arxiv: 2604.13191 · v1 · submitted 2026-04-14 · 💻 cs.GR · cs.LG

Recognition: unknown

Fast Voxelization and Level of Detail for Microgeometry Rendering

Javier Fabre , Carlos Castillo , Carlos Rodriguez-Pardo , Jorge Lopez-Moreno

Authors on Pith no claims yet

Pith reviewed 2026-05-10 13:21 UTC · model grok-4.3

classification 💻 cs.GR cs.LG

keywords voxelizationlevel of detailmicrogeometryanisotropic scatteringSGGX clusteringpath tracingparallel renderingvolume aggregation

0 comments

The pith

An efficient parallel voxelization method combined with hierarchical SGGX clustering enables accurate multi-resolution rendering of anisotropic microgeometry.

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

Surfaces with sparse micro structures such as fibers or brushed metal ridges produce anisotropic light scattering that requires high-resolution voxel data, yet building and querying such volumes is slow and memory intensive. The paper presents a parallel voxelization algorithm that quickly generates aggregated data across multiple resolution levels from either triangle meshes or explicit fiber models. It then introduces a hierarchical SGGX clustering representation to combine the scattering appearance from different distances with higher fidelity than standard level-of-detail baselines. This matters for path-tracing pipelines because it reduces the number of samples computed per pixel while maintaining visual quality without extra tuning parameters.

Core claim

The work introduces an efficient parallel voxelization method designed to facilitate fast data aggregation at multiple resolution levels, together with a novel representation based on hierarchical SGGX clustering that provides better accuracy than baseline methods. The approach is implemented in CUDA and tested on both triangle meshes and volumetric fabrics modeled with explicit fibers, then used inside a path tracer that follows the proposed level-of-detail rendering model.

What carries the argument

Hierarchical SGGX clustering, a tree-structured grouping of symmetric GGX scattering parameters that aggregates anisotropic appearance data across successive resolution levels.

If this is right

Preprocessing time for high-resolution voxel grids of sparse microgeometry drops because aggregation happens during parallel voxelization.
Distant views of fiber-like or ridged surfaces retain higher visual detail than with conventional mip-map or averaging LoD schemes.
Path tracers can use fewer samples per pixel while still producing comparable results to full-resolution rendering.
The same data structure works for both polygon meshes and models built from explicit fiber geometry.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The preprocessing speed gain may allow on-demand voxelization inside interactive applications that currently rely on precomputed volumes.
The clustering approach could be adapted to other parametric scattering models beyond SGGX if the underlying distribution remains amenable to hierarchical aggregation.
Memory footprints in large scenes might decrease further if the hierarchy is pruned according to view-dependent error bounds.

Load-bearing premise

Hierarchical SGGX clustering can combine anisotropic scattering appearance from different distances while keeping visual fidelity higher than existing level-of-detail techniques and without requiring post-hoc parameter adjustments.

What would settle it

Side-by-side path-traced images of the same microgeometry scene rendered at multiple viewing distances, comparing the new hierarchical SGGX level-of-detail model against both full-resolution voxel data and standard baseline LoD methods using quantitative image-difference metrics such as mean squared error or perceptual hashes.

read the original abstract

Many materials show anisotropic light scattering patterns due to the shape and local alignment of their underlying micro structures: surfaces with small elements such as fibers, or the ridges of a brushed metal, are very sparse and require a high spatial resolution to be properly represented as a volume. The acquisition of voxel data from such objects is a time and memory-intensive task, and most rendering approaches require an additional Level-of-Detail (LoD) data structure to aggregate the visual appearance, as observed from multiple distances, in order to reduce the number of samples computed per pixel (E.g.: MIP mapping). In this work we introduce first, an efficient parallel voxelization method designed to facilitate fast data aggregation at multiple resolution levels, and second, a novel representation based on hierarchical SGGX clustering that provides better accuracy than baseline methods. We validate our approach with a CUDA-based implementation of the voxelizer, tested both on triangle meshes and volumetric fabrics modeled with explicit fibers. Finally, we show the results generated with a path tracer based on the proposed LoD rendering model.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper delivers a practical parallel voxelization scheme plus hierarchical SGGX clustering for faster, more accurate LoD on sparse anisotropic microgeometry.

