINPUG}},
  note         = {Machine review of arXiv:2412.02306}
}
read the original abstract

Non-rigid shape deformations pose significant challenges, and most existing methods struggle to handle partial deformations effectively. We propose to learn deformations at the point level, which allows for localized control of 3D surface meshes, enabling Partial Non-rigid Deformations and interpolations of Surfaces (PaNDaS). Unlike previous approaches, our method can restrict the deformations to specific parts of the shape in a versatile way. Moreover, one can mix and combine various poses from the database, all while not requiring any optimization at inference time. We demonstrate state-of-the-art accuracy and greater locality for shape reconstruction and interpolation compared to approaches relying on global shape representation across various types of human surface data. We also demonstrate several localized shape manipulation tasks and show that our method can generate new shapes by combining different input deformations. Code and data will be made available after the reviewing process.

Figures

Figures reproduced from arXiv: 2412.02306 by the authors.

Figure 1
Figure 1. We present PaNDAS: a robust learning-based approach to learn non-rigid deformations of triangular meshes applied to human body surfaces. By combining per-face learned features on a neutral pose mesh (in blue) with a global encoding of a deformed mesh (in purple), the model enables localized control of deformations (in cyan) with user-chosen regions (masked in red on the left). Notably, this approach permits several … view at source ↗
Figure 2
Figure 2. Shape interpolation between unregistered hand meshes. The top row shows a predicted interpolation sequence from a hand with missing ring finger. The bottom row shows an interpolation between two raw scans. deformations and interpolations without a complex config￾uration. Indeed, maintaining geometric consistency and en￾suring smooth transitions between deformed and anchored regions usually require manual corrections… view at source ↗
Figure 3
Figure 3. Overview of the proposed architecture. A pair of meshes is given as input: a source pose X and a target pose Y. Per-triangle features (ft)t are learned on the source mesh and are enriched with a global encoding z of the target mesh to obtain a feature field ( ˜ft)t over the source mesh. Finally, the deformation generator learns a Jacobian field from the feature field over the neutral mesh to predict a first-order, p… view at source ↗
Figures from the paper (10 more)
Figure 4
Figure 4. Figure 4: Qualitative comparisons of the regularization effect when adding a loss term on the normal map. On the bottom left is a generation mid-interpolation for a model trained only with the MSE loss, on the right is a similar generated mesh but with a model trained with a los…
Figure 5
Figure 5. Figure 5: Qualitative comparisons of partial deformations on DFAUST, COMA and MANO. The As-Rigid-As-Possible deformations handle efficiently deformations that are close to linear (2nd and last row) but fail to adapt to large deviations. MANO DFAUST COMA MSE ↓ HD ↓ CD (10−5 ) ↓ M…
Figure 6
Figure 6. Figure 6: Restriction of the deformation field. We show an ex￾ample of partial deformation, restricted to the red parts on the left. We show the corresponding norm of the predicted jacobians and the resulting gradient of the deformation field (after the Poisson solve). We highli…
Figure 7
Figure 7. Figure 7: Applications of PaNDaS–Pose mixing: By combining the feature fields of different poses, we can combine deformations to generate new poses. Transfer: The deformation encoder is trained to encode the pose which allow partial motion transfer. Statistics: The learned featu…
Figure 8
Figure 8. Figure 8: Pose interpolation examples. We compare results ob￾tained using latent linear interpolation with ARAPReg (top row) to the same interpolations computed using PaNDaS (bottom row). The latter handles better large and highly non linear deformations. ing [15, 44], to train …
Figure 9
Figure 9. Figure 9: Pose mixing using partial interpolations of a neutral pose with two other poses. The first pose raises both arms, the second pose raises the right knee. We mix them and highlight how our method (on the right) computes smoother deformations and avoids limb shrinking (ri…
Figure 10
Figure 10. Figure 10: Qualitative comparison of latent interpolation. Our method shows more realistic interpolation trajectories, with lim￾ited distortion of the hand even with large non-linear deformations. VCMC Source Ours Target [PITH_FULL_IMAGE:figures/full_fig_p011_10.png]
Figure 11
Figure 11. Figure 11: Qualitative comparison of partial latent interpola￾tion. We computed a partial latent interpolation on an unseen iden￾tity between a source pose and a target pose displayed in yellow. The interpolated part is the left arm (dashed case). Previous com￾parable method VCM…
Figure 13
Figure 13. Figure 13: Robustness of the mean shape estimation. We perform mean shape estimation from a collection of 11 poses of hands. We predict the feature field associated with each pose and compute the Euclidean mean. The mean feature field is then used to compute the mean pose. In th…
Figure 14
Figure 14. Figure 14: Estimated mean from the same collection of hands. From 10 different hand pose from MANO, we compute a linear mean (on the left) and the ”latent mean” of the registered meshes along with the latent mean of the raw scans (unregistered) [PITH_FULL_IMAGE:figures/full_fig…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 44 canonical work pages

