Pith. sign in

REVIEW 3 major objections 4 minor 190 references

A Hybrid Neural-Microfacet BRDF Model for Real-Time Rendering

T0 review · 3 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A small MLP correcting a GGX microfacet model beats fully neural BRDFs at equal memory.

desk verdict A useful hybrid BRDF model with a plausible small-MLP win, but the headline comparison relies on a modified unreleased baseline and thinner margins than the abstract implies. read the letter →

arxiv 2608.09604 v1 pith:ZZ4BG4OD submitted 2026-08-10 cs.GR cs.CV

Louis De Oliveira , Anastasia Karpova , Georges Nader , Antoine Houdard , Pierre Mezieres , Damien Rioux-Lavoie , Romain Pacanowski This is my paper · ORCID
classification cs.GRcs.CV
keywords BRDFmicrofacetmodelGGXneuralmeasuredreflectancereal-timerenderingappearanceeditingimportancesampling
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 show that a BRDF model does not need to be fully neural to beat fully neural models: combining a standard GGX microfacet base with a small shared neural network that only learns residual corrections gives better fits to measured materials than a same-size neural network alone. The authors fit one 14 kB MLP plus 12–16 per-material parameters to 312 measured BRDFs from three public databases. At identical memory cost, the hybrid model achieves lower BRDF reconstruction error (MERL SMAPE 0.1974 vs 0.2015) and lower rendering error (MAE 0.0051 vs 0.0062) than the neural baseline; it also keeps the analytic parameters editable and supplies importance sampling for free. If true, this gives real-time renderers a practical way to get neural-level appearance without a large network or a dedicated sampling network.

What carries the argument

The model is defined by $f_t(\omega_i,\omega_o)=f_c(\omega_i,\omega_o,z)+f_g(\omega_i,\omega_o,z)\,f_a(\omega_i,\omega_o,p)$, where $f_a$ is the Lambertian-plus-GGX analytical term with per-material parameters $p=(k_d,\eta,\alpha)$, and a single shared MLP $\psi_w$ emits the non-negative residual $f_c$ and the gate $f_g\in[0,1]^3$ from input directions and a per-material latent code $z$. Training minimizes $\sum_i \sum_\omega \mathcal{L}_a(f^i, f_a) + \mathcal{L}_t(f^i, f_t)$ with $\mathcal{L}_{a,t}(f_1,f_2)=\|\log(1+\cos\theta_i f_1)-\log(1+\cos\theta_i f_2)\|_2$, so the analytic term is supervised to stay close to the target while the network learns only residual effects. GGX is the microfacet normal-distribution model that provides the dominant reflectance lobe, the per-material parameterization, and the sampling distribution used with multiple importance sampling.

What would settle it

Run the original neural appearance model with its encoder intact at the same 12 per-material parameters and 32×3 MLP on the same 100 MERL BRDFs. If its average SMAPE falls to or below 0.1974, or its rendering MAE to or below 0.0051, the paper's central claim is overturned.

Watch

Extended reading notes

Core claim

The paper's central claim is that a composite BRDF model—a Lambertian-plus-GGX analytical base with per-material parameters, plus a single small shared MLP that outputs a multiplicative gate and an additive residual—can fit measured materials better than a fully neural model of the same memory footprint. On 100 MERL isotropic materials, the hybrid model with a 32×3 MLP and 12 total per-material parameters reaches SMAPE 0.1974 versus 0.2015 for the neural baseline, and rendering MAE 0.0051 versus 0.0062; across the 100 MERL materials the hybrid wins on 88 of them. The same pattern holds on the RGL and UTIA datasets, with the gap shrinking as network size grows, because a large network can approximate the BRDF on its own. Because the analytical component is trained to stay close to the reference, the neural part carries only residual effects, which keeps the analytic parameters meaningful for editing and lets the GGX term serve as an importance-sampling proxy without extra overhead.

Load-bearing premise

The central comparison assumes that the paper's re-implementation of the leading neural model, with its encoder replaced by a simple per-material code, is just as strong as the original; if that replacement weakens the baseline, the reported margin could shrink or reverse.

Editorial extensions

If this is right

  • A renderer can replace a fully neural material network with one 14 kB shared MLP plus 12–16 per-material parameters and get lower reconstruction error at equal memory.
  • Because the analytic GGX term remains close to the fitted BRDF, importance sampling can use standard microfacet and cosine sampling with MIS, approaching tabulated reference efficiency without any additional network evaluation.
  • After fitting, editing diffuse albedo, roughness, or index of refraction produces coherent new appearances because the network only corrects residual effects, and interpolating between two materials gives plausible intermediate results.
  • Neural inference can be restricted to the first one or two path-tracing bounces without visible loss, cutting path-tracing frame time from 5.57 ms to about 1.9 ms in the paper's test scene (and from 2.94 ms to 1.6 ms with cooperative-vector acceleration).
  • New materials can be added after training by optimizing only the per-material latent code and analytic parameters, with rendering MAE going from 0.0087 to 0.0099 on MERL, so the shared network acts as a material prior.

