Pith. sign in

REVIEW 3 major objections 5 minor 52 references

Validity-first automatic polycube labeling for CAD models

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

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

Pith's one-line read This paper claims that a validity-first labeling pipeline removes the main obstacle to automatic polycube-based hex meshing for CAD models, producing valid labelings on 100% of the easier benchmark families and 97.2% of the harder CAD…

desk verdict Solid engineering contribution with honest limitations; the validity criteria are not proven sufficient, so read the 100% claims as 'valid under their local rules.' read the letter →

arxiv 2502.05712 v1 pith:Y4373XOJ submitted 2025-02-08 cs.CG

Sébastien Mestrallet , Christophe Bourcier , Franck Ledoux This is my paper · ORCID
classification cs.CG MSC 68U0565D1865N50
keywords polycubelabelinghexahedralmeshingCADmodelsvaliditygraphcutturning-pointfeaturepreservationmonotoneboundaries
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

This paper tries to remove the labeling bottleneck in automatic polycube-based hexahedral meshing for CAD models. It argues that previous methods fail because they use over-constrained validity criteria and mix validity with quality optimization, so it separates the two: a validity-oriented routine first repairs the labeling graph, and only then a monotonicity-oriented routine removes turning-points. The core claim, measured on two public CAD model collections, is that this pipeline produces valid labelings on essentially all tested models—100% valid and all-monotone on the easier benchmark families, 97.2% valid on the harder full collection—while running roughly 10 to 570 times faster than earlier automatic methods. If correct, the main obstacle to fully automatic all-hex mesh generation for industrial CAD geometry is substantially removed.

What carries the argument

The central object is the labeling graph: charts formed by connected triangles with the same axis label, the boundaries between them, and the corners where several boundaries meet. A boundary is monotone when it has no turning-points, meaning no vertices where it markedly drifts from its assigned direction. The argument is carried by semi-global operators that edit this graph locally—increasing a chart's valence by inserting a new chart, joining two turning-points with a chart, pulling a corner toward a turning-point, moving a boundary, and straightening boundaries—combined with the existing chart-removal, opposite-boundary and high-valence-corner repairs. These operators are orchestrated in two routines that separate validity repair from monotonicity repair, with loop detection to avoid infinite backtracking.

What would settle it

Take any CAD model whose labeling passes the paper's local validity checks, push the labeled surface to its assigned axis directions, and quantize the result onto the integer grid; if doing that ever produces overlapping faces, a collapsed chart, or a grid that cannot be unwrapped into a solid polycube, then the local criteria are not sufficient to guarantee a true polycube.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes that a surface labeling of a tetrahedral CAD mesh can be made a valid polycube labeling by a fully automatic, deterministic pipeline, without trading away feature preservation. The validity notion is deliberately relaxed: a boundary is accepted when it separates different axes, or even same-axis charts when the solid angle is greater than 180 degrees, and a corner is accepted when its incident boundaries are associable in pairs or form an XYZ trio, which subsumes the higher-valence configurations known to be realizable. Starting from a graph-cut labeling whose node costs are tilted on near-ambiguous normals, the pipeline alternates between fixing invalid charts, boundaries and corners and then suppressing turning-points. The reported outcome is 100% valid all-monotone labelings on the 74-model and 30-model benchmark families, 97.2% valid labelings on the 109-model CAD family, and hex meshes whose scaled-Jacobian statistics are comparable to or better than those of the earlier baselines.

Load-bearing premise

The paper's validity checks are local checks on the outer surface only, and the authors themselves note that those checks cannot catch global conflicts such as contradictory surface-normal requirements or twists hidden inside the volume; if such conflicts occur, a labeling recorded as valid may not actually correspond to any real polycube.

Editorial extensions

If this is right

  • Most CAD models in the tested collections no longer need manual or evolutionary labeling: the whole labeling stage finishes in tens of seconds rather than minutes to hours.
  • Valid all-monotone labelings on the basic and simple benchmark families mean downstream hex extraction can start directly, without a human repair step.
  • Relaxed corner and boundary criteria unlock configurations, such as 6-connected corners and same-axis boundaries with solid angle above 180 degrees, that earlier validity rules wrongly rejected.
  • Feature edges are preserved on the large majority of sharp edges, so the resulting hex meshes retain the CAD model's intended sharp features.
  • Because the routines are deterministic and fast, the method is suitable for design-loop use where many variants of a part must be meshed in sequence.