  1. [1]

    Kim, Siddhartha Chaudhuri, Jun Saito, and Thibault Groueix

    Noam Aigerman, Kunal Gupta, Vladimir G. Kim, Siddhartha Chaudhuri, Jun Saito, and Thibault Groueix. Neural jacobian fields: learning intrinsic mappings of arbitrary meshes. ACM Trans. Graph., 41(4), 2022. 2, 3, 5, 6, 7

    work page 2022

  2. [2]

    As- rigid-as-possible shape interpolation

    Marc Alexa, Daniel Cohen-Or, and David Levin. As- rigid-as-possible shape interpolation. In Proceedings of the 27th Annual Conference on Computer Graphics and Interactive Techniques, page 157–164, USA, 2000. ACM Press/Addison-Wesley Publishing Co. 2, 3

    work page 2000

  3. [3]

    Ncp: Neural cor- respondence prior for effective unsupervised shape match- ing

    Souhaib Attaiki and Maks Ovsjanikov. Ncp: Neural cor- respondence prior for effective unsupervised shape match- ing. Advances in Neural Information Processing Systems , 35:28842–28857, 2022. 4

    work page 2022

  4. [4]

    Martin Bauer, Philipp Harms, and Peter W. Michor. Sobolev metrics on shape space of surfaces, 2011. 2, 3

    work page 2011

  5. [5]

    A numerical framework for elastic surface match- ing, comparison, and interpolation

    Martin Bauer, Nicolas Charon, Philipp Harms, and Hsi-Wei Hsieh. A numerical framework for elastic surface match- ing, comparison, and interpolation. International Journal of Computer Vision, 129:2425 – 2444, 2020. 3

    work page 2020

  6. [6]

    Computing large deformation metric mappings via geodesic flows of diffeomorphisms

    Mirza Faisal Beg, Michael Miller, Alain Trouv´e, and Laurent Younes. Computing large deformation metric mappings via geodesic flows of diffeomorphisms. International Journal of Computer Vision, 61:139–157, 2005. 3

    work page 2005

  7. [7]

    Federica Bogo, Javier Romero, Gerard Pons-Moll, and Michael J. Black. Dynamic FAUST: Registering human bod- ies in motion. InIEEE Conf. on Computer Vision and Pattern Recognition (CVPR), 2017. 6

    work page 2017

  8. [8]

    Primo: coupled prisms for intuitive surface modeling

    Mario Botsch, Mark Pauly, Markus Gross, and Leif Kobbelt. Primo: coupled prisms for intuitive surface modeling. In Proceedings of the Fourth Eurographics Symposium on Ge- ometry Processing, page 11–20, Goslar, DEU, 2006. Euro- graphics Association. 2, 3

    work page 2006

Show all 48 references
  1. [9]

    Neural 3d mor- phable models: Spiral convolutional networks for 3d shape representation learning and generation

    Giorgos Bouritsas, Sergiy Bokhnyak, Stylianos Ploumpis, Michael Bronstein, and Stefanos Zafeiriou. Neural 3d mor- phable models: Spiral convolutional networks for 3d shape representation learning and generation. In The IEEE Inter- national Conference on Computer Vision (ICCV), 2019. 3

    2019

  2. [10]

    D. Cao, M. Eisenberger, N. El Amrani, D. Cremers, and F. Bernard. Spectral meets spatial: Harmonising 3d shape matching and interpolation. In 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 3658–3668, Los Alamitos, CA, USA, 2024. IEEE Computer...