Reading between the lines

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

  • The same residual architecture could be applied to richer analytic bases—multi-lobe, layered, or diffraction-aware models—which would likely recover iridescence and similar effects that the single GGX lobe misses, at the cost of more per-material parameters.
  • A natural next experiment is SVBRDF acquisition: fitting per-texel latent and analytic parameter maps with the shared network fixed, then aggressively quantizing and clustering the analytic parameters, could make noisy multi-view captures more robust than either a fully analytic or fully neural fit.
  • Because latent-space edits are not guaranteed to stay BRDF-like, the model's editability is probably bounded to a neighborhood of the training distribution; a regularization term encouraging a structured latent space would be the direct test.
  • The selective-evaluation result hints at a simple runtime LOD rule—skip the neural correction for distant or rough surfaces—that could be validated on game-like scenes and would make the model attractive on hardware without neural acceleration.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper proposes a hybrid BRDF model that combines a GGX microfacet component with a small shared MLP that predicts a multiplicative gate and an additive residual. The model is fitted jointly over a collection of measured BRDFs by optimizing per-material analytical parameters (kd, eta, alpha) and latent codes together with the shared network weights. The authors evaluate on MERL, RGL, and UTIA datasets, comparing BRDF-space SMAPE and render-space MAE against a modified version of the neural appearance model of Zeltner et al. [ZRW*24]. They also present ablations, an importance-sampling scheme based on the analytical component, and a selective-evaluation strategy that disables the MLP after a chosen bounce depth. The main claim is that at equal per-material memory cost, the hybrid model outperforms the fully neural baseline for small MLP sizes.

Significance. If the central comparison held up, the paper would make a useful practical contribution: a small hybrid network plus analytically meaningful parameters that beats a same-size fully neural latent model, with editability and importance sampling as added benefits. The formulation in Eq. (1)-(3) is clean, the joint training loss in Eq. (4) is a sensible mechanism for keeping the analytical component meaningful, and the ablations in Table 2, the importance-sampling results in Figures 9-10, and the selective-evaluation tables are informative. However, the strength of the main claim is currently limited by three issues: the baseline is a modified implementation of the prior work; the reported margins are small and no variance information is given; and some tabulated results (RGL 16x3 and 64x3) run counter to the headline. These are fixable with additional experiments and a more carefully qualified claim.

major comments (3)
  1. [Section 4.1, Table 1, Abstract] The abstract's claim that 'at identical memory cost, our model approximates measurements better than state-of-the-art neural models' is broader than the reported data. Table 1 shows that for the RGL dataset at 16x3 and 64x3 MLP sizes (12-parameter configuration), the neural baseline achieves lower render MAE (0.0256 vs 0.0267 and 0.0143 vs 0.0149, respectively). The sentence in Section 4.1 stating that 'with small MLPs (16x3 and 32x3), our model consistently outperforms the neural model' is therefore inaccurate for RGL 16x3. The claim should be qualified to the datasets and configurations where it holds, and the RGL exceptions should be discussed rather than glossed over.
  2. [Section 4.1, comparison protocol] The neural baseline is not the original [ZRW*24] system; the authors state that they 'replace the encoder part with a trainable latent code' to equalize memory. No code, trained weights, or detailed training hyperparameters (learning rate, schedule, number of steps, loss weighting) for this modified baseline are provided. Because the reported margins are small (MERL SMAPE 0.1974 vs 0.2015; render MAE 0.0051 vs 0.0062), the central comparison rests on the unverified premise that this modified baseline is a faithful and well-tuned representative of the state of the art. The authors should either compare against the original implementation, release the modified baseline code and weights, or include a validation that the modified baseline reproduces the original method's accuracy on a shared subset.
  3. [Section 4.1, statistical robustness] The paper reports single-run results for every configuration. The observed differences are often small (e.g., MERL SMAPE 0.1974 vs 0.2015; RGL 16x3 MAE 0.0267 vs 0.0256 in the opposite direction), and training involves stochastic mini-batch optimization, so run-to-run variation could be comparable to the reported gaps. The authors should report means and standard deviations over multiple training runs (at least for the flagship 32x3 configuration) and indicate whether the differences are consistent across seeds. Without this information, the reader cannot assess whether the claimed advantage is systematic or within the noise.
minor comments (4)
  1. [Section 3.3, Eq. (4)] The summation in Eq. (4) is written as a sum over 'omega in B', but B is defined as a set of direction pairs (omega_i, omega_o) in the previous sentence. The notation should be clarified to sum over pairs of directions, e.g., sum_{(omega_i, omega_o) in B}.
  2. [Table 2 and Abstract] The value 0.2015 appears both as the MERL SMAPE for the 'Analytical Input' variant in Table 2 and as the neural baseline's MERL SMAPE in the abstract for the same 32x3 configuration. Please verify that this is not a copy/paste error and, if it is a coincidence, state so explicitly to avoid confusion.
  3. [Section 5.2, Figure 10] In the text preceding Figure 10, 'Figures9 and 10' is missing a space. In addition, the caption of Figure 10 would benefit from a definition of the plotted variance quantity and the reference sampling strategy.
  4. [Section 3.3, training details] The training description gives the optimizer, learning rate, and gradient clipping, but does not specify how materials are selected within a mini-batch (e.g., whether each batch contains all N BRDFs or a subsample), nor how the per-material parameters are initialized. Adding these details would improve reproducibility.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the accuracy claims are measured fits against external BRDF data, with only a minor non-load-bearing self-citation for shader-evaluation feasibility.

full rationale

The paper's central claim—that at identical memory cost the hybrid model approximates measured BRDFs better than a fully neural baseline—is supported by direct optimization against external measured data (MERL, UTIA, RGL) and evaluated with SMAPE and rendering MAE. The hybrid model is defined by Eq. (1), and the fitting loss (4) jointly supervises both the analytical and neural terms; no quantity used in the headline comparison is defined in terms of the claim itself. The only self-citations ([WDOHN24] for fragment-shader MLP feasibility and [BB25] for cooperative vectors) are performance/implementation plausibility arguments and do not enter the accuracy comparison, so they are not load-bearing. The replacement of the [ZRW*24] encoder with a trainable latent code to equalize memory (Section 4.1) is a benchmarking assumption rather than a circular reduction: the comparison is an empirical measurement against a modified external baseline, and any concern about baseline fidelity belongs to correctness risk, not circularity. No equation reduces to its own inputs, and no fitted parameter is renamed as a prediction. Score 1 reflects a minor, non-load-bearing self-citation.

