Pith. sign in

REVIEW 4 major objections 6 minor 59 references

NeuraLeaf: Neural Parametric Leaf Models with Shape and Deformation Disentanglement

T0 review · 4 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read NeuraLeaf models each leaf as a flat 2D base shape plus a learned 3D deformation, and fits partial 3D observations of leaves with much lower Chamfer distance than existing parametric leaf models.

desk verdict A solid first neural parametric leaf model with a valuable dataset; the quantitative edge over NPM is real but overstated by an unfair baseline and small test set. read the letter →

arxiv 2507.12714 v2 pith:YQFA5JIU submitted 2025-07-17 cs.CV cs.GR

classification cs.CVcs.GR
keywords neuralparametricmodelleafreconstruction3Ddeformationmodelingskeleton-freeskinninglinearblendshapeanddisentanglementDeformdatasetsigneddistancefunction
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

NeuraLeaf is a neural parametric model built around the observation that a flattened leaf is nearly a 2D plane. The paper argues that splitting leaf geometry into a 2D base shape and a 3D deformation makes it possible to learn realistic leaf shape from abundant 2D leaf-image datasets while learning deformation from a comparatively small 3D dataset. To support the deformation side, the authors introduce a skeleton-free skinning model with up to a thousand control points and a new captured dataset, DeformLeaf, of roughly 300 base-to-deformed leaf pairs. On the 30-leaf DeformLeaf test set, fitting NeuraLeaf to single-view point clouds reaches a Chamfer distance of 2.1 mm and normal consistency of 0.973, versus 15.1 mm and 0.961 for the neural parametric baseline. If this stands, 3D leaf reconstruction and generation become practical from inputs as simple as depth maps and point clouds.

What carries the argument

The load-bearing object is the skeleton-free neural blend skinning model. Instead of a skeletal joint hierarchy, NeuraLeaf places up to 1,000 control points uniformly over the leaf's UV plane; the skinning weight decoder maps each base-shape vertex, conditioned on the shape latent code, to a distribution over control points, and the transformation decoder maps each control point, conditioned only on the deformation latent code, to a rigid transformation. Linear blend skinning then deforms the base mesh. Because leaves have no common skeleton and deform continuously, the many-control-point design is what lets the model capture fine surface details such as venation while keeping the deformation latent space independent of species. The accompanying deformation mapping loss ties the norm of the deformation code to the chamfer distance between base and deformed shape, which gives the latent space a meaningful zero-mean Gaussian structure.

What would settle it

Take a set of strongly curled leaves, scan them with an independent high-resolution laser or structured-light scanner, fit NeuraLeaf to those scans, and compare against the PCA, B-spline, and NPM baselines; if the Chamfer advantage over the NPM baseline shrinks to noise, or if deformation codes no longer order by actual bending amount, the claim that the learned deformation space is accurate is refuted.

Watch

Extended reading notes

Core claim

The paper's central claim is that a leaf's full 3D geometry can be faithfully represented by two independent latent codes: a shape latent code that controls the flattened base shape, and a deformation latent code that controls how that base mesh bends in 3D. The base shape is a signed distance function in the plane—each point records its signed distance to the leaf outline—converted to a base mesh, so the shape space can be trained on 2D leaf-image data. The deformation is produced by a skeleton-free linear blend skinning model: a dense set of up to 1,000 control points on the base plane, with skinning weights predicted from the shape latent and rigid transformations (quaternion plus translation) predicted from the deformation latent alone. Because the deformation space is shared across all base shapes, the same deformation code applies to different leaf species, and fitting an observed point cloud or depth map amounts to inverting the two latent codes and then refining them by direct optimization. The paper reports that this representation, trained with a chamfer-based deformation mapping loss on the new DeformLeaf dataset, reconstructs test leaves with substantially lower Chamfer distance than PCA, B-spline, and human-oriented neural parametric baselines.

Load-bearing premise

The reported accuracy rests on the assumption that the DeformLeaf ground truth—photogrammetric 3D scans of deformed leaves aligned to flattened 2D photos through rigid and non-rigid registration—is unbiased; if that alignment is wrong for strongly curved leaves, both the learned deformation space and the reported Chamfer improvements inherit the error.

Editorial extensions

If this is right

  • Because base shapes are 2D, the shape space can be trained on large existing collections of 2D leaf scans, sidestepping the scarcity of 3D leaf data.
  • Because the deformation code is shared across species, a deformation learned on one leaf can be transferred to another base shape, giving a simple route to generating deformed foliage in computer graphics.
  • Reconstruction from partial observations reduces to latent-code inversion followed by direct refinement, so depth maps and point clouds from RGB-D cameras can be turned into complete leaf models.
  • Texture is generated as a UV map aligned with the base shape, so appearance stays consistent with geometry during deformation.
  • When reconstructing multiple occluded leaves from a top-view RGB-D image, sharing the shape latent across instances recovers leaf areas hidden by occlusion.