Reading between the lines

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

  • Editorial inference: the speedup makes labeling cheap enough to embed in interactive design loops, where a user edits a CAD model and receives a fresh valid labeling in seconds; this use is not discussed in the paper.
  • Editorial inference: the same operator set could plausibly be adapted to repair invalid configurations in frame-field-based hex meshing, where local invalid charts also need fixing; the paper does not explore this transfer.
  • Editorial inference: the sufficiency weakness flagged by the authors suggests the validity rates should be re-verified on high-genus models or models with near-self-intersecting features, where in-volume twists are more likely to hide; the tested collections may under-represent those cases.
  • Editorial inference: the 0.05-radian normal tilt and the compactness-over-fidelity ratio of one-third are calibration choices whose sensitivity to mesh resolution and feature-angle distributions is untested and could be quantified by ablation.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper presents a fully automatic method for the labeling stage of polycube-based hexahedral meshing, targeting CAD models. The method combines a modified graph-cut initial labeling with two post-processing routines: a validity-oriented routine that repairs invalid charts, boundaries, and corners using both existing and new semi-global operators, and a monotonicity-oriented routine that removes turning-points while preserving validity. The validity criteria are relaxed relative to earlier work by allowing opposite-axis boundaries along sufficiently obtuse solid edges and by permitting corner valences higher than 3 under an associativity condition. The method is evaluated on the MAMBO and OctreeMeshing CAD datasets, reporting high rates of valid and all-monotone labelings, comparable fidelity, and speedups of one to two orders of magnitude over PolyCut and Evocube. The implementation is open source.

Significance. If the central validity claim holds, this is a practically significant advance: it largely automates a bottleneck in polycube-based hex meshing for CAD models and does so at much lower cost than existing optimization-based labeling methods. The paper's strengths are its extensive empirical evaluation on two public datasets, comparison with two established baselines, use of downstream hex-mesh quality measures (Scaled Jacobian), and the release of open-source code. The authors also make a useful conceptual separation between validity (topological feasibility) and quality (distortion and feature preservation). The main risk is that the validity predicate used for the headline statistics is local and is not proven sufficient for actual polycube realizability; the authors themselves flag this residual gap in Section 8.

major comments (3)
  1. [Section 4 and Section 8] The validity predicate defined in Section 4 consists of local conditions on charts, boundaries, and corners, and the authors state in Section 8 that 'by working on the input surface, we cannot detect conflicting normal constraints [36] or in-volume twists [37] that would classify a labeling as valid despite not representing a polycube.' Because Tables 1 and 2 report validity rates using this predicate, the headline claim that the method produces valid labelings is only as strong as the unproven sufficiency of these local rules. Please either (a) provide a proof of sufficiency for the class of genus-0 CAD boundary meshes under the expanded rules, or (b) add an experimental check on cases with known ground truth (e.g., the counterexamples of [36] and [37]) and quantify how many of the reported 'valid' labelings would fail a full polycube-map or volumetric deformation test. Without this, the 'valid' counts should be described as 'locally valid under the Section 4 criteria' rather than as proven polycube-feasible.
  2. [Section 4] The optional boundary rule allows opposite-axis labels when the solid angle is greater than 180°, and the corner rule allows pair-associable boundaries or an XYZ trio. The paper does not give a precise, reproducible definition of how the solid angle is measured on the input triangle mesh: is it the angle between chart normals, a dihedral angle accumulated along the boundary, or an angle computed after a local parameterization? This is load-bearing because the expanded valid space is a central contribution and because the classification of a given labeling can change with the geometric discretization. Please provide the exact formula or algorithm, including any numerical tolerances.
  3. [Section 5 and Section 7] The method introduces several free parameters: the graph-cut sensitivity threshold (10^-10), the normal tilt angle (0.05 rad), the compactness/fidelity ratio (1/3), and the maximum iteration count Nmax in Algorithm 1. No ablation or sensitivity study is reported, so the robustness of the headline rates (e.g., the 100% valid/all-monotone results on MAMBO/Basic and Simple) to these choices is unclear. Please report a sensitivity analysis on a representative subset of the datasets, or justify each value by a prior reference and a stability test.
minor comments (5)
  1. [Section 5] In the paragraph describing the graph-cut construction, 'each edge of E connects two nodes of E' should read 'two nodes of N'.
  2. [Section 8] The sentence 'No hex-meshing was attempted our initial labeling' is missing the word 'on' before 'our initial labeling'.
  3. [Section 6.4.1] The phrase 'the closer to v is moved on v' should be reworded, e.g., 'the equilibrium point closer to v is moved to v'.
  4. [Tables 1 and 2] The header 'Overall average minSJ; avgSJ ratio cumulative' is unclear; please define 'ratio cumulative' in the caption or split the column into two clearly named statistics.
  5. [Figure 5 caption] The caption 'A labeled genus-0 shape satisfying criteria from [23] despite not leading to a polycube if the labels' sign is enforced: cuboid slope branches will be mixed-up on the same z-coordinates' is hard to parse; consider rephrasing for clarity.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the validity criteria are externally benchmarked and the admitted global-conflict limitation is a correctness caveat, not a definitional loop.