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

No new physical entities are introduced. The fitted quantities are the per-material latent codes, GGX parameters, shared network weights, and training hyperparameters; none of these is an independent physical constant. The assumptions are the typical domain assumptions of data-driven BRDF fitting.

free parameters (4)
  • per-material latent code z = dim 4 (isotropic), 8 (anisotropic)
    Optimized per material during training; each BRDF is reduced to this code plus the analytic parameters.
  • per-material analytical parameters p = kd in [0,1]^3, eta in [1,10]^3, alpha in [0,1]^2
    Fitted jointly with the network; these are the meaningful GGX parameters and one of the two axes compared at equal memory.
  • shared MLP weights w = e.g., 14 kB for a 32x3 MLP
    A single network is trained over all BRDFs in a dataset and is part of the memory budget; its size is a core design variable.
  • training hyperparameters = lr 0.005, cosine decay, 200k steps, batch 1024, grad clip 0.01
    Chosen by tuning; they affect the fitted quality and are not derived from theory.
assumptions (4)
  • domain assumption Measured BRDFs from MERL, UTIA, and RGL are accurate ground truth.
    All fitting and evaluation assume the acquisition of these databases is reliable; acquisition noise would bias both the training targets and the reported errors (Section 4).
  • ad hoc to paper The GGX single-lobe model with Lambertian diffuse is an adequate analytic prior for most target materials.
    The model architecture fixes fa to this form; materials outside this class (e.g., iridescence, Figure 12) degrade, showing that the whole approach leans on this assumption (Sections 3.2 and 6).
  • domain assumption The chosen loss (log-cosine) and metrics (SMAPE, MAE) reflect the quality that matters for rendering.
    The authors optimize and then evaluate on these functions without perceptual validation; a different metric could change rankings (Sections 3.3 and 4.1).
  • ad hoc to paper A low-dimensional latent code plus a shared MLP can capture the residual appearance differences across the collection.
    The paper does not prove this capacity; it is empirically validated by the fitting results, and its failure on iridescent materials is acknowledged (Section 6).

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Hybrid Neural-Microfacet BRDF Model for Real-Time Rendering." pith.science (2026). https://pith.science/paper/ZZ4BG4OD

@misc{pith2026260809604,
  author       = {Pith},
  title        = {Pith review of: A Hybrid Neural-Microfacet BRDF Model for Real-Time Rendering},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZZ4BG4OD}},
  note         = {Machine review of arXiv:2608.09604}
}
read the original abstract

Over the past decade, microfacet-based BRDF models have formed the foundation of real-time rendering pipelines. Despite their widespread use, they often fail to reproduce subtle appearance effects arising from complex light-surface interactions, which have led to the emergence of specialized physics-based models for specific optical phenomena (e.g., diffraction, iridescence, multilayers). Although more accurate, these models lose versatility and lack performance for real-time rendering. Recently introduced, neural models have demonstrated their ability to approximate BRDF reference data coming from measurements, simulations, or even complex shading networks. However, most current neural models require relatively large networks, making them costly for real-time rendering. In this paper, we introduce a hybrid model that combines a GGX-type microfacet model and a neural model to leverage the best features of both representations. The neural component corrects the appearance approximated by the microfacet component, allowing much smaller network than in existing neural models. We show that, at identical memory cost, our model approximates measurements better than state-of-the-art neural models for a low evaluation overhead compared to a microfacet-based model. Furthermore, our hybrid model remains easily editable by artists and benefits from an important sampling scheme, making it attractive for both offline and real-time rendering.

Figures

Figures reproduced from arXiv: 2608.09604 by the authors.