Reading between the lines

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

  • If the disentanglement holds, the deformation space should transfer not only to unseen species but to synthetic or edited base shapes; a cheap test is to deform procedurally generated leaves with latent codes fitted to real scans and check physical plausibility.
  • The paper's own reported limitation that rare deformation types are undercovered suggests the bottleneck is dataset species diversity rather than representation; adding scans of strongly curled, lobed, or needle-like leaves would show whether the 1,000-point skinning model scales.
  • The reported accuracy depends on photogrammetric ground truth registered by ARAP and CPD; an independent evaluation against laser scans would separate representation quality from dataset bias.
  • Because the transformation decoder is conditioned only on the deformation code, the model implicitly assumes a canonical alignment between base and deformed shapes; future work might test whether the learned space remains coherent when leaves are observed from arbitrary initial poses.
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 / 6 minor

Summary. The paper proposes NeuraLeaf, a neural parametric model for 3D leaves that separates leaf geometry into a 2D base shape (learned from 2D leaf silhouette datasets) and a 3D deformation (learned from a newly captured DeformLeaf dataset of paired flattened and deformed leaves). Deformation is modeled by a skeleton-free linear blend skinning scheme with up to 1,000 control points, where a shape-conditioned decoder predicts skinning weights and a deformation-conditioned decoder predicts rigid transformations. The paper further contributes a correspondence-free training objective for deformation, a regularizer that links the deformation latent norm to Chamfer magnitude, and latent-code inversion for fitting to depth/point-cloud observations. Experiments on single-leaf and multi-leaf reconstruction compare NeuraLeaf with PCA, B-spline, and neural parametric model (NPM) baselines, reporting substantially lower Chamfer distances, plus qualitative results for generation, interpolation, deformation transfer, and comparison with image/text-to-3D methods.

Significance. NeuraLeaf is a sensible and potentially useful representation for leaf modeling: the disentanglement leverages abundant 2D leaf data, the skeleton-free skinning avoids the need for a common leaf skeleton, and the authors release both code and a new dataset (DeformLeaf) with dense correspondences. The correspondence-free training is a genuine advantage over displacement-field NPMs, and the qualitative deformation-transfer results support the plausibility of a shared deformation space. However, the strength of the contribution is not yet matched by the evidence: the quantitative claims rest on a small test set with no error bars, and the main baseline comparison may be confounded by the very registration pipeline the baseline depends on, so the current evaluation cannot fully support the magnitude of the reported improvement.