full rationale

The paper's central claim is that its labeling algorithm produces high rates of 'valid' labelings on CAD datasets. The validity predicate is a proposed local graph criterion defined in Section 4 and justified by a combination of external references ([23], [32]) and the authors' own prior analysis ([37]). The reported success rates are measurements against that predicate, but the paper does not stop at the predicate: it independently generates hex meshes with HexEx and reports scaled Jacobian statistics (Tables 1-3), and it explicitly concedes that the local criteria cannot detect conflicting normal constraints or in-volume twists (Section 8). That concession means the criteria are presented as a practical, possibly incomplete definition rather than as a theorem derived from themselves. Self-citations to Evocube [31] and to [37] are present, but they are not the sole justification for the central claim: the boundary/corner criteria are tied to the external valid-space expansion of [32], and the method is benchmarked against the external OctreeMeshing dataset in addition to the co-authored MAMBO dataset. No equation is fitted and then renamed as a prediction, and no uniqueness or ansatz is imported from a self-citation chain. The residual concern that local validity may overstate true polycube feasibility is a correctness risk explicitly acknowledged by the authors, not a circular derivation.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claim rests on the paper's own validity criteria, which are local and heuristic, plus a set of hand-tuned parameters in the graph-cut initialization and routine loops. No new physical or geometric entities are postulated; the method reuses the standard polycube labeling vocabulary.

free parameters (4)
  • graph-cut sensitivity threshold = 1e-10
    Used in Section 5 to flag triangles whose two closest labels have nearly equal weight; chosen ad hoc to avoid fragmentation.
  • normal tilt angle = 0.05 rad
    Applied to normals of ambiguous triangles in Section 5 to force consistent label choice; no sensitivity analysis.
  • compactness/fidelity ratio = 1/3
    Graph-cut weighting parameter in Section 5; hand-selected.
  • max iterations Nmax = not specified numerically
    Bound on the validity routine loop in Algorithm 1; never stated, so the loop terminates by other conditions or implicitly by the tracking set.
assumptions (5)
  • domain assumption The local validity criteria of Section 4 (chart valence, boundary solid angle, corner associability) are assumed to characterize labelings that map to a valid polycube.
    The paper relies on this to call labelings 'valid'; it is not proven, and the authors admit that conflicting normal constraints and in-volume twists can slip through (Section 8).
  • domain assumption The labeling graph extracted from the triangle mesh correctly captures the charts, boundaries, and corners of the underlying polycube structure.
    Section 3.2 assumes the union-find aggregation and boundary tracing faithfully represent the topology.
  • standard math Known results on orthogonal polyhedra (Eppstein and Mumford [23]) and the expanded space of He et al. [32] are correct.
    The new corner criterion is justified by reference to [32] and the geometric analysis in Section 4.
  • standard math Graph-cut optimization minimizes the stated energy and produces a labeling from which the labeling graph can be reconstructed.
    Section 5 relies on the standard multi-label graph-cut machinery of Boykov et al. and the validity of its outputs.
  • domain assumption The downstream pipeline (morphing, quantization, HexEx) can extract a hex mesh from a valid labeling.
    Section 8 uses HexEx [21] and the work of Protais et al. [22]; the paper does not re-derive this and assumes these tools behave as documented.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Validity-first automatic polycube labeling for CAD models." pith.science (2026). https://pith.science/paper/Y4373XOJ