Figure 1
Figure 1. Our hybrid model is fitted from measured BRDFs and represented as a compact set of microfacet parameters (kd,η,α) and a low￾dimensional latent code z shared across a single neural network. For each material, the neural correction improves upon the microfacet model alone, capturing subtle appearance effects, while the analytical component remains a faithful and useful approximation. From left to right irid_flake_pain… view at source ↗
Figure 2
Figure 2. Overview of our method. We jointly optimize a latent code z and analytical parameters p = (kd,η,α) for each BRDF and a shared neural network. The output of the neural network (here, a 32 × 3 MLP) is split into two parts: the correction fc(ωi ,ωo, z) ∈ R 3 + with an exponential activation and the gate fg(ωi ,ωo, z) ∈ [0,1] 3 with a sigmoid activation. These outputs are combined with the evaluation of the analytical m… view at source ↗
Figure 3
Figure 3. Impact of the analytical loss La term. Top-part: with full loss (4), fa remains close to the target reflectance. Bottom-part: without La, fa is allowed to deviate from the target. This impacts the overall quality and weakens the editability, as the analytic pa￾rameters p no longer match the final appearance. Tested Material: alumn-bronze from MERL. z for the selected BRDF. The last linear layer is split into two par… view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Average SMAPE on BRDF reconstruction for the 3 datasets for multiple MLP size. Since the analytical part of our hybrid model requires 7/8 parameters, we compare it to a fully neural model with 8 additional latent dimensions to provide a comparison with the same number …
Figure 5
Figure 5. Figure 5: Per BRDF SMAPE on the MERL database, which consists of 100 measured materials. Our model with the 32×3 MLP and a 4D latent space is compared to the Neural model [ZRW∗ 24] with a 32 × 3 MLP and a 12D latent space, as well as a microfacet-based GGX model. Of the 100 mate…
Figure 7
Figure 7. Figure 7: Editing the satin-gold BRDF from the RGL dataset by modifying the fitted parameters (roughness α, albedo kd, IOR η) of the analytical component of our model. ψ were trained on and the others. When training on only 50% of the MERL BRDFs, we observe a moderate degradatio…
Figure 6
Figure 6. Figure 6: Per-material rendering comparisons between our model, its underlying analytical component, and the neural model. The in￾sets show the per-pixel absolute error as well as the color-averaged MAE for the whole image. The number of parameters (12) and the MLP size (32 × 3)…
Figure 8
Figure 8. Figure 8: BRDF interpolation example. From left to right, we linearly interpolate (parameter t) between two materials from the RGL dataset, the latent vector, and the analytical parameters of our hybrid model, showing very plausible results [PITH_FULL_IMAGE:figures/full_fig_p00…
Figure 9
Figure 9. Figure 9: Importance Sampling (IS) comparisons in image space for the red-phenolic material (MERL). From left to right: IS based on the GGX distribution and a cosine lobe distribution, tabulated IS (similarly to [LRR04]), with a memory footprint of 5.62 MB, the cosine lobe IS, a…
Figure 10
Figure 10. Figure 10: Color-averaged variance of the absolute difference be￾tween an image rendered with importance sampling and a reference image computed with 16k samples. Analytical Importance Sam￾pling corresponds to applying either cosine sampling or GGX sam￾pling using a MIS procedur…
Figure 11
Figure 11. Figure 11: Path-Tracing performance evaluation. All materials of the scene’s objects are rendered using our hybrid microfacet-neural model, which is fitted from three datasets. We show the effect in terms of MAE and rendering time (green and yellow curves show the time per sampl…
Figure 12
Figure 12. Figure 12: Our model struggles to reproduce complex phenomena such as iridescence (from cc_amber_citrine from RGL) accu￾rately without using a large neural network. At the bottom of each image is displayed, (Left) the color-averaged MAE and (Right) a map of the absolute error. […

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

190 extracted references · 57 canonical work pages

  1. [1]

    2023 , month = dec, journal =

    Bangaru, Sai Praveen and Wu, Lifan and Li, Tzu-Mao and Munkberg, Jacob and Bernstein, Gilbert and. 2023 , month = dec, journal =. doi:10.1145/3618353 , urldate =

    work page doi:10.1145/3618353 2023

  2. [2]

    2021 , month = aug, journal =

    An Inverse Method for the Exploration of Layered Material Appearance , author =. 2021 , month = aug, journal =. doi:10.1145/3450626.3459857 , urldate =

    arXiv 2021

  3. [3]

    Interpretable Disentangled Parametrization of Measured BRDF with $\beta$-VAE

    Benamira, Alexis and Shah, Sachin and Pattanaik, Sumanta , year =. Interpretable. doi:10.48550/arXiv.2208.03914 , urldate =. arXiv , langid =:2208.03914 , primaryclass =

    work page Pith review arXiv doi:10.48550/arxiv.2208.03914

  4. [5]

    Multi-View

    Bruneau, Robin and Brument, Baptiste and Qu. Multi-View. 2025 , month = jun, number =. doi:10.48550/arXiv.2506.04115 , urldate =. arXiv , keywords =:2506.04115 , primaryclass =

    work page doi:10.48550/arxiv.2506.04115 2025

  5. [6]

    Physically-

    Burley, Brent , year =. Physically-

  6. [7]

    2025 , month = jul, number =

    Weighted. 2025 , month = jul, number =. doi:10.48550/arXiv.2507.22270 , urldate =. arXiv , keywords =:2507.22270 , primaryclass =

    work page doi:10.48550/arxiv.2507.22270 2025

  7. [8]

    Deep Image-Based

    Cao, Wen , year =. Deep Image-Based. Proceedings of the 20th. doi:10.5220/0013201000003912 , urldate =. arXiv , keywords =:2410.02917 , primaryclass =

    arXiv

  8. [9]

    Lightplane:

    Cao, Ang and Johnson, Justin and Vedaldi, Andrea and Novotny, David , year =. Lightplane:. doi:10.48550/arXiv.2404.19760 , urldate =. arXiv , langid =:2404.19760 , primaryclass =

    arXiv doi:10.48550/arxiv.2404.19760

Show all 190 references
  1. [10]

    Cao, Xu and Taketomi, Takafumi , year =. Neural. doi:10.48550/arXiv.2507.23162 , urldate =. arXiv , langid =:2507.23162 , primaryclass =

    doi arXiv

  2. [11]

    doi:10.48550/arXiv.2508.17811 , urldate =

    Chang, Hanzhi and Zhu, Ruijie and Chang, Wenjie and Yu, Mulin and Liang, Yanzhe and Lu, Jiahao and Li, Zhuoyuan and Zhang, Tianzhu , year =. doi:10.48550/arXiv.2508.17811 , urldate =. arXiv , keywords =:2508.17811 , primaryclass =

    doi

  3. [12]

    , year =

    Chen, Guanying and Han, Kai and Shi, Boxin and Matsushita, Yasuyuki and Wong, Kwan-Yee K. , year =. Deep. doi:10.48550/arXiv.2007.13145 , urldate =. arXiv , langid =:2007.13145 , primaryclass =

    2007 doi arXiv

  4. [13]

    doi:10.48550/arXiv.2507.18231 , urldate =

    Chen, Yixiao and Liang, Bin and Guo, Hanzhi and Cheng, Yongqing and Zhao, Jiayi and Weng, Dongdong , year =. doi:10.48550/arXiv.2507.18231 , urldate =. arXiv , langid =:2507.18231 , primaryclass =

    doi arXiv

  5. [14]