    2024

  3. [11]

    Limp: Learning latent shape rep- resentations with metric preservation priors

    Luca Cosmo, Antonio Norelli, Oshri Halimi, Ron Kimmel, and Emanuele Rodol `a. Limp: Learning latent shape rep- resentations with metric preservation priors. In Computer Vision – ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part III, page 1...

    2020

  4. [12]

    Neuromorph: Unsupervised shape interpo- lation and correspondence in one go

    Marvin Eisenberger, David Novotny, Gael Kerchenbaum, Patrick Labatut, Natalia Neverova, Daniel Cremers, and An- drea Vedaldi. Neuromorph: Unsupervised shape interpo- lation and correspondence in one go. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern R...

    2021

  5. [13]

    Spiralnet++: A fast and highly efficient mesh convolution operator

    Shunwang Gong, Lei Chen, Michael Bronstein, and Stefanos Zafeiriou. Spiralnet++: A fast and highly efficient mesh convolution operator. In Proceedings of the IEEE Interna- tional Conference on Computer Vision Workshops, pages 0– 0, 2019. 3

    2019

  6. [14]

    Kim, Bryan Russell, and Mathieu Aubry

    Thibault Groueix, Matthew Fisher, Vladimir G. Kim, Bryan Russell, and Mathieu Aubry. 3d-coded : 3d correspondences by deep deformation. In ECCV, 2018. 3

    2018

  7. [15]

    Unsupervised representation learning for di- verse deformable shape collections

    Sara Hahner, Souhaib Attaiki, Jochen Garcke, and Maks Ovsjanikov. Unsupervised representation learning for di- verse deformable shape collections. In International Con- ference on 3D Vision 2024, 2024. 3, 8

    2024

  8. [16]

    BaRe-ESA: A rieman- nian framework for unregistered human body shapes

    Emmanuel Hartman, Emery Pierson, Martin Bauer, Nico- las Charon, and Mohamed Daoudi. BaRe-ESA: A rieman- nian framework for unregistered human body shapes. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 14181–14191, 2023. 2, 3

    2023

  9. [17]

    Elastic shape analysis of sur- faces with second-order sobolev metrics: A comprehensive numerical framework

    Emmanuel Hartman, Yashil Sukurdeep, Eric Klassen, Nico- las Charon, and Martin Bauer. Elastic shape analysis of sur- faces with second-order sobolev metrics: A comprehensive numerical framework. International Journal of Computer Vision, 2023. 3

    2023

  10. [18]

    Basis restricted elastic shape analysis on the space of unregistered surfaces

    Emmanuel Hartman, Emery Pierson, Martin Bauer, Mo- hamed Daoudi, and Nicolas Charon. Basis restricted elastic shape analysis on the space of unregistered surfaces. Inter- national Journal of Computer Vision, 2024. 2

    2024

  11. [19]

    Huang, X

    Q. Huang, X. Huang, B. Sun, Z. Zhang, J. Jiang, and C. Ba- jaj. Arapreg: An as-rigid-as possible regularization loss for learning deformable shape generators. In 2021 IEEE/CVF International Conference on Computer Vision (ICCV), pages 5795–5805, Los Alamitos, CA, USA, 2021. IEE...

    2021

  12. [20]

    Elastic shape matching of parameterized surfaces using square root normal fields

    Ian H Jermyn, Sebastian Kurtek, Eric Klassen, and Anuj Sri- vastava. Elastic shape matching of parameterized surfaces using square root normal fields. In European conference on computer vision, pages 804–817. Springer, 2012. 3

    2012

  13. [21]

    Geo- metric modeling in shape space

    Martin Kilian, Niloy J Mitra, and Helmut Pottmann. Geo- metric modeling in shape space. In ACM SIGGRAPH 2007 papers, pages 64–es. 2007. 2, 3

    2007

  14. [22]

    Kingma and Jimmy Ba

    Diederik P. Kingma and Jimmy Ba. Adam: A method for stochastic optimization. In 3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, May 7-9, 2015, Conference Track Proceedings, 2015. 10