@misc{pith2026250205712,
  author       = {Pith},
  title        = {Pith review of: Validity-first automatic polycube labeling for CAD models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Y4373XOJ}},
  note         = {Machine review of arXiv:2502.05712}
}
read the original abstract

For many simulation codes, block-structured hex meshes remain preferred while their automatic generation is unsolved. We investigate the usage of a polycube-based approach. More specifically, we focus on the labeling stage, which consists in assigning each boundary facet to one of the 6 signed principal axis. Similar works are confronted with 2 challenges: over-constraining validity criteria, and the conflated processing of validity criteria with quality metrics. We tackle these obstacles with automatic routines based on semi-global labeling operators. Our approach is successfully tested on CAD models, which are of interest for many numerical simulation problems.

Figures

Figures reproduced from arXiv: 2502.05712 by the authors.

Figure 1
Figure 1. Collection of polycube labelings generated using our algorithm. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Our pipeline steps: (a) CAD model; (b) tetrahedral [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Diagram of our method. After a graph-cut opti [PITH_FULL_IMAGE:figures/full_fig_p002_3.png] view at source ↗
Figures from the paper (14 more)
Figure 4
Figure 4. Figure 4: Labelings that do not correspond to simple orthog [PITH_FULL_IMAGE:figures/full_fig_p003_4.png]
Figure 5
Figure 5. Figure 5: (left) 4 cubes joined to form a valid polycube with a [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]
Figure 6
Figure 6. Figure 6: Two shapes with non-orthogonal boundaries: (a) [PITH_FULL_IMAGE:figures/full_fig_p004_6.png]
Figure 8
Figure 8. Figure 8: (left) a triangle mesh with feature edges (thick [PITH_FULL_IMAGE:figures/full_fig_p004_8.png]
Figure 9
Figure 9. Figure 9: Better discrimination between valid and invalid [PITH_FULL_IMAGE:figures/full_fig_p005_9.png]
Figure 10
Figure 10. Figure 10: (a) naive labeling, resulting in charts fragmenta [PITH_FULL_IMAGE:figures/full_fig_p005_10.png]
Figure 12
Figure 12. Figure 12: Split of a non-planar chart at a turning-point (left, [PITH_FULL_IMAGE:figures/full_fig_p006_12.png]
Figure 14
Figure 14. Figure 14: The increase chart valence operator applied on MAMBO [17] B23 (left) and B49 (right). not coincide with one of them, the closer to 𝑣 is moved on 𝑣 to enforce the placement of a corner on 𝑣: between 𝑜𝑝↻ and 𝑜𝑝↺ will be the new boundary around 𝒞𝑖 . Afterwards, we greedi…
Figure 15
Figure 15. Figure 15: The join turning-points pair operator applied on MAMBO [17] B29 [PITH_FULL_IMAGE:figures/full_fig_p007_15.png]
Figure 16
Figure 16. Figure 16: The pull closest corner operator applied on MAMBO [17] S36. The turning-point is in yellow [PITH_FULL_IMAGE:figures/full_fig_p008_16.png]
Figure 17
Figure 17. Figure 17: The move boundary near turning-point operator applied on CAD5 [45]. Here the two turning-points are removed in one iteration. with 𝑙, and we extend the relabeling to adjacent facet of the same side, effectively creating a new chart (Fig.15). Unlike Evocube [31] (Secti…
Figure 18
Figure 18. Figure 18: (left to right) Comparisons between PolyCut [27], Evocube [31] and our algorithm, with hexahedral meshes gen￾erated with [21], [46]. Models MAMBO [17] B76, S35 and B49 [PITH_FULL_IMAGE:figures/full_fig_p011_18.png]
Figure 20
Figure 20. Figure 20: Invalid labeling on MAMBO [17] M8 [PITH_FULL_IMAGE:figures/full_fig_p012_20.png]
Figure 19
Figure 19. Figure 19: Labelings of Evocube [31] (left) and our method (right) on the cheese2 model from [45]. It has 6-connected corners that Evocube classifies as invalid and tries to avoid, whereas we rightfully classify them as valid. Evocube lasts 4min 36s, while we return a result in …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 48 canonical work pages

