Pith. sign in

REVIEW 5 major objections 5 minor 1 cited by

Vector Representations of Vessel Trees

T0 review · 5 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read VeTTA encodes a full vascular tree into a single vector and decodes it back into a geometrically accurate, topologically valid tree.

desk verdict Real architectural idea, but Table 2 overclaims: VeTTA's advantage is confounded by giving it centerlines while voxel baselines see only blurry grids. read the letter →

arxiv 2506.11163 v1 pith:FFT6O2CT submitted 2025-06-11 eess.IV cs.CVcs.GR

James Batten , Michiel Schaap , Matthew Sinclair , Ying Bai , Ben Glocker This is my paper · ORCID
classification eess.IVcs.CVcs.GR
keywords treeautoencodersrecursivedecodingcoronaryarterymodelingvesseltreesvascularnetworkstransformerautoencodertopologypreservationlatentvectorrepresentation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper claims that an entire vascular tree—its branching pattern together with the continuous geometry of every vessel segment—can be compressed into a single vector and rebuilt faithfully from that vector. VeTTA trains two Transformer autoencoders in sequence: the first learns a compact embedding of individual vessel curves with their radii, and the second pools the tree's edges into one global latent code. A recursive decoder rebuilds the tree node by node, and because every step attaches new child nodes to an already reconstructed node, the output is guaranteed to be a valid tree rather than a disconnected tangle. On a synthetic 2D tree dataset and a 3D coronary artery dataset, the paper reports more accurate reconstructions than voxel-based convolutional autoencoders, substantially lower GPU memory use, and plausible trees from latent-space interpolation. If correct, this gives medical imaging a compact, topologically reliable representation of vessel anatomy for downstream tasks such as image-to-geometry prediction and generative modeling.

What carries the argument

The mechanism that carries the argument is recursive decoding conditioned on a global tree code. The latent code $z_t$, produced by pooling transformer-encoded edge features, is concatenated into the edge encodings of the partial tree at each decoding step, so the decoder branch always sees both the full-tree context and the current reconstruction state. A transformer decoder with 32 learned slot vectors outputs per-slot predictions of position, topology, and (in 3D) vessel embedding, log radius, and a skip-vessel flag; agglomerative clustering of these slots in the sinusoidally lifted domain groups several slots into each child node, with the number of children fixed by the query node's predicted topology. Because every appended child connects to an existing queried node, the output cannot contain cycles, which is the paper's topological guarantee. The per-segment geometry itself is carried through the first-stage Vessel Autoencoder and decoded by a small MLP residual model $g_d(z_v, t)$ that refines a linear interpolation between endpoint positions and radii, with an evaluation-time windowing function $m(t)$ that forces exact endpoint matching.

What would settle it

Re-run the VesselTrees evaluation with matched inputs: train the convolutional baselines on the same explicit centerline-and-radius data that VeTTA receives, or feed VeTTA only its own 128³ voxelized reconstruction and extract centerlines from it, then compare Dice and centerline-distance scores on the identical test trees. If the Table 2 gaps vanish under either symmetry, the central claim reduces to a statement about input representation rather than about the autoencoder itself.

Watch

Extended reading notes

Core claim

The central claim is that tree-structured geometry, exemplified by coronary artery networks, admits a single-vector representation from which the original branching structure and continuous vessel geometry can be recovered. VeTTA realizes this in two stages: a Vessel Autoencoder encodes a single vessel segment, represented as 64 samples of position, radius, and arc-length parameter lifted into sinusoidal Fourier features, into a 64-dimensional embedding $z_v$; a Vessel Tree Autoencoder then takes the set of directed edges of the whole tree—each carrying Fourier-lifted endpoint positions, one-hot topology vectors, and (in 3D) the vessel embedding—and pools their transformer encodings into a global code $z_t$. The decoder is the key move: at every recursion step it encodes the partially reconstructed tree with a query flag marking the node to expand, concatenates $z_t$ as context, and uses a transformer decoder over 32 learned slots to predict that node's children, clustering slot predictions in the lifted Fourier domain to form discrete child nodes. A learned start token and a self-loop 'semi-edge' bootstrap the process from an empty partial tree, and the loss is a two-directional matching (right-hand matching plus top-k matching) between slots and ground-truth targets. The authors' claim is that this recursive scheme converts the latent code into a 'program' for rebuilding the tree, giving a structural guarantee of valid tree topology and reconstructions that, in their experiments on a 2D synthetic dataset and a 3D coronary dataset, exceed convolutional autoencoder baselines in Dice score, Hausdorff and centerline distances, and latent interpolation quality.