read the letter

The core contribution is a voxelizer built from the start for quick multi-resolution aggregation, paired with a hierarchical SGGX clustering representation that targets better directional fidelity than standard LoD baselines for things like fibers or brushed surfaces. That combination is new enough on its own terms and directly tackles the memory and sampling cost of high-resolution volumes for anisotropic scattering. The CUDA implementation tested on both triangle meshes and explicit-fiber volumes, plus the path-tracer integration, shows they worked through the actual pipeline constraints rather than stopping at an abstract idea. That is useful engineering work for anyone who has to render these materials at varying distances. The claim of improved accuracy rests on the clustering step preserving scattering patterns without heavy post-processing, and the stress-test note indicates the validation setup matches the sparsity and resolution needs of the problem. Still, the abstract gives no error metrics, baseline tables, or scene counts, so the size of the accuracy gain and whether it holds without hidden tuning remain open until the results section is checked. If the comparisons are only qualitative or limited to a few materials, the advantage could turn out narrower than stated. This is aimed at graphics researchers and practitioners who already work with volume representations of microgeometry and need concrete speed-ups for LoD. Readers who care about implemented voxel pipelines or anisotropic LoD will find usable details here. The work is concrete and reproducible enough to deserve a serious referee, even if the quantitative claims need tightening in revision. I would send it out for review.

Referee Report

2 major / 2 minor

Summary. The paper introduces an efficient parallel voxelization method for fast multi-resolution data aggregation of sparse anisotropic microgeometry (e.g., fibers, brushed metals) and a novel hierarchical SGGX clustering representation for Level-of-Detail (LoD) that is claimed to provide better accuracy than baseline methods. Validation consists of a CUDA implementation tested on triangle meshes and explicit-fiber volumetric fabrics, together with qualitative results from a path tracer using the proposed LoD model.

Significance. If the accuracy and performance claims are substantiated, the work would offer a practical advance for rendering high-resolution anisotropic volumes by accelerating voxelization and improving directional scattering fidelity in LoD structures, with direct relevance to offline and real-time graphics pipelines handling microgeometry.

major comments (2)

[Abstract / Validation] Abstract and validation description: the central claim that hierarchical SGGX clustering 'provides better accuracy than baseline methods' is unsupported because the text provides no quantitative metrics, error bars, PSNR/SSIM values, or explicit baseline comparisons (e.g., against standard MIP-mapping or prior SGGX LoD), which is load-bearing for the accuracy contribution.
[Results] Results section (path-tracer demonstrations): without reported timing breakdowns, memory usage, or convergence comparisons between the proposed voxelizer/LoD and existing methods, it is impossible to assess whether the parallel voxelization delivers the claimed efficiency gains for multi-resolution aggregation.

minor comments (2)

Clarify the exact clustering objective and distance metric used in the hierarchical SGGX step; the current description leaves the aggregation rule for anisotropic scattering ambiguous.
Add explicit pseudocode or complexity analysis for the parallel voxelization kernel to make the CUDA implementation reproducible.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback and for recognizing the potential relevance of our work to rendering pipelines handling anisotropic microgeometry. We address each major comment below and outline the revisions we will make to strengthen the quantitative validation.

read point-by-point responses

Referee: [Abstract / Validation] Abstract and validation description: the central claim that hierarchical SGGX clustering 'provides better accuracy than baseline methods' is unsupported because the text provides no quantitative metrics, error bars, PSNR/SSIM values, or explicit baseline comparisons (e.g., against standard MIP-mapping or prior SGGX LoD), which is load-bearing for the accuracy contribution.

Authors: We agree that the accuracy claim requires quantitative backing. The current manuscript supports the claim through qualitative visual comparisons in the path-tracer results, where the hierarchical SGGX LoD exhibits improved directional scattering fidelity over baselines. To address this, we will add explicit quantitative evaluations in the revised manuscript, including PSNR and SSIM metrics against high-resolution ground-truth renders, error bars across multiple scenes, and direct numerical comparisons to MIP-mapping and prior SGGX aggregation methods. revision: yes
Referee: [Results] Results section (path-tracer demonstrations): without reported timing breakdowns, memory usage, or convergence comparisons between the proposed voxelizer/LoD and existing methods, it is impossible to assess whether the parallel voxelization delivers the claimed efficiency gains for multi-resolution aggregation.