  1. [36]

    Fixing normal constraints for generation of polycubes,

    D. Sokolov and N. Ray, “Fixing normal constraints for generation of polycubes,” HAL:hal-1211408, 2015. [Online]. Available: https://hal.inria.fr/hal-01211408

    work page 2015

  2. [37]

    Limits and prospects of polycube labelings,

    S. Mestrallet , F. Protais, C. Bourcier , and F. Ledoux, “Limits and prospects of polycube labelings,” 2023, HAL:cea-4169841. [Online]. Available: https://cea. hal.science/cea-04169841

    work page 2023

  3. [1]

    Hexahedral mesh generation constraints,

    J. F. Shepherd and C. R. Johnson , “Hexahedral mesh generation constraints,” Engineering with Com- puters, vol. 24, pp. 195–213, 2008, doi: 10.1007/ s00366-008-0091-4

    work page 2008

  4. [2]

    Hex-Mesh Generation and Pro- cessing: A Survey,

    N. Pietroni et al., “Hex-Mesh Generation and Pro- cessing: A Survey,” ACM Trans. Graph., vol. 42, no. 2, pp. 1–44, 2022, doi: 10.1145/3554920

    doi:10.1145/3554920 2022

  5. [3]

    Interactive all-hex meshing via cuboid decomposition,

    L. Li, P. Zhang, D. Smirnov, S. Mazdak Abulnaga, and J. Solomon , “Interactive all-hex meshing via cuboid decomposition,” ACM Trans. Graph., vol. 40, no. 6, 2021, doi: 10.1145/3478513.348056

    work page arXiv 2021

  6. [4]

    HexBox: Interactive Box Modeling of Hexahedral Meshes,

    F. Zoccheddu, E. Gobbetti, M. Livesu, N. Pietroni, and G. Cherchi, “HexBox: Interactive Box Modeling of Hexahedral Meshes,” Computer Graphics Forum, vol. 42, no. 5, 2023, doi: 10.1111/cgf.14899

    work page doi:10.1111/cgf.14899 2023

  7. [5]

    Hex-dominant meshing: Mind the gap!,

    N. Ray, D. Sokolov, M. Reberol, F. Ledoux, and B. Lévy , “Hex-dominant meshing: Mind the gap!,” Computer-Aided Design, vol. 102, pp. 94–103, 2018, doi: 10.1016/j.cad.2018.04.012

    work page doi:10.1016/j.cad.2018.04.012 2018

  8. [6]

    A grid-based algorithm for the gener- ation of hexahedral element meshes,

    R. Schneiders, “A grid-based algorithm for the gener- ation of hexahedral element meshes,” Engineering with Computers, vol. 5, no. 3, pp. 168–177, 1996, doi: 10.1007/ BF01198732

    work page 1996

Show all 52 references
  1. [7]

    Feature Pre- serving Octree-Based Hexahedral Meshing,

    X. Gao , H. Shen , and D. Panozzo , “Feature Pre- serving Octree-Based Hexahedral Meshing,” Computer Graphics Forum, vol. 38, no. 5, pp. 135–149, 2019, doi: 10.1111/cgf.13795

    2019 doi

  2. [8]

    Advances in Octree-Based All-Hexahe- dral Mesh Generation: Handling Sharp Features,

    L. Maréchal, “Advances in Octree-Based All-Hexahe- dral Mesh Generation: Handling Sharp Features,” in Proceedings of International Meshing Roundtable, 2009, pp. 65–84. doi: 10.1007/978-3-642-04319-2_5

    2009 doi

  3. [9]

    Hexahedral Mesh Generation using the Embedded Voronoi Graph,

    A. Sheffer , M. Etzion , A. Rappoport , and M. Bercovier, “Hexahedral Mesh Generation using the Embedded Voronoi Graph,” Engineering with Com- puters, vol. 15, pp. 248–262, 1999, doi: 10.1007/ s003660050020