major comments (4)
  1. [§3.3, Eq. (7)] Equation (7) is not a valid linear blend skinning formula as written. In homogeneous coordinates, tilde{v}'_i - tilde{c}_k = [v'_i - c_k, 0]^T, so applying T_k and taking the weighted sum yields a vector with zero homogeneous component (a direction), not a point in 3D. The standard formulation is tilde{v}_i = sum_k w_{i,k} T_k tilde{v}'_i, or, for transformations about control points, tilde{v}_i = sum_k w_{i,k}(T_k(tilde{v}'_i - tilde{c}_k) + tilde{c}_k). Please correct the equation and confirm that the implemented code uses the corrected form; because Eq. (7) defines the entire deformation model, this is a load-bearing issue.
  2. [§6.2, Tables 1 and 2] Table 1 reports a single aggregate Chamfer distance over 30 test leaves, and Table 2 reports results on 5 plants, with no error bars, confidence intervals, or significance tests. Given the large claimed improvement (15.1 → 2.1 mm), the authors should report per-leaf/per-plant distributions (or at least standard deviations) and a paired test, and ideally a per-species breakdown, to show that the gain is systematic rather than driven by outliers. Without this, the headline quantitative claim is not statistically grounded.
  3. [Suppl. B.3 vs. §3.3–§4] The comparison with the NPM baseline is potentially unfair because the baseline is trained using the dense correspondences in DeformLeaf (Suppl. B.3), which are produced by the ARAP+CPD pipeline (Sec. 4, Suppl. A.2–A.3), whereas NeuraLeaf is trained with a correspondence-free Chamfer loss (Eq. 11). If ARAP/CPD alignments are inaccurate for strongly curved or occluded leaves, the NPM baseline is penalized by corrupted displacement targets, while NeuraLeaf is not. The paper should quantify registration accuracy (e.g., against manual landmarks) on the test set, and/or train an NPM variant with a fitting objective that does not rely on those correspondences, to demonstrate that the 7× improvement reflects the representation rather than registration noise. Relatedly, because training and test deformed shapes both come from the same Metashape reconstruction pipeline, an evaluation on independent data (e.g., synthetic leaves with known ground-truth deformation) would help rule out that the reported numbers are tied to the specific reconstruction biases of that pipeline.
  4. [§3.3, Eq. (10)] Equation (10) defines Lmap = (dcham(Sb,Sd)/||zd||2 − phi)^2, but the Chamfer distance between the base and deformed shape is not a pure deformation-magnitude measure: for the same physical deformation, a larger or differently shaped leaf yields a different dcham. The regularizer therefore couples the deformation latent norm to base-shape geometry, which may weaken the claimed disentanglement between z_s and z_d. The paper should analyze this coupling and ensure the denominator is safely away from zero (e.g., add a small epsilon or a lower bound on ||zd||2).
minor comments (6)
  1. [§3.1, Eq. (5)] The loss term L_sil is used in the training objective but is never defined in the text; please define it.
  2. [§6.1, Table 1 caption] The meaning of the columns 'Corres-free' and 'Temp-free' is not explained; please state explicitly whether each method requires dense correspondences or a template.
  3. [§4 and Fig. 2] The text says DeformLeaf contains approximately 300 base-deformation pairs, while the figure caption reads '≈ 300 species & 4000 instances'; please disambiguate which quantity refers to DeformLeaf and which to the 2D dataset.
  4. [§6.2, Table 2] The PCA baseline reports a Chamfer distance of 1311.3 mm, an order of magnitude larger than the other entries; please check for a misalignment or typo and, if real, discuss why it is so large.
  5. [§3.2 and Fig. 2] The texture generator is denoted f_theta_t in the text but the figure uses 'E/D'; please unify the notation.
  6. [§6.2] The multi-leaf reconstruction section says the shape latent is shared for observations of the same plant species, but the clustering and anchor details are only in Suppl. C.4; consider moving the essential mechanism (K-means, anchor regularization) to the main text or giving a one-sentence summary, since it is important for interpreting Table 2.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity found: NeuraLeaf's core claims are supported by held-out fitting experiments and a newly captured dataset, with no load-bearing self-citation or construction-equivalent prediction.

full rationale

The paper's central derivation chain is self-contained. The deformation space is trained on the newly captured DeformLeaf dataset by minimizing a correspondence-free Chamfer loss against real 3D scans (Sec. 3.3, Eq. 11), while the reported reconstruction accuracy is measured on a held-out 30-leaf test set (Sec. 6.2, Table 1) using Chamfer distance and normal consistency against ground-truth surfaces reconstructed by Metashape. The training objective and the evaluation metric are both Chamfer-based, but the targets are raw deformed scans during training versus ground-truth surface reconstructions at test time; the test shapes are not the same instances used for optimization, so there is no fit-to-target circularity. The shape space is learned from external 2D leaf image datasets, and the texture space uses a standard CycleGAN formulation; neither is defined in terms of the deformation claims. The NPM baseline is trained using the dense correspondences produced by ARAP and CPD registration (Suppl. B.3), whereas NeuraLeaf is correspondence-free, which raises a legitimate question of baseline fairness or sensitivity to registration quality, but this is an experimental-comparison concern rather than a circular-reasoning defect: NeuraLeaf's reported numbers are not forced by the baseline's construction or by the paper's own definitions. The paper contains no load-bearing self-citations and does not import a uniqueness theorem or ansatz from the authors' prior work; the skeleton-free skinning model, deformation mapping loss, and DeformLeaf dataset are introduced in this paper with stated assumptions. The limitation section acknowledges restricted deformation diversity and texture-fitting difficulty, which does not indicate circularity. Overall, no specific reduction of a claimed result to its inputs by construction could be exhibited, so the appropriate finding is no significant circularity.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central reconstruction accuracies rest on hyperparameters (control point count, latent dimensionality, Gaussian priors) and domain assumptions (planar base shape, LBS expressiveness, skinning-weight similarity). The most consequential is K=1000, which is ablated: dropping it raises Chamfer distance from 2.1 mm to 8.3-13.7 mm, so the headline result is sensitive to this hand-chosen capacity.