Authors: The referee is correct that the results section currently lacks these quantitative performance details. While the manuscript describes the CUDA voxelizer implementation and its testing on triangle meshes and explicit-fiber volumes, no specific timing, memory, or convergence data are reported. In the revision we will add performance tables and analysis, including voxelization runtimes at multiple resolutions, memory consumption of the hierarchical representation, and rendering convergence/speed comparisons against baseline methods. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation is self-contained

full rationale

The paper introduces a parallel voxelization scheme and hierarchical SGGX clustering representation for microgeometry LoD. No equations, parameters, or central claims in the provided abstract or description reduce by construction to fitted inputs, self-definitions, or load-bearing self-citations. The validation via CUDA implementation on triangle meshes and explicit-fiber volumes provides independent empirical support, with no renaming of known results or ansatz smuggling detectable. The method stands as an independent contribution without internal reductions to its own inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only view yields no explicit free parameters, axioms, or invented entities; SGGX is treated as an existing model being extended hierarchically, with no details on any clustering thresholds or aggregation rules.

pith-pipeline@v0.9.0 · 5487 in / 1075 out tokens · 47335 ms · 2026-05-10T13:21:28.712263+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

61 extracted references · 5 canonical work pages

[1]

A fast voxel traversal algorithm for ray tracing,

J. Amanatides and A. Woo, “A fast voxel traversal algorithm for ray tracing,” inProceedings of Euro- graphics, pp. 3–10, Eurographics Association, 1987

1987
[2]

Rendering participating media with bidirectional path trac- ing,

E. P. Lafortune and Y. D. Willems, “Rendering participating media with bidirectional path trac- ing,” inRendering Techniques (Proceedings of the Eurographics Workshop on Rendering), (Vienna), pp. 91–100, Springer-Verlag, June 1996

1996
[3]

Voxel-based visibility analy- sis for safety assessment of urban environments,

M. Aleksandrov, S. Zlatanova, L. Kimmel, J. Bar- ton, and B. Gorte, “Voxel-based visibility analy- sis for safety assessment of urban environments,” ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, vol. 4, pp. 11–17, 2019

2019
[4]

NeRF: Rep- resenting scenes as neural radiance fields for view synthesis,

B. Mildenhall, P. P. Srinivasan, M. Tancik, J. T. Barron, R. Ramamoorthi, and R. Ng, “NeRF: Rep- resenting scenes as neural radiance fields for view synthesis,” inEuropean Conference on Computer Vision (ECCV), 2020

2020
[5]

Nerf++: Analyzing and improving neural radiance fields

K. Zhang, G. Riegler, N. Snavely, and V. Koltun, “Nerf++: Analyzing and improving neural radi- ance fields,”arXiv:2010.07492 [cs.CV], 2020

work page arXiv 2010
[6]

3D gaussian splatting for real-time radiance field rendering,

B. Kerbl, G. Kopanas, T. Leimk¨ uhler, and G. Dret- takis, “3D gaussian splatting for real-time radiance field rendering,”ACM Transactions on Graphics, vol. 42, July 2023

2023
[7]

4D gaussian splatting for real-time dynamic scene rendering,

G. Wu, T. Yi, J. Fang, L. Xie, X. Zhang, W. Wei, W. Liu, Q. Tian, and X. Wang, “4D gaussian splatting for real-time dynamic scene rendering,” inIEEE/CVF International Conference on Com- puter Vision and Pattern Recognition (CVPR), pp. 20310–20320, June 2024. 12 Javier Fabre et al

2024
[8]

3D gaussian ray tracing: Fast tracing of particle scenes,

N. Moenne-Loccoz, A. Mirzaei, O. Perel, R. de Lu- tio, J. M. Esturo, G. State, S. Fidler, N. Sharp, and Z. Gojcic, “3D gaussian ray tracing: Fast tracing of particle scenes,”ACM Transactions on Graph- ics (Proceedings of SIGGRAPH Asia), 2024

2024
[9]

Don’t splat your gaus- sians: Volumetric ray-traced primitives for model- ing and rendering scattering and emissive media,