    1999

  4. [10]

    Skeleton-driven Adaptive Hexahedral Meshing of Tubu- lar Shapes,

    M. Livesu, A. Muntoni, E. Puppo, and R. Scateni, “Skeleton-driven Adaptive Hexahedral Meshing of Tubu- lar Shapes,” Computer Graphics Forum , 2014, doi: 10.1111/cgf.13021

    2014 doi

  5. [11]

    Meso- Skeleton Guided Hexahedral Mesh Design,

    P. Viville, P. Kraemer, and D. Bechmann, “Meso- Skeleton Guided Hexahedral Mesh Design,” Computer Graphics Forum, vol. 42, no. 7, 2023, doi: 10.1111/ cgf.14932

    2023

  6. [12]

    LayTracks3D: A New Approach to Meshing General Solids using Medial Axis Trans- form,

    Quadros W. R. , “LayTracks3D: A New Approach to Meshing General Solids using Medial Axis Trans- form,” Procedia Engineering , 2014, doi: 10.1016/ j.proeng.2014.10.374

    2014

  7. [14]

    Polycube-Maps,

    M. Tarini, K. Hormann, P. Cignoni, and C. Mon- tani, “Polycube-Maps,” in SIGGRAPH, ACM, 2004, pp. 853–860. doi: 10.1145/1186562.1015810

    2004

  8. [15]

    Cube- Cover - Parameterization of 3D Volumes,

    M. Nieser, U. Reitebuch, and K. Polthier, “Cube- Cover - Parameterization of 3D Volumes,” Computer Graphics Forum , vol. 30, pp. 1397–1406, 2011, doi: 10.1111/j.1467-8659.2011.02014.x

    2011

  9. [16]

    Hexahedral-Dominant Meshing,

    D. Sokolov , N. Ray , L. Untereiner , and Lévy Bruno, “Hexahedral-Dominant Meshing,” ACM Trans. Graph., vol. 35, no. 5, 2016, doi: 10.1145/2930662

    2016 doi

  10. [17]

    Ledoux, MAMBO: Model dAtabase Mesh BlOcking

    F. Ledoux, MAMBO: Model dAtabase Mesh BlOcking. Accessed: May 20, 2024. [Online]. Available: https:// gitlab.com/franck.ledoux/mambo

    2024

  11. [18]

    Gmsh: a three- dimensional finite element mesh generator with built-in pre- and post-processing facilities,

    C. Geuzaine and J.-F. Remacle , “Gmsh: a three- dimensional finite element mesh generator with built-in pre- and post-processing facilities,” 2008. doi: 10.1002/ nme.2579

    2008

  12. [19]

    3D Precise Mesh

    Dassault Systèmes - Spatial Corp. , “3D Precise Mesh.” [Online]. Available: https://www.spatial.com/ products/3d-precise-mesh

  13. [20]

    SALOME: The open source platform for numerical simulation

    EDF, CEA, and OpenCascade, “SALOME: The open source platform for numerical simulation.” [Online]. Available: https://www.salome-platform.org/?lang=en

  14. [22]

    Robust Quantization for Polycube Maps,

    F. Protais, M. Reberol , N. Ray , E. Corman , F. Ledoux, and D. Sokolov, “Robust Quantization for Polycube Maps,” Computer-Aided Design, vol. 150, 2022, doi: 10.1016/j.cad.2022.103321

    2022

  15. [23]

    Steinitz Theorems for Orthogonal Polyhedra,

    D. Eppstein and E. Mumford , “Steinitz Theorems for Orthogonal Polyhedra,” in Proceedings of the 26th annual symposium on Computational Geometry , ACM, pp. 429–438. doi: 10.1145/1810959.1811030

  16. [24]

    Hexahedral shell mesh construction via volumetric polycube map,

    S. Han , J. Xia , and Y. He , “Hexahedral shell mesh construction via volumetric polycube map,” ACM Sym- posium on Solid and Physical Modeling , 2010, doi: 10.1145/1839778.1839796

    2010

  17. [25]