    Differentiable

    Chung, Hoon-Gyu and Choi, Seokjun and Baek, Seung-Hwan , year =. Differentiable. doi:10.48550/arXiv.2411.17994 , urldate =. arXiv , langid =:2411.17994 , primaryclass =

    doi arXiv

  6. [15]

    Differentiable

    Chung, Hoon-Gyu and Choi, Seokjun and Baek, Seung-Hwan , year =. Differentiable. doi:10.48550/arXiv.2312.02480 , urldate =. arXiv , langid =:2312.02480 , primaryclass =

    doi arXiv

  7. [16]

    Cook, R. L. and Torrance, K. E. , year =. A. ACM Transactions on Graphics , volume =. doi:10.1145/357290.357293 , urldate =

  8. [17]

    Coomans, Arno and Dominci, Edoardo A. and D. Real-time. 2024 , month = may, journal =. doi:10.1111/cgf.15014 , urldate =

    2024 doi

  9. [18]

    doi:10.48550/arXiv.2506.20979 , urldate =

    Dai, Weichen and Ma, Kangcheng and Wang, Jiaxin and Pan, Kecen and Ming, Yuhang and Zhang, Hua and Kong, Wanzeng , year =. doi:10.48550/arXiv.2506.20979 , urldate =. arXiv , keywords =:2506.20979 , primaryclass =

    doi arXiv

  10. [19]

    2022 , month = may, journal =

    Rendering. 2022 , month = may, journal =. doi:10.1145/3522620 , urldate =

    2022 doi

  11. [21]

    Deschaintre, Valentin and Aittala, Miika and Durand, Fredo and Drettakis, George and Bousseau, Adrien , year =. Single-. ACM Transactions on Graphics , volume =. doi:10.1145/3197517.3201378 , urldate =

  12. [22]

    Dou, Yishun and Zheng, Zhong and Jin, Qiaoqiao and Ni, Bingbing and Chen, Yugang and Ke, Junxiang , year =. Real-. 2024. doi:10.1109/cvpr52733.2024.00415 , urldate =

    2024

  13. [24]

    Fan, Jiahui and Wang, Beibei and Ha. Neural. 2021 , month = nov, number =. doi:10.48550/arXiv.2111.03797 , urldate =. arXiv , keywords =:2111.03797 , primaryclass =

    2021 doi arXiv

  14. [27]

    and Gruson, Adrien and Nowrouzezahrai, Derek , year =

    Gamboa, Luis E. and Gruson, Adrien and Nowrouzezahrai, Derek , year =. An. Computer Graphics Forum , volume =. doi:10.1111/cgf.13936 , urldate =

    doi

  15. [28]

    doi:10.48550/arXiv.2507.15602 , urldate =

    Gao, Zihui and Bian, Jia-Wang and Lin, Guosheng and Chen, Hao and Shen, Chunhua , year =. doi:10.48550/arXiv.2507.15602 , urldate =. arXiv , keywords =:2507.15602 , primaryclass =

    doi arXiv

  16. [29]

    Implicit

    Genest, B and Gueth, P and Levallois, J and Wang, S , abstract =. Implicit

  17. [30]

    doi:10.48550/arXiv.2412.07371 , urldate =

    Ge, Wenhang and Lin, Jiantao and Shen, Guibao and Feng, Jiawei and Hu, Tao and Xu, Xinli and Chen, Ying-Cong , year =. doi:10.48550/arXiv.2412.07371 , urldate =. arXiv , keywords =:2412.07371 , primaryclass =

    doi arXiv

  18. [31]

    Hypernetworks for

    Gokbudak, Fazilet and Sztrajman, Alejandro and Zhou, Chenliang and Zhong, Fangcheng and Mantiuk, Rafal and Oztireli, Cengiz , year =. Hypernetworks for. Computer. doi:10.1007/978-3-031-73116-7_5 , urldate =

    doi

  19. [32]

    Hypernetworks for

    Gokbudak, Fazilet and Sztrajman, Alejandro and Zhou, Chenliang and Zhong, Fangcheng and Mantiuk, Rafal and Oztireli, Cengiz , year =. Hypernetworks for. doi:10.48550/arXiv.2311.15783 , urldate =. arXiv , keywords =:2311.15783 , primaryclass =

    doi arXiv

  20. [33]

    doi:10.48550/arXiv.2504.17728 , urldate =

    Gong, Shucheng and Zhao, Lingzhe and Li, Wenpu and Xie, Hong and Zhang, Yin and Zhao, Shiyu and Liu, Peidong , year =. doi:10.48550/arXiv.2504.17728 , urldate =. arXiv , langid =:2504.17728 , primaryclass =

    doi arXiv

  21. [34]

    Uncertainty

    Gottwald, Thomas and Heinert, Edgar and Rottmann, Matthias , year =. Uncertainty. doi:10.48550/arXiv.2508.02443 , urldate =. arXiv , langid =:2508.02443 , primaryclass =

    doi

  22. [35]

    Govindarajan, Shrisudhan and Rebain, Daniel and Yi, Kwang Moo and Tagliasacchi, Andrea , year =. Radiant. doi:10.48550/arXiv.2502.01157 , urldate =. arXiv , keywords =:2502.01157 , primaryclass =

    doi arXiv

  23. [36]

    Gaussian

    Gu. Gaussian. 2024 , month = mar, number =. arXiv , keywords =:2403.14554 , primaryclass =