Load-bearing premise

The load-bearing premise is that the comparison is fair: VeTTA is given the ground-truth centerlines and radii of each vessel as input, while the convolutional baselines receive only voxel grids, so the reported reconstruction advantage could shrink or disappear if every method had to start from the same input representation.

Editorial extensions

If this is right

  • Whole coronary trees become fixed-length vectors, so storing, transmitting, and comparing anatomy no longer requires voxel grids or meshes, and discretization artifacts disappear by construction.
  • Because the two stages are trained sequentially on large datasets of vessels, the resulting embeddings can be reused as priors for downstream tasks such as predicting vessel geometry directly from cardiac CT images.
  • Interpolating between latent codes yields intermediate trees that remain valid branching structures, opening a generative route to synthetic-but-plausible coronary anatomies for data augmentation and population shape analysis.
  • The structural guarantee of tree validity means downstream analyses, such as hemodynamic assessment, start from a connected branching geometry rather than a skeletonized voxel segmentation that may contain spurious branches or gaps.
  • GPU memory requirements drop substantially relative to 3D convolutional autoencoders, which the authors report makes large-batch training on whole trees feasible on a single 32 GB GPU.

Reading between the lines

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

  • The paper acknowledges an input asymmetry as unfair only for the VesselVAE comparison (Appendix D.4), not for the main voxel-CNN comparison; a matched-input experiment would separate the architecture's contribution from the richer centerline input VeTTA receives.
  • The authors defer ablation studies to future work, so which components—Fourier lifting, slot clustering, the recursive conditioning, or merely the richer input—drive the reconstruction gains is not yet established.
  • Because the mechanism assumes only a rooted tree with bounded branching (here at most two children per node), the same two-stage scheme should transfer to other branching anatomies such as the lung airways or hepatic and renal trees, and to any rooted tree-shaped data.
  • The two-stage split suggests a generative upgrade: place a richer prior over the tree code $z_t$ and sample whole trees whose branch-length and radius statistics could then be validated against the empirical distribution of coronary anatomies.
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

5 major / 5 minor

Summary. The paper proposes VeTTA, a two-stage transformer autoencoder that encodes a vascular tree into a single vector. A vessel autoencoder first maps each vessel segment into an embedding; a tree autoencoder then encodes the whole tree connectivity and geometry into a latent code. A recursive slot-based decoder reconstructs the tree one node at a time, guaranteeing a tree-structured output by construction. Experiments on the public 2D SSA synthetic dataset and a private 3D coronary dataset compare VeTTA with convolutional autoencoders on reconstruction metrics (Dice, Hausdorff, average surface/centerline/point-cloud distances) and show latent interpolations.

Significance. If substantiated, the compact and topology-consistent tree representation could be useful for downstream image-to-geometry tasks and generative modeling of vasculature. The recursive decoder with fixed slots is a sensible design, and the use of continuous centerline geometry is a distinctive contribution. However, the evaluation as presented does not support the headline claim of superior reconstruction fidelity: the comparison against voxel baselines is confounded by input representation, no uncertainty quantification is provided, topology correctness is not measured, and key design choices are not ablated. These gaps are addressable within the scope of a revision.