    Optimizing polycube domain construction for hexahedral remesh- ing,

    W. Yu, K. Zhang, S. Wan, and X. Li, “Optimizing polycube domain construction for hexahedral remesh- ing,” Computer-Aided Design, vol. 46, pp. 58–68, 2014, doi: 10.1016/j.cad.2013.08.018

    2014 doi

  18. [26]

    Efficient Volumetric PolyCube-Map Construction,

    X.-M. Fu, C.-Y. Bai, and Y. Liu, “Efficient Volumetric PolyCube-Map Construction,” vol. 35, no. 7, pp. 97–106, 2016, doi: 10.1111/cgf.13007

    2016 doi

  19. [27]

    PolyCut: Monotone Graph-Cuts for PolyCube Base-Complex Construction,

    M. Livesu , N. Vining , A. Sheffer , J. Gregson , and R. Scateni , “PolyCut: Monotone Graph-Cuts for PolyCube Base-Complex Construction,” in Pro- ceedings of SIGGRAPH Asia , ACM, 2013. doi: 10.1145/2508363.2508388

    2013

  20. [28]

    ℓ1-Based Construction of Polycube Maps from Complex Shapes,

    J. Huang, T. Jiang, Z. Shi, Y. Tong, H. Bao, and M. Desbrun, “ℓ1-Based Construction of Polycube Maps from Complex Shapes,” ACM Trans. Graph., vol. 33, no. 3, 2014, doi: 10.1145/2602141

    2014 doi

  21. [29]

    All- Hex Meshing using Closed-Form Induced Polycube,

    X. Fang , W. Xu , H. Bao , and J. Huang , “All- Hex Meshing using Closed-Form Induced Polycube,” ACM Trans. Graph. , vol. 35, no. 4, 2016, doi: 10.1145/2897824.2925957

    2016

  22. [30]

    Computing Surface PolyCube-Maps by Constrained Voxelization,

    Y. Yang, X.-M. Fu, and L. Liu, “Computing Surface PolyCube-Maps by Constrained Voxelization,” Com- puter Graphics Forum, vol. 38, no. 7, pp. 299–309, 2019, doi: 10.1111/cgf.13838

    2019 doi

  23. [31]

    Evocube: a Genetic Label- ing Framework for Polycube-Maps,

    C. Dumery , F. Protais , S. Mestrallet , C. Bourcier, and F. Ledoux, “Evocube: a Genetic Label- ing Framework for Polycube-Maps,” vol. 41, no. 6, pp. 467–479, 2022, doi: 10.1111/cgf.14649

    2022 doi

  24. [32]

    Expanding the Solvable Space of Polycube- Map via Validity-Enhanced Construction,

    L. He, N. Lei, Z. Wang, C. Wang, X. Zheng, and Z. Luo , “Expanding the Solvable Space of Polycube- Map via Validity-Enhanced Construction,” in Interna- tional Meshing Roundtable , 2024, pp. 40–52. doi: 10.1137/1.9781611978001.4

    2024 doi

  25. [33]

    Cut- enhanced PolyCube-maps for feature-aware all-hex meshing,

    H.-X. Guo , X. Liu , D.-M. Yan , and Y. Liu , “Cut- enhanced PolyCube-maps for feature-aware all-hex meshing,” ACM Trans. Graph. , vol. 39, no. 4, doi: 10.1145/3386569.3392378

  26. [34]

    Selective Padding for Polycube-Based Hexa- hedral Meshing,

    G. Cherchi, P. Alliez, R. Scateni, M. Lyon, and D. Bommes, “Selective Padding for Polycube-Based Hexa- hedral Meshing,” Computer Graphics Forum, vol. 38, no. 1, pp. 580–591, 2019, doi: 10.1111/cgf.13593

    2019 doi

  27. [35]

    Fun sheet matching: towards automatic block de- composition for hexahedral meshes,

    N. Kowalski, F. Ledoux, M. L. Staten, and Owen S. J., “Fun sheet matching: towards automatic block de- composition for hexahedral meshes,” Eng. with Comput, 2012

    2012

  28. [38]

    Lévy and contributors, Geogram: a program- ming library with geometric algorithms

    B. Lévy and contributors, Geogram: a program- ming library with geometric algorithms . Accessed: Jun. 24, 2024. [Online]. Available: https://github.com/ BrunoLevy/geogram