J. Condor, S. Speierer, L. Bode, A. Bozic, S. Green, P. Didyk, and A. Jarabo, “Don’t splat your gaus- sians: Volumetric ray-traced primitives for model- ing and rendering scattering and emissive media,” ACM Transactions on Graphics, vol. 44, Feb. 2025

2025
[10]

Radiant foam: Real-time differen- tiable ray tracing.arXiv:2502.01157, 2025

S. Govindarajan, D. Rebain, K. M. Yi, and A. Tagliasacchi, “Radiant foam: Real-time differ- entiable ray tracing,”arXiv:2502.01157 [cs.CV], 2025

work page arXiv 2025
[11]

Sparse voxels rasterization: Real-time high- fidelity radiance field rendering,

C. Sun, J. Choe, C. Loop, W.-C. Ma, and Y.-C. F. Wang, “Sparse voxels rasterization: Real-time high- fidelity radiance field rendering,”arXiv:2412.04459 [cs.CV], 2024

work page arXiv 2024
[12]

Splatvoxel: History-aware novel view streaming without tem- poral training,

Y. Wang, L. Chai, X. Luo, M. Niemeyer, M. Lagu- nas, S. Lombardi, S. Tang, and T. Sun, “Splatvoxel: History-aware novel view streaming without tem- poral training,”arXiv:2503.14698 [cs.CV], 2025

work page arXiv 2025
[13]

Voxelisation algorithms and data structures: A re- view,

M. Aleksandrov, S. Zlatanova, and D. J. Heslop, “Voxelisation algorithms and data structures: A re- view,”Sensors, vol. 21, no. 8241, 2021

2021
[14]

Fast parallel surface and solid voxelization on GPUs,

M. Schwarz and H.-P. Seidel, “Fast parallel surface and solid voxelization on GPUs,”ACM Transac- tions on Graphics, vol. 29, no. 6, pp. 1–10, 2010

2010
[15]

Real-time voxelization for complex polygonal models,

Z. Dong, W. Chen, H. Bao, H. Zhang, and Q. Peng, “Real-time voxelization for complex polygonal models,” inPacific Conference on Com- puter Graphics and Applications, (Washington, DC, USA), pp. 43–50, IEEE Computer Society, 2004

2004
[16]

Real-time high- resolution sparse voxelization with application to image-based modeling,

C. Loop, C. Zhang, and Z. Zhang, “Real-time high- resolution sparse voxelization with application to image-based modeling,” inProceedings of High Per- formance Graphics, (New York, NY, USA), pp. 73– 79, ACM Press, July 2013

2013
[17]

VDB: High-resolution sparse volumes with dynamic topology,

K. Museth, “VDB: High-resolution sparse volumes with dynamic topology,”ACM Transactions on Graphics, vol. 32, pp. 27:1–27:22, July 2013

2013
[18]

The SGGX microflake distribution,

E. Heitz, J. Dupuy, C. Crassin, and C. Dachs- bacher, “The SGGX microflake distribution,”ACM Transactions on Graphics (Proceedings of SIG- GRAPH), vol. 34, pp. 48:1–48:11, July 2015

2015
[19]

OpenVDB: An open-source data struc- ture and toolkit for high-resolution volumes,

K. Museth, J. Lait, J. Johanson, J. Budsberg, R. Henderson, M. Alden, P. Cucka, D. Hill, and A. Pearce, “OpenVDB: An open-source data struc- ture and toolkit for high-resolution volumes,” in ACM SIGGRAPH Courses, ACM Press, 2013

2013
[20]

OpenVDB,

Academy Software Foundation, “OpenVDB,” Oct. 2024

2024
[21]

NanoVDB: A GPU-friendly and portable VDB data structure for real-time render- ing and simulation,

K. Museth, “NanoVDB: A GPU-friendly and portable VDB data structure for real-time render- ing and simulation,” inACM SIGGRAPH Talks, (New York, NY, USA), ACM Press, 2021

2021
[22]

NanoVDB,

Nvidia Corporation, “NanoVDB,” Oct. 2024

2024
[23]

3D line voxeliza- tion and connectivity control,

D. Cohen-Or and A. Kaufman, “3D line voxeliza- tion and connectivity control,”IEEE Computer Graphics & Applications, vol. 17, no. 6, pp. 80–87, 1997