major comments (5)
  1. [§4, Tables 1 and 2; §5] The claimed 'superior reconstruction fidelity' is not supported by the current comparison because VeTTA and the convolutional baselines receive different input representations. VeTTA is given the ground-truth tree (centerline coordinates, radii, connectivity, and in 3D the pre-trained vessel embeddings), while the baselines receive only 128^3 voxel grids from which centerlines are extracted by skeletonization (Section 4). The improved Dice, ACD, and related metrics in Tables 1 and 2 could therefore stem from the richer input rather than from the transformer autoencoder. Please add a controlled comparison—for example, a variant of VeTTA that takes voxel input, or voxel baselines that are given the same centerline/radius input—or re-frame the conclusion to state that, given an explicit tree representation, VeTTA reconstructs it more accurately than voxel autoencoders reconstruct segmentations.
  2. [§5, Tables 1 and 2] No error bars, confidence intervals, or statistical significance tests are reported for any reconstruction metric. With a single training run per configuration, the differences between VeTTA and the strongly performing Conv-3D-AE baseline (e.g., Mesh HD 2.09 vs 2.22) may not be reliable. Please report the mean and standard deviation across multiple training runs or bootstrap resampling of the test set, and test the main differences.
  3. [§5, abstract] The abstract and Section 5 claim 'accurate topology preservation' and 'topologically correct' reconstructions, but no quantitative topology metric is presented. The recursive decoder guarantees that the output is some tree, but it does not guarantee that the branching pattern matches the ground truth. Please add a quantitative topology evaluation (e.g., branch count error, tree edit distance, or bifurcation match rate) to support the claim.
  4. [§6, Appendix B] The method depends on many un-ablated design choices: the number of slots (M=32), the matching parameter k=3, the loss reweighting α and position/radius weights, the curvature segmentation sensitivity, the skip-vessel threshold, the evaluation mask m(t), and the Fourier octave choice. The conclusion states that ablation studies are left to future work, which means the current paper does not establish which components drive the reported results. Please include at least a limited set of ablations or sensitivity analyses on the most consequential hyperparameters (slots, k, KL weight, and input feature composition).
  5. [§1, §6] The abstract and introduction state that VeTTA 'substantially lowers GPU memory requirements' compared to 3D convolutional models, but no measured memory usage or computational cost is reported anywhere. Since this is one of the two headline advantages claimed, please provide quantitative memory measurements (e.g., peak GPU memory during training/inference for VeTTA and the baselines) or soften the claim to a qualitative statement about architectural capacity.
minor comments (5)
  1. [§4] The SSA dataset is referenced only by URL without a formal citation; please provide a bibliographic reference or versioned DOI.
  2. [§5] The caveat that VesselVAE is a generative method and that comparing it on reconstruction metrics is 'somewhat unfair' appears only in Appendix D.4; it should also be stated in the main text where Table 2 is discussed.
  3. [§3] The evaluation mask m(t) in Eq. (1) is introduced without explanation in the main text; please describe its role when it is first used.
  4. [Appendix A] The equation numbering in Appendix A appears to be inconsistent with the main text (e.g., Eqs. (5)-(8) are locally defined); please renumber and unify cross-references.
  5. [§4] The VesselTrees dataset is private and its annotation protocol is not described; please add details on annotation quality, inclusion criteria, and availability, or state that it cannot be released.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: VeTTA's reconstruction claims are empirical and self-contained, and the only 'by construction' element is the recursive decoder's tree-structured output guarantee, which the paper explicitly identifies rather than disguises as a learned result.

full rationale

The paper's central derivation is an empirical autoencoder pipeline. The Vessel Autoencoder learns embeddings from sampled points along each curve, and the Tree Autoencoder learns a global code from edge features; the reconstruction losses (Eqs. 2 and 4) are standard fitted objectives, not quantities that reduce to their inputs by definition. The 'topological guarantee' (Section 1) is explicitly a property of the recursive decoder design, and the paper does not claim that tree-structuredness is learned from data. The comparison against voxel baselines is asymmetric (VeTTA receives ground-truth centerlines and radii while baselines receive voxel grids), but that is a correctness or experimental-design concern about the headline claim of 'superior reconstruction fidelity,' not a circularity: no equation or fitted parameter is renamed as a prediction. The only overlapping-author citation, Rasal et al. 2023 (co-authored by Glocker), is used as contextual motivation and is not load-bearing. Therefore no circular step can be exhibited with a quote and a specific reduction.