    2015

  15. [23]

    Elastic geodesic paths in shape space of parameterized surfaces

    Sebastian Kurtek, Eric Klassen, John C Gore, Zhaohua Ding, and Anuj Srivastava. Elastic geodesic paths in shape space of parameterized surfaces. IEEE transactions on pattern anal- ysis and machine intelligence, 34(9):1717–1730, 2011. 3

    2011

  16. [24]

    Numerical inversion of srnf maps for elastic shape analysis of genus-zero surfaces

    Hamid Laga, Qian Xie, Ian H Jermyn, and Anuj Srivastava. Numerical inversion of srnf maps for elastic shape analysis of genus-zero surfaces. IEEE transactions on pattern analy- sis and machine intelligence, 39(12):2451–2464, 2017. 3

    2017

  17. [25]

    Tianye Li, Timo Bolkart, Michael. J. Black, Hao Li, and Javier Romero. Learning a model of facial shape and ex- pression from 4D scans. ACM Transactions on Graphics, (Proc. SIGGRAPH Asia), 36(6):194:1–194:17, 2017. 2

    2017

  18. [26]

    Matthew Loper, Naureen Mahmood, Javier Romero, Ger- ard Pons-Moll, and Michael J. Black. SMPL: A skinned multi-person linear model. ACM Trans. Graphics (Proc. SIGGRAPH Asia), 34(6):248:1–248:16, 2015. 2

    2015

  19. [27]

    Controllable mesh generation through sparse latent point diffusion models

    Zhaoyang Lyu, Jinyi Wang, Yuwei An, Ya Zhang, Dahua Lin, and Bo Dai. Controllable mesh generation through sparse latent point diffusion models. In 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 271–280, 2023. 3

    2023

  20. [28]

    Smooth non-rigid shape matching via effective dirichlet energy optimization

    Robin Magnet, Jing Ren, Olga Sorkine-Hornung, and Maks Ovsjanikov. Smooth non-rigid shape matching via effective dirichlet energy optimization. In 2022 International Con- ference on 3D Vision (3DV), pages 495–504, Los Alamitos, CA, USA, 2022. IEEE Computer Society. 6

    2022

  21. [29]

    Sanjeev Muralikrishnan, Niladri Shekhar Dutt, Siddhartha Chaudhuri, Noam Aigerman, Vladimir Kim, Matthew Fisher, and Niloy J. Mitra. Temporal residual jacobians for rig-free motion transfer, 2024. 3

    2024

  22. [30]

    Neural face rigging for animating and retargeting facial meshes in the wild

    Dafei Qin, Jun Saito, Noam Aigerman, Groueix Thibault, and Taku Komura. Neural face rigging for animating and retargeting facial meshes in the wild. In SIGGRAPH 2023 Conference Papers, 2023. 3

    2023

  23. [31]

    Anurag Ranjan, Timo Bolkart, Soubhik Sanyal, and Michael J. Black. Generating 3D faces using convolutional mesh autoencoders. In European Conference on Computer Vision (ECCV), pages 725–741, 2018. 6

    2018

  24. [32]

    Javier Romero, Dimitrios Tzionas, and Michael J. Black. Embodied hands: Modeling and capturing hands and bod- ies together. ACM Transactions on Graphics, (Proc. SIG- GRAPH Asia), 36(6), 2017. 2, 6

    2017

  25. [33]

    A generalized solution of the orthogonal procrustes problem

    Peter Sch ¨onemann. A generalized solution of the orthogonal procrustes problem. Psychometrika, 31(1):1–10, 1966. 6

    1966

  26. [34]

    Diffusionnet: Discretization agnostic learning on surfaces

    Nicholas Sharp, Souhaib Attaiki, Keenan Crane, and Maks Ovsjanikov. Diffusionnet: Discretization agnostic learning on surfaces. ACM Trans. Graph., 01(1), 2022. 4

    2022

  27. [35]

    Dragdiffusion: Harnessing diffu- sion models for interactive point-based image editing