free parameters (6)
  • K, number of control points = 1000
    Chosen by hand; ablation (Table 3) shows reconstruction quality degrades with 100 or 500 points, so the central reconstruction result depends on this capacity choice.
  • Latent dimensions Ns, Nd, Nt = 256 each
    Set without sensitivity analysis; capacity of the shape, deformation, and texture spaces.
  • Latent regularization sigmas sigma_s, sigma_d = 10
    Spherical covariance of the zero-mean Gaussian priors in Eqs. (4) and (11); controls how far latent codes can move during fitting.
  • SDF truncation delta_tau = 0.01
    Truncation width in the SDF loss, Eq. (3); set without ablation.
  • Learnable scaling phi in Lmap = learned
    Adaptable scale mapping Chamfer distance between base and deformed shapes to latent norm in Eq. (10); a fitted value, not derived.
  • Number of similar base shapes m in second-stage training = not specified
    The second training stage selects m similar base shapes by mask IoU; m and the similarity threshold are not reported.
assumptions (4)
  • domain assumption Flattened leaves can be approximated as a 2D plane
    Justifies representing base shape as a 2D SDF and learning it from 2D images (Section 3.1).
  • domain assumption Leaf deformation can be represented by linear blend skinning with 1000 control points
    The skeleton-free skinning model in Eq. (7) assumes a fixed control-point layout and learned skinning weights can express natural leaf deformations, including venation detail (Section 3.3).
  • domain assumption Similar base shapes should have similar skinning weights
    Second-stage deformation training encourages paired shapes with high mask IoU to share skinning weights; if false, the regularization could distort the deformation space (Section 3.3, second stage).
  • domain assumption Shape and deformation latent spaces are zero-mean Gaussians
    Gaussian latent priors in Eqs. (4) and (11) are used for regularization and for latent inversion initialization; the paper does not validate that the learned spaces are actually Gaussian.

how reviews work

0 comments
Cite this review

Pith. "Pith review of NeuraLeaf: Neural Parametric Leaf Models with Shape and Deformation Disentanglement." pith.science (2026). https://pith.science/paper/YQFA5JIU

@misc{pith2026250712714,
  author       = {Pith},
  title        = {Pith review of: NeuraLeaf: Neural Parametric Leaf Models with Shape and Deformation Disentanglement},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YQFA5JIU}},
  note         = {Machine review of arXiv:2507.12714}
}
read the original abstract

We develop a neural parametric model for 3D leaves for plant modeling and reconstruction that are essential for agriculture and computer graphics. While neural parametric models are actively studied for humans and animals, plant leaves present unique challenges due to their diverse shapes and flexible deformation. To this problem, we introduce a neural parametric model for leaves, NeuraLeaf. Capitalizing on the fact that flattened leaf shapes can be approximated as a 2D plane, NeuraLeaf disentangles the leaves' geometry into their 2D base shapes and 3D deformations. This representation allows learning from rich sources of 2D leaf image datasets for the base shapes, and also has the advantage of simultaneously learning textures aligned with the geometry. To model the 3D deformation, we propose a novel skeleton-free skinning model and create a newly captured 3D leaf dataset called DeformLeaf. We show that NeuraLeaf successfully generates a wide range of leaf shapes with deformation, resulting in accurate model fitting to 3D observations like depth maps and point clouds. Our implementation and dataset are available at https://neuraleaf-yang.github.io/.

Figures

Figures reproduced from arXiv: 2507.12714 by the authors.

Figure 1
Figure 1. Our neural parametric model for leaves, NeuraLeaf, represents (a) [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of NeuraLeaf. We first learn the 2D base shape of leaves from a large-scale 2D dataset [9]. The base shape and texture spaces are modeled by a shape decoder fθs and texture generator fθt conditioned on shape & texture latent codes {zs, zt}. Based on the learned shape space, we then train two deformation decoders to predict skinning parameters, called skinning weight decoder fθw and transformation decoder fθ… view at source ↗
Figure 3
Figure 3. Deformation mapping in the zero-mean latent space. The colormap indicates the Chamfer distance between each de￾formed shape and its corresponding base shape. Samples closer to the center exhibit smaller deformations, while those farther from the center show greater deformations. the chamfer distance between predicted deformed points V and real points V¯ without any correspondence information. To ensure the smoothnes… view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Capturing pipeline for paired base and deformed shapes. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Visual comparisons of reconstructed leaf samples. We compare the ground truth with predicted shapes. Our method shows accurate reconstruction compared to baseline methods and has significantly more surface detail, such as leaf veins [PITH_FULL_IMAGE:figures/full_fig_p…
Figure 6
Figure 6. Figure 6: Visual comparisons of multiple leaf reconstruction. We use top-view RGB-D observations as input. Carefully captured multi￾view images are used to create the ground truths (denoted as “dense point cloud”). Our method achieves faithful reconstruction (e.g., vein shapes o…
Figure 7
Figure 7. Figure 7: Deformation transfer results. Each row in (a) shows different leaf species with the same deformation latent. (b) demon￾strates that each deformed leaf shape in a different color can be represented by a general surface, supporting the idea of a general￾ized deformation …