Assumptions & free parameters 12 free parameters · 6 assumptions · 0 invented entities

The method relies on a number of explicit domain assumptions and hand-chosen hyperparameters. The largest burden is the proprietary, semi-automatically annotated dataset, whose quality is taken as ground truth without analysis. The architectural guarantees (valid tree output) are built in by construction, which lowers the risk of invalid outputs but also means 'topology preservation' is partly a design constraint rather than a learned property.

free parameters (12)
  • Latent code size z_t (3D) = 8192
    Manual choice for the tree-level representation; the paper uses a single large bottleneck for all 3D models, which directly affects reconstruction capacity.
  • Latent code size z_t (2D) = 256
    Manual choice on SSA dataset; same for all compared models.
  • Vessel embedding dimension z_v = 64
    Manual choice for per-segment embeddings; used by the vessel autoencoder and concatenated into tree edges.
  • KL weight λ_KL = 1e-6
    Manual weighting of the KL term in variational variants; drastically low value makes the prior nearly uninformative.
  • Number of slots M_slots = 32
    Manual choice for the set-prediction decoder; bounds the number of child nodes that can be predicted at each recursive step.
  • Matching parameter k = 3
    Manual choice in Top-K matching loss; requires each target to be matched to at least k slots.
  • Loss reweighting α = 0.3
    Manual coefficient in the vessel loss reweighting term α·da/√de, designed to compensate for longer/tortuous vessels.
  • Position/radius loss weights = 1.0 and 0.01
    Manual separation of MSE terms for coordinates vs radius in vessel autoencoder loss.
  • Evaluation mask m(t) coefficients = 0.5, 0.2, 0.1
    Manually chosen function m(t)=0.5−0.2·t^0.1(1−t)^0.1 to enforce endpoint matching at evaluation.
  • Curvature sensitivity = 0.75
    Manual parameter in the segment-computation algorithm that controls vessel segmentation into 64 parts.
  • Fourier feature octaves = [1,2,4,8,16,32]
    Manual choice of frequencies for sinusoidal lifting of coordinates.
  • Skip-vessel length threshold = 2.5mm
    Manual threshold for deciding which short vessel segments are represented by a flag instead of a full vessel embedding.
assumptions (6)
  • domain assumption Vessels are represented as generalized cylinders parameterized by centerline coordinates (x,y,z) and radius r.
    Section 3: 'Our representation models vessels as generalized cylinders parameterized by coordinates and radius, sufficient for accurate mesh reconstruction.' This reduces the problem to centerline plus radius prediction and ignores non-cylindrical features like vessel wall thickness or plaque.
  • domain assumption Each node in the tree has at most two children (binary trees).
    Section 4: 'we limit each node in the tree connectivity structure to having either 0, 1, or 2 children.' This is enforced on both datasets; trifurcations are treated as two close bifurcations.
  • domain assumption The VesselTrees dataset is an accurate representation of coronary anatomy.
    Section 4: data is 'semi-automatically annotated 3D coronary meshes (and centerlines) derived from cardiac CT angiography through HeartFlow's commercial processing pipeline.' No annotation error analysis is provided.
  • standard math Transformer self-attention with sinusoidal Fourier features can represent the mapping from tree edges to latent codes and back.
    The method uses standard transformer architectures (Vaswani, 2017) and Fourier features (Tancik et al., 2020); these are established but unproven for this specific task.
  • ad hoc to paper The recursive decoding procedure with slots and clustering yields a valid tree for any predicted topology.
    Appendix B.2: the decoder always appends children to the query node and the cluster count equals the predicted number of children, so output is a tree by construction.
  • ad hoc to paper The loss matching algorithm (Top-K matching) is a correct surrogate for reconstruction quality.
    Appendix E Algorithms 1-3 define a custom bipartite matching loss; the paper does not prove it is unbiased or that it encourages globally consistent tree reconstruction.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Vector Representations of Vessel Trees." pith.science (2026). https://pith.science/paper/FFT6O2CT