    2024 arXiv

  24. [37]

    and Bagautdinov, Timur and Baque, Pierre and Fua, Pascal , year =

    Guillard, Benoit and Remelli, Edoardo and Lukoianov, Artem and Richter, Stephan R. and Bagautdinov, Timur and Baque, Pierre and Fua, Pascal , year =. doi:10.48550/arXiv.2106.11795 , urldate =. arXiv , keywords =:2106.11795 , primaryclass =

    doi arXiv

  25. [38]

    doi:10.48550/arXiv.2504.17062 , urldate =

    Guo, Yu and Lao, Zhiqiang and Song, Xiyun and Zhou, Yubin and Lin, Zongfang and Yu, Heather , year =. doi:10.48550/arXiv.2504.17062 , urldate =. arXiv , langid =:2504.17062 , primaryclass =

    doi arXiv

  26. [40]

    Neural. Lecture. 2024 , pages =. doi:10.1007/978-3-031-50069-5_9 , urldate =

    2024 doi

  27. [41]

    Polarimetric

    Ha, Hyunho and Hwang, Inseung and Monzon, Nestor and Cho, Jaemin and Kim, Donggun and Baek, Seung-Hwan and Mu. Polarimetric. 2024 , month = dec, journal =. doi:10.1145/3687767 , urldate =

    2024 doi

  28. [42]

    He, Zexin and Wang, Tengfei and Huang, Xin and Pan, Xingang and Liu, Ziwei , year =. Neural. doi:10.48550/arXiv.2412.09593 , urldate =. arXiv , keywords =:2412.09593 , primaryclass =

    doi arXiv

  29. [43]

    Hofherr, Florian and Haefner, Bjoern and Cremers, Daniel , year =. On. Proceedings - 2025 IEEE Winter Conference on Applications of Computer Vision, WACV 2025 , series =. doi:10.1109/WACV61041.2025.00181 , urldate =

    2025

  30. [44]

    Computer Graphics Forum , volume =

    Hu, Bingyang and Guo, Jie and Chen, Yanjun and Li, Mengtian and Guo, Yanwen , year =. Computer Graphics Forum , volume =. doi:10.1111/cgf.13920 , urldate =

    doi

  31. [45]

    2022 , month = jul, journal =

    Jakob, Wenzel and Speierer, S. 2022 , month = jul, journal =. doi:10.1145/3528223.3530099 , urldate =

    2022

  32. [46]

    Jakob, Wenzel , urldate =. A

  33. [47]

    Mitsuba 3 Renderer , author =

  34. [48]

    and Levoy, Marc and Hanrahan, Pat , year =

    Jensen, Henrik Wann and Marschner, Stephen R. and Levoy, Marc and Hanrahan, Pat , year =. A Practical Model for Subsurface Light Transport , booktitle =. doi:10.1145/383259.383319 , urldate =

  35. [49]

    2025 , month = apr, number =

    A. 2025 , month = apr, number =. doi:10.48550/arXiv.2504.15028 , urldate =. arXiv , langid =:2504.15028 , primaryclass =

    2025 doi arXiv

  36. [50]

    Ju, Yakun and Lam, Kin-Man and Xie, Wuyuan and Zhou, Huiyu and Dong, Junyu and Shi, Boxin , year =. Deep. doi:10.48550/arXiv.2212.08414 , urldate =. arXiv , langid =:2212.08414 , primaryclass =

    doi arXiv

  37. [51]

    2023 , month = oct, journal =

    Unified Shape and Appearance Reconstruction with Joint Camera Parameter Refinement , author =. 2023 , month = oct, journal =. doi:10.1016/j.gmod.2023.101193 , urldate =

    2023

  38. [52]

    Karis, Brian , langid =. Real

  39. [53]

    Differentiable

    Kato, Hiroharu and Beker, Deniz and Morariu, Mihai and Ando, Takahiro and Matsuoka, Toru and Kehl, Wadim and Gaidon, Adrien , year =. Differentiable. arXiv , keywords =:2006.12057 , primaryclass =

    2006 arXiv

  40. [54]

    Perceived Quality of

    Kavoosighafi, Behnaz and Mantiuk, Rafal and Hajisharif, Saghi and Miandji, Ehsan and Unger, Jonas , year =. Perceived Quality of. doi:10.17863/CAM.118953 , urldate =

    doi

  41. [55]

    Poisson Surface Reconstruction , booktitle =

    Kazhdan, Michael and Bolitho, Matthew and Hoppe, Hugues , editor =. Poisson Surface Reconstruction , booktitle =. 2006 , publisher =. doi:/10.2312/SGP/SGP06/061-070 , isbn =

    2006 doi

  42. [56]

    ACM Transactions on Graphics , volume =

    Kerbl, Bernhard and Kopanas, Georgios and Leimkuehler, Thomas and Drettakis, George , year =. ACM Transactions on Graphics , volume =. doi:10.1145/3592433 , urldate =

    doi

  43. [57]

    and Welling, Max , year =

    Kingma, Diederik P. and Welling, Max , year =. Auto-. doi:10.48550/arXiv.1312.6114 , urldate =. arXiv , keywords =:1312.6114 , primaryclass =

    doi arXiv

  44. [58]

    Intrinsic

    Koestler, Lukas and Grittner, Daniel and Moeller, Michael and Cremers, Daniel and L. Intrinsic. Computer. 2022 , volume =. doi:10.1007/978-3-031-20086-1_36 , urldate =

    2022 doi

  45. [59]

    Provably

    Kolluri, Ravikrishna , abstract =. Provably

  46. [60]

    2021 , month = apr, number =

    Kuznetsov, Alexandr and Mullia, Krishna and Xu, Zexiang and Ha. 2021 , month = apr, number =. arXiv , langid =:2104.02789 , primaryclass =

    2021 arXiv

  47. [61]