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

59 extracted references · 56 canonical work pages

  1. [1]

    SCAPE: Shape completion and animation of people

    Dragomir Anguelov, Praveen Srinivasan, Daphne Koller, Se- bastian Thrun, Jim Rodgers, and James Davis. SCAPE: Shape completion and animation of people. ACM Transac- tions on Graphics (TOG), 24(3):408–416, 2005. 2

  2. [2]

    SAL: Sign agnos- tic learning of shapes from raw data

    Matan Atzmon and Yaron Lipman. SAL: Sign agnos- tic learning of shapes from raw data. In Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2020. 10

  3. [3]

    Editable parametric dense foliage from 3D capture

    Paul Beardsley and Gaurav Chaurasia. Editable parametric dense foliage from 3D capture. In Proceedings of IEEE/CVF International Conference on Computer Vision (ICCV), 2017. 1, 2, 6

  4. [4]

    A morphable model for the synthesis of 3D faces

    V olker Blanz and Thomas Vetter. A morphable model for the synthesis of 3D faces. In Proceedings of ACM Annual Con- ference on Computer Graphics and Interactive Techniques (SIGGRAPH), 1999. 2

  5. [5]

    Image-based reconstruction and synthesis of dense foliage

    Derek Bradley, Derek Nowrouzezahrai, and Paul Beardsley. Image-based reconstruction and synthesis of dense foliage. ACM Transactions on Graphics (TOG) , 32(4):74:1–74:10,

  6. [6]

    Recovering non-rigid 3D shape from image streams

    Christoph Bregler, Aaron Hertzmann, and Henning Bier- mann. Recovering non-rigid 3D shape from image streams. In Proceedings of IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition (CVPR), 2000. 2

  7. [7]

    SNARF: Differentiable forward skin- ning for animating non-rigid neural implicit shapes

    Xu Chen, Yufeng Zheng, Michael J Black, Otmar Hilliges, and Andreas Geiger. SNARF: Differentiable forward skin- ning for animating non-rigid neural implicit shapes. In Pro- ceedings of IEEE/CVF International Conference on Com- puter Vision (ICCV), 2021. 1

  8. [8]

    Implicit functions in feature space for 3D shape reconstruc- tion and completion

    Julian Chibane, Thiemo Alldieck, and Gerard Pons-Moll. Implicit functions in feature space for 3D shape reconstruc- tion and completion. In Proceedings of IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition (CVPR) ,

Show all 59 references
  1. [9]

    A data repository of leaf images: Practice towards plant conservation with plant pathology

    Siddharth Singh Chouhan, Ajay Kaul, Uday Pratap Singh, and Sanjeev Jain. A data repository of leaf images: Practice towards plant conservation with plant pathology. InProceed- ings of International Conference on Information Systems and Computer Networks (ISCON), 2019. 3

  2. [10]

    Coherent point drift revisited for non-rigid shape match- ing and registration

    Aoxiang Fan, Jiayi Ma, Xin Tian, Xiaoguang Mei, and Wei Liu. Coherent point drift revisited for non-rigid shape match- ing and registration. In Proceedings of IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition (CVPR) ,

  3. [11]

    Learning neural parametric head models

    Simon Giebenhain, Tobias Kirschstein, Markos Georgopou- los, Martin R ¨unz, Lourdes Agapito, and Matthias Nießner. Learning neural parametric head models. In Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023. 1, 2

  4. [12]

    Gordon and Richard F

    William J. Gordon and Richard F. Riesenfeld. B-spline curves and surfaces. Computer Aided Geometric Design , 1974:95–126, 1974. 6, 7

  5. [13]

    Implicit geometric regularization for learn- ing shapes

    Amos Gropp, Lior Yariv, Niv Haim, Matan Atzmon, and Yaron Lipman. Implicit geometric regularization for learn- ing shapes. In Proceedings of International Conference on Machine Learning (ICML), 2020. 4

  6. [14]

    Shape modeling of plant leaves with un- structured meshes

    Sung Min Hong. Shape modeling of plant leaves with un- structured meshes. Master’s thesis, University of Waterloo,

  7. [15]

    SPAR3D: Stable point-aware reconstruction of 3D objects from single images

    Zixuan Huang, Mark Boss, Aaryaman Vasishta, James M Rehg, and Varun Jampani. SPAR3D: Stable point-aware reconstruction of 3D objects from single images. arXiv preprint arXiv:2501.04689, 2025. 13

  8. [16]

    Human3.6M: Large scale datasets and pre- dictive methods for 3D human sensing in natural environ- ments

    Catalin Ionescu, Dragos Papava, Vlad Olaru, and Cristian Sminchisescu. Human3.6M: Large scale datasets and pre- dictive methods for 3D human sensing in natural environ- ments. IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI), 36(7):1325–1339, 2014. 4