@misc{pith2026250611163,
  author       = {Pith},
  title        = {Pith review of: Vector Representations of Vessel Trees},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FFT6O2CT}},
  note         = {Machine review of arXiv:2506.11163}
}
read the original abstract

We introduce a novel framework for learning vector representations of tree-structured geometric data focusing on 3D vascular networks. Our approach employs two sequentially trained Transformer-based autoencoders. In the first stage, the Vessel Autoencoder captures continuous geometric details of individual vessel segments by learning embeddings from sampled points along each curve. In the second stage, the Vessel Tree Autoencoder encodes the topology of the vascular network as a single vector representation, leveraging the segment-level embeddings from the first model. A recursive decoding process ensures that the reconstructed topology is a valid tree structure. Compared to 3D convolutional models, this proposed approach substantially lowers GPU memory requirements, facilitating large-scale training. Experimental results on a 2D synthetic tree dataset and a 3D coronary artery dataset demonstrate superior reconstruction fidelity, accurate topology preservation, and realistic interpolations in latent space. Our scalable framework, named VeTTA, offers precise, flexible, and topologically consistent modeling of anatomical tree structures in medical imaging.

Figures

Figures reproduced from arXiv: 2506.11163 by the authors.

Figure 1
Figure 1. Proposed method diagram. First stage (left): the Vessel Autoencoder which [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Example reconstruction with the proposed model. Left: ground truth mesh from [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. (First row) Interpolations produced by the proposed 3D vessel tree autoencoder [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Diagram of the Vessel Tree autoencoder model. The encoder branch outputs [PITH_FULL_IMAGE:figures/full_fig_p015_4.png]
Figure 5
Figure 5. Figure 5: Diagram of the Single Edge Encoder in the 2D case. On the 3D VesselTrees [PITH_FULL_IMAGE:figures/full_fig_p016_5.png]
Figure 6
Figure 6. Figure 6: Diagram of the Edges Encoder Pooling Head. We include a pooling head at the end of the encoder pathway that first applies a two-layer MLP to map the edge encodings {eˆi} to the target embedding size zdim before taking the mean of the resulting representations over the …
Figure 7
Figure 7. Figure 7: Diagram of the Encoder Branch. The Edges Encoder in this branch is the Full [PITH_FULL_IMAGE:figures/full_fig_p018_7.png]
Figure 8
Figure 8. Figure 8: Diagram of the Decoder Branch. The Edges Encoder in this branch is the Partial [PITH_FULL_IMAGE:figures/full_fig_p018_8.png]
Figure 9
Figure 9. Figure 9: (First row) Interpolations produced using the baseline model Conv-2D-VAE (GN). [PITH_FULL_IMAGE:figures/full_fig_p022_9.png]
Figure 10
Figure 10. Figure 10: (Left) Original mesh of the preprocessed sample passed as input to the VesselVAE [PITH_FULL_IMAGE:figures/full_fig_p025_10.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. VesselTok: Tokenizing Vessel-like 3D Biomedical Graph Representations for Reconstruction and Generation

    cs.CV 2026-03 conditional novelty 6.0 of 10

    VesselTok learns compact continuous tokens of large tubular biomedical graphs from centerline points plus a fixed pseudo-radius, enabling reconstruction, generation, and link prediction across anatomies.