    Rendering

    Kuznetsov, Alexandr and Wang, Xuezheng and Mullia, Krishna and Luan, Fujun and Xu, Zexiang and Hasan, Milos and Ramamoorthi, Ravi , year =. Rendering. Special. doi:10.1145/3528233.3530721 , urldate =

  48. [62]

    Computer Graphics Forum , volume =

    Rainer, Gilles and Jakob, Wenzel and Ghosh, Abhijeet and Weyrich, Tim , title =. Computer Graphics Forum , volume =. doi:https://doi.org/10.1111/cgf.13633 , abstract =

    doi

  49. [63]

    Computer Graphics Forum , volume =

    Rainer, Gilles and Ghosh, Abhijeet and Jakob, Wenzel and Weyrich, Tim , title =. Computer Graphics Forum , volume =. doi:https://doi.org/10.1111/cgf.13921 , abstract =

    doi

  50. [64]

    2020 , journal =

    Modular Primitives for High-Performance Differentiable Rendering , author =. 2020 , journal =

    2020

  51. [65]

    Navigating the

    Lanza, Dario and Masia, Belen and Jarabo, Adrian , year =. Navigating the. Computer Graphics Forum , volume =. doi:10.1111/cgf.15035 , urldate =

    doi

  52. [66]

    ArXiv , year=

    Hardware Accelerated Neural Block Texture Compression with Cooperative Vectors , author=. ArXiv , year=

  53. [67]

    doi:10.48550/arXiv.2508.01312 , urldate =

    Lee, Seong Hun and Vandewalle, Patrick and Civera, Javier , year =. doi:10.48550/arXiv.2508.01312 , urldate =. arXiv , langid =:2508.01312 , primaryclass =

    doi arXiv

  54. [68]

    2003 , month = apr, journal =

    Image-Based Reconstruction of Spatial Appearance and Geometric Detail , author =. 2003 , month = apr, journal =. doi:10.1145/636886.636891 , urldate =

    2003

  55. [69]

    Differentiable

    Li, Tzu-Mao and Aittala, Miika and Durand, Fr. Differentiable. 2018 , month = dec, journal =. doi:10.1145/3272127.3275109 , urldate =

    2018

  56. [70]

    Light of

    Li, Hong and Chen, Houyuan and Ye, Chongjie and Chen, Zhaoxi and Li, Bohan and Xu, Shaocong and Guo, Xianda and Liu, Xuhui and Wang, Yikai and Zhang, Baochang and Ikehata, Satoshi and Shi, Boxin and Rao, Anyi and Zhao, Hao , year =. Light of. doi:10.48550/arXiv.2506.18882 , ur...

    doi

  57. [71]

    Stochastic

    Ling, Selena and. Stochastic. 2025 , month = may, eprint =. doi:10.1145/3731161 , urldate =

    2025 doi

  58. [73]

    Lipman, Yaron and Chen, Ricky T. Q. and. Flow. 2023 , month = feb, number =. doi:10.48550/arXiv.2210.02747 , urldate =. arXiv , keywords =:2210.02747 , primaryclass =

    2023 doi arXiv

  59. [74]

    Lipman, Yaron and Havasi, Marton and Holderrieth, Peter and Shaul, Neta and Le, Matt and Karrer, Brian and Chen, Ricky T. Q. and. Flow. 2024 , month = dec, number =. doi:10.48550/arXiv.2412.06264 , urldate =. arXiv , keywords =:2412.06264 , primaryclass =

    2024 doi arXiv

  60. [75]

    2025 , month = aug, number =

    Li, Zixuan and Wang, Zixiong and Yang, Jian and Ha. 2025 , month = aug, number =. doi:10.48550/arXiv.2508.07240 , urldate =. arXiv , keywords =:2508.07240 , primaryclass =

    2025 doi arXiv

  61. [76]

    Litalien, Joey and Ha. Neural. 2024 , month = dec, pages =. doi:10.1145/3680528.3687566 , urldate =

    2024

  62. [77]

    ACM Transactions on Graphics , volume =

    Liu, Yuan and Wang, Peng and Lin, Cheng and Long, Xiaoxiao and Wang, Jiepeng and Liu, Lingjie and Komura, Taku and Wang, Wenping , year =. ACM Transactions on Graphics , volume =. doi:10.1145/3592134 , urldate =

    doi

  63. [78]

    2012 , month = jan, journal =

    L. 2012 , month = jan, journal =. doi:10.1145/2077341.2077350 , urldate =

    2012

  64. [79]

    Luan, Fujun and Zhao, Shuang and Bala, Kavita and Dong, Zhao , year =. Unified. Computer Graphics Forum , volume =. doi:10.1111/cgf.14344 , urldate =

    doi

  65. [80]

    Photoreal

    Lv, Zhaoyang and Monge, Maurizio and Chen, Ka and Zhu, Yufeng and Goesele, Michael and Engel, Jakob and Dong, Zhao and Newcombe, Richard , year =. Photoreal. doi:10.48550/arXiv.2506.04444 , urldate =. arXiv , keywords =:2506.04444 , primaryclass =

    doi arXiv

  66. [81]

    Mahajan, Dhruv and Ramamoorthi, Ravi and Curless, Brian , editor =. A. Computer. 2006 , volume =. doi:10.1007/11744085_4 , urldate =

    2006 doi

  67. [82]

    2003 , month = jul, journal =

    A Data-Driven Reflectance Model , author =. 2003 , month = jul, journal =. doi:10.1145/882262.882343 , urldate =

    2003

  68. [83]

    and Ang, Jason and Ahmad, Anis , year =

    McCool, Michael D. and Ang, Jason and Ahmad, Anis , year =. Homomorphic Factorization of. Proceedings of the 28th Annual Conference on. doi:10.1145/383259.383276 , urldate =