1997
[24]

Incremental triangle voxelization,

F. D. IX and A. Kaufman, “Incremental triangle voxelization,” inProceedings of Graphics Interface, pp. 205–212, 2000

2000
[25]

3D scan-conversion algorithms for voxel-based graphics,

A. Kaufman and E. Shimony, “3D scan-conversion algorithms for voxel-based graphics,” inProceed- ings of the Workshop on Interactive 3D Graphics, pp. 45–75, 1987

1987
[26]

Fast scene voxeliza- tion and applications,

E. Eisemann and X. D´ ecoret, “Fast scene voxeliza- tion and applications,” inProceedings of the Sym- posium on Interactive 3D Graphics and Games, pp. 71–78, 2006

2006
[27]

Crassin,GigaVoxels: A Voxel-Based Rendering Pipeline for Efficient Exploration of Large and De- tailed Scenes

C. Crassin,GigaVoxels: A Voxel-Based Rendering Pipeline for Efficient Exploration of Large and De- tailed Scenes. PhD thesis, Universit´ e de Grenoble, July 2011

2011
[28]

Interactive indirect illumination us- ing voxel cone tracing,

C. Crassin, F. Neyret, M. Sainz, S. Green, and E. Eisemann, “Interactive indirect illumination us- ing voxel cone tracing,”Computer Graphics Forum (Proceedings of Pacific Graphics), vol. 30, pp. 207– 207, Sept. 2011

2011
[29]

Matching real fabrics with micro- appearance models,

P. Khungurn, D. Schroeder, S. Zhao, K. Bala, and S. Marschner, “Matching real fabrics with micro- appearance models,”ACM Transactions on Graph- ics, vol. 35, Dec. 2016

2016
[30]

Re- cent advances in fabric appearance reproduction,

C. Castillo, J. L´ opez-Moreno, and C. Aliaga, “Re- cent advances in fabric appearance reproduction,” Computers & Graphics, vol. 84, pp. 103–121, Nov. 2019

2019
[31]

Sparse GPU Voxelization of Yarn-Level Cloth: Sparse GPU Voxelization of Yarn-Level Cloth,

J. Lopez-Moreno, D. Miraut, G. Cirio, and M. A. Otaduy, “Sparse GPU Voxelization of Yarn-Level Cloth: Sparse GPU Voxelization of Yarn-Level Cloth,”Computer Graphics Forum, vol. 36, pp. 22– 34, Jan. 2017

2017
[32]

Photorealistic rendering of knitwear using the lumislice,

Y.-Q. Xu, Y. Chen, S. Lin, H. Zhong, E. Wu, B. Guo, and H.-Y. Shum, “Photorealistic rendering of knitwear using the lumislice,” inAnnual Con- ference Series (Proceedings of SIGGRAPH), (New York, NY, USA), p. 391–398, ACM Press, 2001

2001
[33]

Pyramidal parametrics,

L. Williams, “Pyramidal parametrics,”Computer Graphics (Proceedings of SIGGRAPH), vol. 17, pp. 1–11, July 1983

1983
[34]

LEAN mapping,

M. Olano and D. Baker, “LEAN mapping,” in Proceedings of the Symposium on Interactive 3D Graphics and Games, pp. 181–188, 2010. Fast Voxelization and Level of Detail for Microgeometry Rendering 13

2010
[35]

MIPNet: Neural normal-to-anisotropic-roughness MIP mapping,

A. Gauthier, R. Faury, J. Levallois, T. Thonat, J.- M. Thiery, and T. Boubekeur, “MIPNet: Neural normal-to-anisotropic-roughness MIP mapping,” ACM Transactions on Graphics (Proceedings of SIGGRAPH Asia), vol. 41, pp. 246:1–246:12, Nov. 2022

2022
[36]

Surface simpli- fication using quadric error metrics,

M. Garland and P. S. Heckbert, “Surface simpli- fication using quadric error metrics,” inAnnual Conference Series (Proceedings of SIGGRAPH), pp. 209–216, ACM Press, 1997

1997
[37]