Reference graph

Works this paper leans on

37 extracted references · 10 canonical work pages · cited by 1 Pith paper

  1. [1]

    Seeing implicit neural representations as fourier series

    Nuri Benbarka, Timon H \"o fer, Andreas Zell, et al. Seeing implicit neural representations as fourier series. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 2041--2050, 2022

    work page 2022

  2. [2]

    Andrew Brock, Theodore Lim, J. M. Ritchie, and Nick Weston. Generative and discriminative voxel modeling with convolutional neural networks, 2016. URL https://arxiv.org/abs/1608.04236

    arXiv 2016

  3. [3]

    Traj-MAE: Masked Autoencoders for Trajectory Prediction

    Hao Chen, Jiaze Wang, Kun Shao, Furui Liu, Jianye Hao, Chenyong Guan, Guangyong Chen, and Pheng-Ann Heng. Traj-MAE : Masked autoencoders for trajectory prediction, 2023. URL https://arxiv.org/abs/2303.06697

    work page Pith review arXiv 2023

  4. [4]

    TrajVAE : A variational autoencoder model for trajectory generation

    Xinyu Chen, Jiajie Xu, Rui Zhou, Wei Chen, Junhua Fang, and Chengfei Liu. TrajVAE : A variational autoencoder model for trajectory generation. Neurocomputing, 428: 0 332--339, March 2021. doi:10.1016/j.neucom.2020.03.120

    doi:10.1016/j.neucom.2020.03.120 2021

  5. [5]

    Co-Reyes, YuXuan Liu, Abhishek Gupta, Benjamin Eysenbach, Pieter Abbeel, and Sergey Levine

    John D. Co-Reyes, YuXuan Liu, Abhishek Gupta, Benjamin Eysenbach, Pieter Abbeel, and Sergey Levine. Self-consistent trajectory autoencoder: Hierarchical reinforcement learning with trajectory embeddings, 2018. URL https://arxiv.org/abs/1806.02813

    arXiv 2018

  6. [6]

    MolGAN : An implicit generative model for small molecular graphs, 2018

    Nicola De Cao and Thomas Kipf. MolGAN : An implicit generative model for small molecular graphs, 2018. URL https://arxiv.org/abs/1805.11973

    arXiv 2018

  7. [7]

    Bert: Pre-training of deep bidirectional transformers for language understanding

    Jacob Devlin. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805, 2018

    arXiv 2018

  8. [8]

    VesselVAE: Recursive Variational Autoencoders for 3D Blood Vessel Synthesis

    Paula Feldman, Miguel Fainstein, Viviana Siless, Claudio Delrieux, and Emmanuel Iarussi. VesselVAE : Recursive variational autoencoders for 3d blood vessel synthesis, 2023. URL https://arxiv.org/abs/2307.03592

    work page Pith review arXiv 2023

Show all 37 references
  1. [9]

    Local deep implicit functions for 3d shape, 2019

    Kyle Genova, Forrester Cole, Avneesh Sud, Aaron Sarna, and Thomas Funkhouser. Local deep implicit functions for 3d shape, 2019. URL https://arxiv.org/abs/1912.06126

    2019 arXiv

  2. [10]

    Wei, David Duvenaud, Jos \'e Miguel Hern \'a ndez-Lobato, Benjam \'i n S \'a nchez-Lengeling, Dennis Sheberla, Jorge Aguilera-Iparraguirre, Timothy D

    Rafael G \'o mez-Bombarelli, Jennifer N. Wei, David Duvenaud, Jos \'e Miguel Hern \'a ndez-Lobato, Benjam \'i n S \'a nchez-Lengeling, Dennis Sheberla, Jorge Aguilera-Iparraguirre, Timothy D. Hirzel, Ryan P. Adams, and Al \'a n Aspuru-Guzik. Automatic chemical design using a d...