  9. [17]

    Skinning: Real-time shape deformation

    Alec Jacobson, Zhigang Deng, Ladislav Kavan, and JP Lewis. Skinning: Real-time shape deformation. In ACM SIGGRAPH 2014 Courses, 2014. 3, 4

  10. [18]

    LAESI: Leaf area estimation with synthetic im- agery

    Jacek Kału ˙zny, Yannik Schreckenberg, Karol Cyganik, Peter Annigh¨ofer, S¨oren Pirk, Dominik L Michels, Mikolaj Cies- lak, Farhah Assaad-Gerbert, Bedrich Benes, and Wojciech Pałubicki. LAESI: Leaf area estimation with synthetic im- agery. arXiv preprint arXiv:2404.00593, 2024. 1

  11. [19]

    Procedural modeling and visualization of multiple leaves

    Daeyeoul Kim and Jinmo Kim. Procedural modeling and visualization of multiple leaves. Multimedia Systems , 23: 435–449, 2017. 2

  12. [20]

    Segment any- thing

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer White- head, Alexander C Berg, Wan-Yen Lo, et al. Segment any- thing. In Proceedings of IEEE/CVF International Confer- ence on Computer Vision (ICCV), 2023. 3, 6, 9

  13. [21]

    Deep non-rigid structure from motion

    Chen Kong and Simon Lucey. Deep non-rigid structure from motion. In Proceedings of IEEE/CVF International Confer- ence on Computer Vision (ICCV), 2019. 2

  14. [22]

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

  15. [23]

    Matthew Loper, Naureen Mahmood, Javier Romero, Gerard Pons-Moll, and Michael J. Black. SMPL: A skinned multi- person linear model. ACM Transactions on Graphics (TOG), 34(6):248:1–248:16, 2015. 1, 2

  16. [24]

    Venation skeleton-based modeling plant leaf wilting

    Shenglian Lu, Chunjiang Zhao, and Xinyu Guo. Venation skeleton-based modeling plant leaf wilting. International Journal of Computer Games Technology , 2009:890917:1– 890917:8, 2009. 2

  17. [25]

    Qianli Ma, Jinlong Yang, Anurag Ranjan, Sergi Pujades, Gerard Pons-Moll, Siyu Tang, and Michael J. Black. Learn- ing to dress 3D people in generative clothing. InProceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2020. 2

  18. [26]

    Troje, Ger- ard Pons-Moll, and Michael J

    Naureen Mahmood, Nima Ghorbani, Nikolaus F. Troje, Ger- ard Pons-Moll, and Michael J. Black. AMASS: Archive of motion capture as surface shapes. In Proceedings of IEEE/CVF International Conference on Computer Vision (ICCV), 2019. 4

  19. [27]

    A framework for plant leaf modeling and shading.Math- ematical and Computer Modelling , 58(3-4):710–718, 2013

    Teng Miao, ChunJiang Zhao, XinYu Guo, and ShengLian Lu. A framework for plant leaf modeling and shading.Math- ematical and Computer Modelling , 58(3-4):710–718, 2013. 2

  20. [28]

    LEAP: Learning articulated occupancy of people

    Marko Mihajlovic, Yan Zhang, Michael J Black, and Siyu Tang. LEAP: Learning articulated occupancy of people. In Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021. 2

  21. [29]

    NeRF: Representing scenes as neural radiance fields for view syn- thesis

    Ben Mildenhall, Pratul P Srinivasan, Matthew Tancik, Jonathan T Barron, Ravi Ramamoorthi, and Ren Ng. NeRF: Representing scenes as neural radiance fields for view syn- thesis. In Proceedings of European Conference on Computer Vision (ECCV), 2020. 1

  22. [30]

    Morphological feature based maturity level identification of kalmegh and tulsi leaves

    Gunjan Mukherjee, Arpitam Chatterjee, and Bipan Tudu. Morphological feature based maturity level identification of kalmegh and tulsi leaves. In Proceedings of International Conference on Research in Computational Intelligence and Communication Networks (ICRCICN), 2017. 9

  23. [31]

    Instant neural graphics primitives with a mul- tiresolution hash encoding

    Thomas M ¨uller, Alex Evans, Christoph Schied, and Alexan- der Keller. Instant neural graphics primitives with a mul- tiresolution hash encoding. ACM Transactions on Graphics (TOG), 41(4):102:1–102:15, 2022. 1

  24. [32]