A topology modifying progres- sive decimation algorithm,

W. J. Schroeder, “A topology modifying progres- sive decimation algorithm,” inProceedings. Visu- alization’97 (Cat. No. 97CB36155), pp. 205–212, IEEE Computer Society, 1997

1997
[38]

Model simplification us- ing vertex-clustering,

K.-L. Low and T.-S. Tan, “Model simplification us- ing vertex-clustering,” inProceedings of the 1997 Symposium on Interactive 3D Graphics, pp. 75–ff, 1997

1997
[39]

Modular primitives for high- performance differentiable rendering,

S. Laine, J. Hellsten, T. Karras, Y. Seol, J. Lehti- nen, and T. Aila, “Modular primitives for high- performance differentiable rendering,”ACM Trans- actions on Graphics (Proceedings of SIGGRAPH Asia), vol. 39, pp. 194:1–194:14, Nov. 2020

2020
[40]

Appearance-driven automatic 3D model simplification.,

J. Hasselgren, J. Munkberg, J. Lehtinen, M. Ait- tala, and S. Laine, “Appearance-driven automatic 3D model simplification.,” inProceedings of the Eurographics Symposium on Rendering (EGSR), pp. 85–97, 2021

2021
[41]

Locally- adaptive level-of-detail for hardware-accelerated ray tracing,

J. Haydel, C. Yuksel, and L. Seiler, “Locally- adaptive level-of-detail for hardware-accelerated ray tracing,”ACM Transactions on Graphics (Pro- ceedings of SIGGRAPH Asia), vol. 42, pp. 196:1– 196:15, Dec. 2023

2023
[42]

Neural prefiltering for correlation- aware levels of detail,

P. Weier, T. Zirr, A. Kaplanyan, L.-Q. Yan, and P. Slusallek, “Neural prefiltering for correlation- aware levels of detail,”ACM Transactions on Graphics (Proceedings of SIGGRAPH), vol. 42, pp. 78:1–78:16, July 2023

2023
[43]

Downsampling scattering parameters for ren- dering anisotropic media,

S. Zhao, L. Wu, F. Durand, and R. Ramamoor- thi, “Downsampling scattering parameters for ren- dering anisotropic media,”ACM Transactions on Graphics (Proceedings of SIGGRAPH Asia), vol. 35, pp. 166:1–166:11, Nov. 2016

2016
[44]

A new microflake model with microscopic self-shadowing for accurate vol- ume downsampling,

G. Loubet and F. Neyret, “A new microflake model with microscopic self-shadowing for accurate vol- ume downsampling,”Computer Graphics Forum (Proceedings of Eurographics), vol. 37, pp. 111–121, May 2018

2018
[45]

Appearance-preserving scene aggrega- tion for level-of-detail rendering,

Y. Zhou, T. Huang, R. Ramamoorthi, P. Sen, and L.-Q. Yan, “Appearance-preserving scene aggrega- tion for level-of-detail rendering,”ACM Transac- tions on Graphics, vol. 44, Jan. 2025

2025
[46]

Hybrid mesh-volume lods for all-scale pre-filtering of complex 3d assets,

G. Loubet and F. Neyret, “Hybrid mesh-volume lods for all-scale pre-filtering of complex 3d assets,” Computer Graphics Forum, vol. 36, no. 2, pp. 431– 442, 2017

2017
[47]

A non-exponential transmittance model for volumet- ric scene representations,

D. Vicini, W. Jakob, and A. Kaplanyan, “A non-exponential transmittance model for volumet- ric scene representations,”ACM Transactions on Graphics (Proceedings of SIGGRAPH), vol. 40, pp. 136:1–136:16, July 2021

2021
[48]

Neural geometric level of detail: Real-time rendering with implicit 3D shapes,

T. Takikawa, J. Litalien, K. Yin, K. Kreis, C. Loop, D. Nowrouzezahrai, A. Jacobson, M. McGuire, and S. Fidler, “Neural geometric level of detail: Real-time rendering with implicit 3D shapes,” IEEE/CVF International Conference on Computer Vision and Pattern Recognition (CVPR), Jan. 2021

2021
[49]

A class of local interpo- lating splines,

E. Catmull and R. Rom, “A class of local interpo- lating splines,” inComputer Aided Geometric De- sign(R. E. Barnhill and R. F. Riesenfeld, eds.), pp. 317–326, PubAP, 1974