    Yujun Shi, Chuhui Xue, Jiachun Pan, Wenqing Zhang, Vin- cent YF Tan, and Song Bai. Dragdiffusion: Harnessing diffu- sion models for interactive point-based image editing. arXiv preprint arXiv:2306.14435, 2023. 3

    2023 arXiv

  28. [36]

    Fast blended transformations for partial shape registration

    Alon Shtern, Matan Sela, and Ron Kimmel. Fast blended transformations for partial shape registration. Journal of Mathematical Imaging and Vision, 60:913 – 928, 2016. 2

    2016

  29. [37]

    As-rigid-as-possible surface modeling

    Olga Sorkine and Marc Alexa. As-rigid-as-possible surface modeling. In Proceedings of the Fifth Eurographics Sympo- sium on Geometry Processing, page 109–116, Goslar, DEU,

  30. [38]

    Deformation recovery: Localized learning for detail-preserving deformations

    Ramana Sundararaman, Nicolas Donati, Simone Melzi, Eti- enne Corman, and Maks Ovsjanikov. Deformation recovery: Localized learning for detail-preserving deformations. ACM Transactions on Graphics (TOG), 43(6):1–16, 2024. 2

    2024

  31. [39]

    Vari- ational autoencoders for deforming 3d mesh models

    Qingyang Tan, Lin Gao, Yu-Kun Lai, and Shihong Xia. Vari- ational autoencoders for deforming 3d mesh models. In 2018 IEEE/CVF Conference on Computer Vision and Pat- tern Recognition, pages 5841–5850, 2018. 3

    2018

  32. [40]

    Variational autoencoders for localized mesh de- formation component analysis

    Qingyang Tan, Ling-Xiao Zhang, Jie Yang, Yu-Kun Lai, and Lin Gao. Variational autoencoders for localized mesh de- formation component analysis. IEEE Transactions on Pat- tern Analysis and Machine Intelligence , 44(10):6297–6310,

  33. [41]

    Dragd3d: Realistic mesh editing with rigidity control driven by 2d diffusion priors, 2024

    Tianhao Xie, Eugene Belilovsky, Sudhir Mudur, and Tiberiu Popa. Dragd3d: Realistic mesh editing with rigidity control driven by 2d diffusion priors, 2024. 3, 6

    2024

  34. [42]

    Drag your gan: Interactive point-based manipulation on the generative image manifold

    Pan Xingang, Tewari Ayush, Leimkuhler Thomas, Li Lingjie, Meka Abhimitra, and Theobalt Christian. Drag your gan: Interactive point-based manipulation on the generative image manifold. In ACM SIGGRAPH 2023 Conference Pro- ceedings, 2023. 3

    2023

  35. [43]

    Geolatent: A geometric approach to latent space design for deformable shape generators.ACM Trans

    Haitao Yang, Bo Sun, Liyan Chen, Amy Pavel, and Qixing Huang. Geolatent: A geometric approach to latent space design for deformable shape generators.ACM Trans. Graph., 42(6), 2023. 3

    2023

  36. [44]

    Bajaj, and Qixing Huang

    Haitao Yang, Xiangru Huang, Bo Sun, Chandrajit L. Bajaj, and Qixing Huang. Gencorres: Consistent shape matching via coupled implicit-explicit shape generative models. InThe Twelfth International Conference on Learning Representa- tions, 2024. 3, 8

    2024

  37. [45]

    Kim, and Minhyuk Sung

    Seungwoo Yoo, Kunho Kim, Vladimir G. Kim, and Minhyuk Sung. As-Plausible-As-Possible: Plausibility-Aware Mesh Deformation Using 2D Diffusion Priors. In CVPR, 2024. 3, 6

    2024

  38. [46]

    Fully convolutional mesh autoencoder using efficient spatially varying kernels

    Yi Zhou, Chenglei Wu, Zimo Li, Chen Cao, Yuting Ye, Ja- son Saragih, Hao Li, and Yaser Sheikh. Fully convolutional mesh autoencoder using efficient spatially varying kernels. In Proceedings of the 34th International Conference on Neu- ral Information Processing Systems , Red H...

  39. [2007]

    Eurographics Association. 2, 3

  40. [2020]

    Curran Associates Inc. 3, 6, 7

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.