    Occupancy flow: 4D reconstruction by learning particle dynamics

    Michael Niemeyer, Lars Mescheder, Michael Oechsle, and Andreas Geiger. Occupancy flow: 4D reconstruction by learning particle dynamics. In Proceedings of IEEE/CVF In- ternational Conference on Computer Vision (ICCV) , 2019. 2

  25. [33]

    NPMs: Neural parametric models for 3D deformable shapes

    Pablo Palafox, Alja ˇz Boˇziˇc, Justus Thies, Matthias Nießner, and Angela Dai. NPMs: Neural parametric models for 3D deformable shapes. In Proceedings of IEEE/CVF Interna- tional Conference on Computer Vision (ICCV) , 2021. 1, 2, 4, 6, 7, 10, 11

  26. [34]

    SPAMs: Structured implicit parametric models

    Pablo Palafox, Nikolaos Sarafianos, Tony Tung, and Angela Dai. SPAMs: Structured implicit parametric models. In Pro- ceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022. 1, 2

  27. [35]

    DeepSDF: Learning continuous signed distance functions for shape representa- tion

    Jeong Joon Park, Peter Florence, Julian Straub, Richard Newcombe, and Steven Lovegrove. DeepSDF: Learning continuous signed distance functions for shape representa- tion. In Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2019. 2

  28. [36]

    A 3D face model for pose and illumination invariant face recognition

    Pascal Paysan, Reinhard Knothe, Brian Amberg, Sami Romdhani, and Thomas Vetter. A 3D face model for pose and illumination invariant face recognition. In Proceedings of IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS), 2009. 11

  29. [37]

    Ani- matable neural radiance fields for modeling dynamic human bodies

    Sida Peng, Junting Dong, Qianqian Wang, Shangzhan Zhang, Qing Shuai, Xiaowei Zhou, and Hujun Bao. Ani- matable neural radiance fields for modeling dynamic human bodies. In Proceedings of IEEE/CVF International Confer- ence on Computer Vision (ICCV), 2021. 1

  30. [38]

    Gerard Pons-Moll, Javier Romero, Mahmood Naureen, and Michael J. Black. Dyna: A model of dynamic human shape in motion. ACM Transactions on Graphics (TOG) , 34(4): 120:1–120:14, 2015. 2

  31. [39]

    On the spectral bias of neural networks

    Nasim Rahaman, Aristide Baratin, Devansh Arpit, Felix Draxler, Min Lin, Fred Hamprecht, Yoshua Bengio, and Aaron Courville. On the spectral bias of neural networks. In Proceedings of International Conference on Machine Learn- ing (ICML), 2019. 10

  32. [40]

    Javier Romero, Dimitrios Tzionas, and Michael J. Black. Embodied hands: Modeling and capturing hands and bod- ies together. ACM Transactions on Graphics (TOG), 36(6): 245:1–245:17, 2017. 2

  33. [41]

    Jump flooding in GPU with applications to voronoi diagram and distance transform

    Guodong Rong and Tiow-Seng Tan. Jump flooding in GPU with applications to voronoi diagram and distance transform. In Proceedings of Symposium on Interactive 3D Graphics and Games (I3D), 2006. 9

  34. [42]

    Modeling and visualization of leaf venation patterns

    Adam Runions, Martin Fuhrer, Brendan Lane, Pavol Federl, Anne-Ga ¨elle Rolland-Lagan, and Przemyslaw Prusinkiewicz. Modeling and visualization of leaf venation patterns. ACM Transactions on Graphics (TOG), 24(3):702– 711, 2005. 1, 3

  35. [43]

    SCANimate: Weakly supervised learning of skinned clothed avatar networks

    Shunsuke Saito, Jinlong Yang, Qianli Ma, and Michael J Black. SCANimate: Weakly supervised learning of skinned clothed avatar networks. In Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021. 10

  36. [44]

    A database of leaf images: Practice towards plant conservation with plant pathology

    Uday Pratap Singh Siddharth Singh Chouhan, Ajay Kaul and Sanjeev Jain. A database of leaf images: Practice towards plant conservation with plant pathology. In Proceedings of International Conference on Information Systems and Com- puter Networks (ISCON), 2019. 1

  37. [45]

    Sim- ulation of morphology changes in drying leaves

    Jeong Sohyeon, Si-Hyung Park, and Chang-Hun Kim. Sim- ulation of morphology changes in drying leaves. Computer Graphics Forum, 32(1):204–215, 2013. 3

  38. [46]

    As-rigid-as-possible surface modeling

    Olga Sorkine and Marc Alexa. As-rigid-as-possible surface modeling. In Proceedings of Eurographics Symposium on Geometry Processing (SGP), 2007. 2, 6

  39. [47]

    SIZER: A dataset and model for parsing 3D clothing and learning size sensitive 3D clothing