1974
[50]

A Low-Distortion Map Between Triangle and Square

E. Heitz, “A Low-Distortion Map Between Triangle and Square.” working paper or preprint, 2019

2019
[51]

A radiative transfer framework for non-exponential media,

B. Bitterli, S. Ravichandran, T. M¨ uller, M. Wren- ninge, J. Nov´ ak, S. Marschner, and W. Jarosz, “A radiative transfer framework for non-exponential media,”ACM Transactions on Graphics (Proceed- ings of SIGGRAPH Asia), vol. 37, pp. 225:1– 225:17, Nov. 2018

2018
[52]

Fractional Gaussian fields for mod- eling and rendering of spatially-correlated media,

J. Guo, Y. Chen, B. Hu, L.-Q. Yan, Y. Guo, and Y. Liu, “Fractional Gaussian fields for mod- eling and rendering of spatially-correlated media,” ACM Transactions on Graphics (Proceedings of SIGGRAPH), vol. 38, pp. 45:1–45:13, July 2019

2019
[53]

Algorithms for hier- archical clustering: An overview,

F. Murtagh and P. Contreras, “Algorithms for hier- archical clustering: An overview,”Wiley Interdisci- plinary Reviews: Data Mining and Knowledge Dis- covery, vol. 2, no. 1, pp. 86–97, 2012

2012
[54]

Calculation of the Wasserstein distance between probability distributions on the line,

SS. Vallender, “Calculation of the Wasserstein distance between probability distributions on the line,”Theory of Probability & Its Applications, vol. 18, no. 4, pp. 784–786, 1974

1974
[55]

A sliced wasserstein loss for neural tex- ture synthesis,

E. Heitz, K. Vanhoey, T. Chambon, and L. Bel- cour, “A sliced wasserstein loss for neural tex- ture synthesis,” inIEEE/CVF International Con- ference on Computer Vision and Pattern Recogni- tion (CVPR), 2021

2021
[56]

Improved Training of Wasserstein GANs

I. Gulrajani, F. Ahmed, M. Arjovsky, V. Dumoulin, and A. Courville, “Improved training of wasserstein gans,”arXiv:1704.00028, 2017

work page Pith review arXiv 2017
[57]

A radiative transfer framework for rendering materials with anisotropic structure,

W. Jakob, A. Arbree, J. T. Moon, K. Bala, and S. Marschner, “A radiative transfer framework for rendering materials with anisotropic structure,” ACM Transactions on Graphics (Proceedings of 14 Javier Fabre et al. SIGGRAPH), vol. 29, pp. 53:1–53:13, July 2010

2010
[58]

Techniques used in the GEM code for Monte Carlo neutronics calculations in re- actors and other systems of complex geometry,

E. R. Woodcock, T. Murphy, P. J. Hemmings, and T. C. Longworth, “Techniques used in the GEM code for Monte Carlo neutronics calculations in re- actors and other systems of complex geometry,” inApplications of Computing Methods to Reactor Problems, Argonne National Laboratory, 1965

1965
[59]

Pharr, W

M. Pharr, W. Jakob, and G. Humphreys,Physically Based Rendering: From Theory to Implementation. Cambridge, MA: Morgan Kaufmann, 3 ed., 2016

2016
[60]

The unreasonable effectiveness of deep features as a perceptual metric,

R. Zhang, P. Isola, A. A. Efros, E. Shechtman, and O. Wang, “The unreasonable effectiveness of deep features as a perceptual metric,” inIEEE/CVF In- ternational Conference on Computer Vision and Pattern Recognition (CVPR), pp. 586–595, 2018

2018
[61]

FLIP: A Difference Evaluator for Alternating Images,

P. Andersson, J. Nilsson, T. Akenine-M¨ oller, M. Os- karsson, K. ˚Astr¨ om, and M. D. Fairchild, “FLIP: A Difference Evaluator for Alternating Images,”Pro- ceedings of the ACM on Computer Graphics and Interactive Techniques, vol. 3, no. 2, pp. 15:1–15:23, 2020. Fast Voxelization and Level of Detail for Microgeometry Rendering 15 Ground Truth Averages SGG...

2020