    2018 doi

  3. [11]

    Gaussian error linear units (gelus)

    Dan Hendrycks and Kevin Gimpel. Gaussian error linear units (gelus). arXiv preprint arXiv:1606.08415, 2016

    2016 arXiv

  4. [12]

    Batch normalization: Accelerating deep network training by reducing internal covariate shift

    Sergey Ioffe. Batch normalization: Accelerating deep network training by reducing internal covariate shift. arXiv preprint arXiv:1502.03167, 2015

    2015 arXiv

  5. [13]

    Dominant motion identification of multi-particle system using deep learning from video, 2021

    Yayati Jadhav and Amir Barati Farimani. Dominant motion identification of multi-particle system using deep learning from video, 2021. URL https://arxiv.org/abs/2104.12722

    2021 arXiv

  6. [14]

    Junction tree variational autoencoder for molecular graph generation, 2018

    Wengong Jin, Regina Barzilay, and Tommi Jaakkola. Junction tree variational autoencoder for molecular graph generation, 2018. URL https://arxiv.org/abs/1802.04364

    2018 arXiv

  7. [15]

    Poisson surface reconstruction

    Michael Kazhdan, Matthew Bolitho, and Hugues Hoppe. Poisson surface reconstruction. In Proceedings of the fourth Eurographics symposium on Geometry processing, volume 7, 2006

    2006

  8. [16]

    Representation learning of 3d meshes using an autoencoder in the spectral domain

    Cl \'e ment Lemeunier, Florence Denis, Guillaume Lavou \'e , and Florent Dupont. Representation learning of 3d meshes using an autoencoder in the spectral domain. Computers & Graphics, 107: 0 131--143, October 2022. doi:10.1016/j.cag.2022.07.011

    2022 doi

  9. [17]

    Meshdiffusion: Score-based generative 3d mesh modeling

    Zhen Liu, Yao Feng, Michael J Black, Derek Nowrouzezahrai, Liam Paull, and Weiyang Liu. Meshdiffusion: Score-based generative 3d mesh modeling. arXiv preprint arXiv:2303.08133, 2023

    2023 arXiv

  10. [18]

    Object-centric learning with slot attention

    Francesco Locatello, Dirk Weissenborn, Thomas Unterthiner, Aravindh Mahendran, Georg Heigold, Jakob Uszkoreit, Alexey Dosovitskiy, and Thomas Kipf. Object-centric learning with slot attention. Advances in neural information processing systems, 33: 0 11525--11538, 2020

    2020

  11. [19]

    Decoupled weight decay regularization

    I Loshchilov. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101, 2017

    2017 arXiv

  12. [20]

    GraphiT : Encoding graph structure in transformers, 2021

    Gr \'e goire Mialon, Dexiong Chen, Margot Selosse, and Julien Mairal. GraphiT : Encoding graph structure in transformers, 2021. URL https://arxiv.org/abs/2106.05667

    2021 arXiv

  13. [21]

    Deep trajectory clustering with autoencoders

    Xavier Olive, Luis Basora, Beno\^ i t Viry, and Richard Alligier. Deep trajectory clustering with autoencoders. In Proceedings of SESAR Innovation Days, Aug 2020. URL https://enac.hal.science/hal-02916241/document

    2020

  14. [22]

    Pytorch: An imperative style, high-performance deep learning library

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. Pytorch: An imperative style, high-performance deep learning library. Advances in neural information processing systems, 32, 2019

    2019

  15. [23]

    Paetzold, Hongwei Bran Li, and Bjoern Menze

    Chinmay Prabhakar, Suprosanna Shit, Fabio Musio, Kaiyuan Yang, Tamaz Amiranashvili, Johannes C. Paetzold, Hongwei Bran Li, and Bjoern Menze. 3D Vessel Graph Generation Using Denoising Diffusion, pages 3--13. Springer Nature Switzerland, 2024. ISBN 9783031721205. doi:10.1007/97...

    2024 doi

  16. [24]

    Improving language understanding by generative pre-training

    Alec Radford. Improving language understanding by generative pre-training. 2018

    2018

  17. [25]

    Castro, Nick Pawlowski, and Ben Glocker

    Rajat Rasal, Daniel C. Castro, Nick Pawlowski, and Ben Glocker. Deep structural causal shape models. In Leonid Karlinsky, Tomer Michaeli, and Ko Nishino, editors, Computer Vision -- ECCV 2022 Workshops, pages 400--432, Cham, 2023. Springer Nature Switzerland. ISBN 978-3-031-25075-0

    2022

  18. [26]

    Tissue metabolism driven arterial tree generation

    Matthias Schneider, Johannes Reichold, Bruno Weber, G \'a bor Szekely, and Sven Hirsch. Tissue metabolism driven arterial tree generation. Medical Image Analysis, 16 0 (7): 0 1397--1414, 2012. ISSN 1361-8415. doi:https://doi.org/10.1016/j.media.2012.04.009. URL https://www.sci...