    Garvita Tiwari, Bharat Lal Bhatnagar, Tony Tung, and Ger- ard Pons-Moll. SIZER: A dataset and model for parsing 3D clothing and learning size sensitive 3D clothing. InProceed- ings of European Conference on Computer Vision (ECCV) ,

  40. [48]

    Im- plicit mesh reconstruction from unannotated image collec- tions

    Shubham Tulsiani, Nilesh Kulkarni, and Abhinav Gupta. Im- plicit mesh reconstruction from unannotated image collec- tions. arXiv preprint arXiv:2007.08504, 2020. 2

  41. [49]

    FaceVerse: A fine-grained and detail- controllable 3D face morphable model from a hybrid dataset

    Lizhen Wang, Zhiyuan Chen, Tao Yu, Chenguang Ma, Liang Li, and Yebin Liu. FaceVerse: A fine-grained and detail- controllable 3D face morphable model from a hybrid dataset. In Proceedings of IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition (CVPR), 2022. 2

  42. [50]

    NeuS: Learning neural im- plicit surfaces by volume rendering for multi-view recon- struction

    Peng Wang, Lingjie Liu, Yuan Liu, Christian Theobalt, Taku Komura, and Wenping Wang. NeuS: Learning neural im- plicit surfaces by volume rendering for multi-view recon- struction. In Advances in Neural Information Processing Systems (NeurIPS), 2021. 1

  43. [51]

    A leaf modeling and multi-scale remeshing method for visual computation via hierarchical parametric vein and margin rep- resentation

    Weiliang Wen, Baojun Li, Bao-jun Li, and Xinyu Guo. A leaf modeling and multi-scale remeshing method for visual computation via hierarchical parametric vein and margin rep- resentation. Frontiers in Plant Science, 9(783), 2018. 1, 3

  44. [52]

    MagicPony: Learning articu- lated 3D animals in the wild

    Shangzhe Wu, Ruining Li, Tomas Jakab, Christian Rup- precht, and Andrea Vedaldi. MagicPony: Learning articu- lated 3D animals in the wild. In Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023. 2

  45. [53]

    CASA: Category-agnostic skeletal ani- mal reconstruction

    Yuefan Wu, Zeyuan Chen, Shaowei Liu, Zhongzheng Ren, and Shenlong Wang. CASA: Category-agnostic skeletal ani- mal reconstruction. In Advances in Neural Information Pro- cessing Systems (NeurIPS), 2022. 2

  46. [54]

    RigNet: Neural rigging for articu- lated characters

    Zhan Xu, Yang Zhou, Evangelos Kalogerakis, Chris Lan- dreth, and Karan Singh. RigNet: Neural rigging for articu- lated characters. ACM Transactions on Graphics (TOG), 39 (4):58:1–58:14, 2020. 4, 10

  47. [55]

    Hi- lassie: High-fidelity articulated shape and skeleton discovery from sparse image ensemble

    Chun-Han Yao, Wei-Chih Hung, Yuanzhen Li, Michael Rubinstein, Ming-Hsuan Yang, and Varun Jampani. Hi- lassie: High-fidelity articulated shape and skeleton discovery from sparse image ensemble. In Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CV...

  48. [56]

    Multiview neu- ral surface reconstruction by disentangling geometry and ap- pearance

    Lior Yariv, Yoni Kasten, Dror Moran, Meirav Galun, Matan Atzmon, Basri Ronen, and Yaron Lipman. Multiview neu- ral surface reconstruction by disentangling geometry and ap- pearance. In Advances in Neural Information Processing Systems (NeurIPS), 2020. 10

  49. [57]

    Hunyuan3D 2.0: Scaling dif- fusion models for high resolution textured 3D assets genera- tion

    Zibo Zhao, Zeqiang Lai, Qingxiang Lin, Yunfei Zhao, Haolin Liu, Shuhui Yang, Yifei Feng, Mingxin Yang, Sheng Zhang, Xianghui Yang, et al. Hunyuan3D 2.0: Scaling dif- fusion models for high resolution textured 3D assets genera- tion. arXiv preprint arXiv:2501.12202, 2025. 13

  50. [58]

    Unpaired image-to-image translation using cycle- consistent adversarial networks

    Jun-Yan Zhu, Taesung Park, Phillip Isola, and Alexei A Efros. Unpaired image-to-image translation using cycle- consistent adversarial networks. In Proceedings of IEEE/CVF International Conference on Computer Vision (ICCV), 2017. 4, 10

  51. [59]

    Silvia Zuffi, Angjoo Kanazawa, David Jacobs, and Michael J. Black. 3D menagerie: Modeling the 3D shape and pose of animals. In Proceedings of IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition (CVPR) ,

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.