    2024

  29. [39]

    Polycube Shape Space,

    H. Zhao et al., “Polycube Shape Space,” Computer Graphics Forum , vol. 38, pp. 311–322, 2019, doi: 10.1111/cgf.13839

    2019 doi

  30. [40]

    Poly- cube Layouts via Iterative Dual Loops,

    M. Snoep, B. Speckmann, and K. Verbeek, “Poly- cube Layouts via Iterative Dual Loops,” 2024. doi: 10.48550/arXiv.2402.00652

    2024 doi arXiv

  31. [41]

    Polyeder und Raumeinteilungen,

    E. Steinitz, “Polyeder und Raumeinteilungen,” Ency- clopädie der mathematischen Wissenschaften, 1922

    1922

  32. [42]

    Fast approxi- mate energy minimization via graph cuts,

    Y. Boykov, O. Veksler, and R. Zabih, “Fast approxi- mate energy minimization via graph cuts,” IEEE Trans- actions on Pattern Analysis and Machine Intelligence , pp. 1222–1239, 2001, doi: 10.1109/34.969114

    2001 doi

  33. [43]

    What Energy Func- tions can be Minimized via Graph Cuts?,

    V. Kolmogorov and R. Zabih, “What Energy Func- tions can be Minimized via Graph Cuts?,” IEEE Trans- actions on Pattern Analysis and Machine Intelligence , pp. 147–159, 2004, doi: 10.1109/TPAMI.2004.1262177

    2004 arXiv

  34. [44]

    An Experimental Comparison of Min-Cut/Max-Flow Algorithms for En- ergy Minimization in Vision,

    Y. Boykov and V. Kolmogorov, “An Experimental Comparison of Min-Cut/Max-Flow Algorithms for En- ergy Minimization in Vision,” IEEE Transactions on Pattern Analysis and Machine Intelligence , pp. 1124– 1137, 2004, doi: 10.1109/TPAMI.2004.60

    2004 doi

  35. [45]

    OctreeMeshing dataset

    “OctreeMeshing dataset.” Accessed: Oct. 07, 2024. [On- line]. Available: https://cims.nyu.edu/gcl/papers/2019- OctreeMeshing.zip

    2024

  36. [46]

    Protais, polycube with HexEx

    F. Protais, polycube with HexEx . Accessed: May 25,

  37. [47]

    Protais , hexsmoothing

    F. Protais , hexsmoothing. Accessed: Oct. 04,

  38. [48]

    Intrinsic mixed-integer polycubes for hexahedral mesh- ing,

    M. Mandad, R. Chen, D. Bommes, and M. Campen, “Intrinsic mixed-integer polycubes for hexahedral mesh- ing,” Computer Aided Geometric Design, vol. 94, 2022, doi: 10.1016/j.cagd.2022.102078

    2022

  39. [49]

    Available: https://github.com/fprotais/ hexsmoothing

    [Online]. Available: https://github.com/fprotais/ hexsmoothing

  40. [50]

    Protais , fastbndpolycube

    F. Protais , fastbndpolycube. Accessed: Jun. 18,

  41. [51]

    Hex me if you can,

    P.-A. Beaufort, M. Reberol, D. Kalmykov, H. Liu, F. Ledoux , and D. Bommes , “Hex me if you can,” Computer Graphics Forum, vol. 41, no. 5, pp. 125–134, 2022, doi: 10.1111/cgf.14608

    2022 doi

  42. [53]

    Available: https://github.com/fprotais/ fastbndpolycube

    [Online]. Available: https://github.com/fprotais/ fastbndpolycube

  43. [54]

    Robust edge-preserving surface mesh polycube defor- mation,

    H. Zhao, N. Lei, X. Li, P. Zeng, K. Xu, and X. Gu, “Robust edge-preserving surface mesh polycube defor- mation,” Computational Visual Media, vol. 4, pp. 33–42, 2018, doi: 10.1007/s41095-017-0100-x

    2018 doi

  44. [2024]

    Available: https://github.com/fprotais/ polycube_withHexEx

    [Online]. Available: https://github.com/fprotais/ polycube_withHexEx

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.