    2012 doi

  19. [27]

    GraphAF : a flow-based autoregressive model for molecular graph generation, 2020

    Chence Shi, Minkai Xu, Zhaocheng Zhu, Weinan Zhang, Ming Zhang, and Jian Tang. GraphAF : a flow-based autoregressive model for molecular graph generation, 2020. URL https://arxiv.org/abs/2001.09382

    2020 arXiv

  20. [28]

    MeshGPT : Generating triangle meshes with decoder-only transformers, 2023

    Yawar Siddiqui, Antonio Alliegro, Alexey Artemov, Tatiana Tommasi, Daniele Sirigatti, Vladislav Rosov, Angela Dai, and Matthias Nie ner. MeshGPT : Generating triangle meshes with decoder-only transformers, 2023. URL https://arxiv.org/abs/2311.15475

    2023 arXiv

  21. [29]

    GraphVAE : Towards Generation of Small Graphs Using Variational Autoencoders , pages 412--422

    Martin Simonovsky and Nikos Komodakis. GraphVAE : Towards Generation of Small Graphs Using Variational Autoencoders , pages 412--422. Springer, 2018. doi:10.1007/978-3-030-01418-6_41

    2018 doi

  22. [30]

    Implicit neural representations with periodic activation functions

    Vincent Sitzmann, Julien Martel, Alexander Bergman, David Lindell, and Gordon Wetzstein. Implicit neural representations with periodic activation functions. Advances in neural information processing systems, 33: 0 7462--7473, 2020

    2020

  23. [31]

    Fourier features let networks learn high frequency functions in low dimensional domains

    Matthew Tancik, Pratul Srinivasan, Ben Mildenhall, Sara Fridovich-Keil, Nithin Raghavan, Utkarsh Singhal, Ravi Ramamoorthi, Jonathan Barron, and Ren Ng. Fourier features let networks learn high frequency functions in low dimensional domains. Advances in neural information proc...

    2020

  24. [32]

    Attention is all you need

    A Vaswani. Attention is all you need. Advances in Neural Information Processing Systems, 2017

    2017

  25. [33]

    Graph attention networks

    Petar Veli c kovi \'c , Guillem Cucurull, Arantxa Casanova, Adriana Romero, Pietro Lio, and Yoshua Bengio. Graph attention networks. arXiv preprint arXiv:1710.10903, 2017

    2017 arXiv

  26. [34]

    Wolterink, Tim Leiner, and Ivana Isgum

    Jelmer M. Wolterink, Tim Leiner, and Ivana Isgum. Blood vessel geometry synthesis using generative adversarial networks, 2018. URL https://arxiv.org/abs/1804.04381

    2018 arXiv

  27. [35]

    Group normalization

    Yuxin Wu and Kaiming He. Group normalization. In Proceedings of the European conference on computer vision (ECCV), pages 3--19, 2018

    2018

  28. [36]

    Intra: 3d intracranial aneurysm dataset for deep learning

    Xi Yang, Ding Xia, Taichi Kin, and Takeo Igarashi. Intra: 3d intracranial aneurysm dataset for deep learning. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2020

    2020

  29. [37]

    A survey on deep graph generation: Methods and applications

    Yanqiao Zhu, Yuanqi Du, Yinkai Wang, Yichen Xu, Jieyu Zhang, Qiang Liu, and Shu Wu. A survey on deep graph generation: Methods and applications. In Proceedings of the First Learning on Graphs Conference, volume 198 of Proceedings of Machine Learning Research, pages 47:1--47:21...

    2022 doi arXiv

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.