  69. [84]

    Discontinuity-Aware

    Milano, Francesco and. Discontinuity-Aware. 2025 , month = jul, number =. doi:10.48550/arXiv.2507.06075 , urldate =. arXiv , keywords =:2507.06075 , primaryclass =

    2025 doi arXiv

  70. [85]

    Montes, Rosana and Ure. An

  71. [86]

    Extracting

    Munkberg, Jacob and Chen, Wenzheng and Hasselgren, Jon and Evans, Alex and Shen, Tianchang and Muller, Thomas and Gao, Jun and Fidler, Sanja , year =. Extracting. 2022. doi:10.1109/cvpr52688.2022.00810 , urldate =

    2022

  72. [88]

    and Sutskever, Ilya and Kaiser, Lukasz and Kurach, Karol and Martens, James , year =

    Neelakantan, Arvind and Vilnis, Luke and Le, Quoc V. and Sutskever, Ilya and Kaiser, Lukasz and Kurach, Karol and Martens, James , year =. Adding. doi:10.48550/arXiv.1511.06807 , urldate =. arXiv , langid =:1511.06807 , primaryclass =

    doi arXiv

  73. [89]

    Park, Jeong Joon and Florence, Peter and Straub, Julian and Newcombe, Richard and Lovegrove, Steven , year =. 2019. doi:10.1109/CVPR.2019.00025 , urldate =

    2019

  74. [90]

    2025 , month = apr, number =

    Pataki, Zador and Sarlin, Paul-Edouard and Sch. 2025 , month = apr, number =. doi:10.48550/arXiv.2504.20040 , urldate =. arXiv , keywords =:2504.20040 , primaryclass =

    2025 doi arXiv

  75. [91]

    doi:10.48550/arXiv.2504.08410 , urldate =

    Pei, Mingzhi and Cao, Xu and Wang, Xiangyi and Guo, Heng and Ma, Zhanyu , year =. doi:10.48550/arXiv.2504.08410 , urldate =. arXiv , keywords =:2504.08410 , primaryclass =

    doi arXiv

  76. [92]

    Physically

    Pharr, Matt and Jakob, Wenzel and Humphreys, Greg , year =. Physically

  77. [93]

    and Guillard, Beno

    Remelli, Edoardo and Lukoianov, Artem and Richter, Stephan R. and Guillard, Beno. 2020 , month = oct, number =. doi:10.48550/arXiv.2006.03997 , urldate =. arXiv , keywords =:2006.03997 , primaryclass =

    2020 doi arXiv

  78. [94]

    doi:10.48550/arXiv.2211.12562 , urldate =

    Rosu, Radu Alexandru and Behnke, Sven , year =. doi:10.48550/arXiv.2211.12562 , urldate =. arXiv , keywords =:2211.12562 , primaryclass =

    doi arXiv

  79. [95]

    Relightable

    Saito, Shunsuke and Schwartz, Gabriel and Simon, Tomas and Li, Junxuan and Nam, Giljoo , year =. Relightable. doi:10.48550/arXiv.2312.03704 , urldate =. arXiv , langid =:2312.03704 , primaryclass =

    doi arXiv

  80. [96]

    doi:10.48550/arXiv.2506.06271 , urldate =

    Schmidt, Jonathan and Giebenhain, Simon and Niessner, Matthias , year =. doi:10.48550/arXiv.2506.06271 , urldate =. arXiv , keywords =:2506.06271 , primaryclass =

    doi

  81. [97]

    Schmitt, Carolin and Donne, Simon and Riegler, Gernot and Koltun, Vladlen and Geiger, Andreas , year =. On. 2020. doi:10.1109/CVPR42600.2020.00355 , urldate =

    2020

  82. [98]

    Schmitt, Carolin and Anti. Towards. 2023 , month = jun, number =. doi:10.48550/arXiv.2306.03747 , urldate =. arXiv , keywords =:2306.03747 , primaryclass =

    2023 doi arXiv

  83. [99]

    doi:10.1145/3680532.3689592 , urldate =

    Singh, Gurprit and Jakob, Wenzel , year =. doi:10.1145/3680532.3689592 , urldate =

  84. [100]

    Soler, Cyril and Subr, Kartic and Nowrouzezahrai, Derek , year =. A. Computer Graphics Forum , volume =. doi:10.1111/cgf.13348 , urldate =

    doi

  85. [101]

    and Zhou, Yi , year =

    Son, Sanghyun and Gadelha, Matheus and Zhou, Yang and Xu, Zexiang and Lin, Ming C. and Zhou, Yi , year =. doi:10.48550/arXiv.2404.13445 , urldate =. arXiv , keywords =:2404.13445 , primaryclass =

    doi arXiv

  86. [102]

    2025 , month = apr, number =

    Steiner, Michael and K. 2025 , month = apr, number =. doi:10.48550/arXiv.2504.12811 , urldate =. arXiv , langid =:2504.12811 , primaryclass =

    2025 doi arXiv

  87. [103]

    Daniel and Lagunas, M

    Subias, J. Daniel and Lagunas, M. , year =. In-the-wild. Computer Graphics Forum , volume =. doi:10.1111/cgf.14765 , urldate =

    doi

  88. [104]

    Sun, Cheng and Cai, Guangyan and Li, Zhengqin and Yan, Kai and Zhang, Cheng and Marshall, Carl and Huang, Jia-Bin and Zhao, Shuang and Dong, Zhao , year =. Neural-. doi:10.48550/arXiv.2304.13445 , urldate =. arXiv , langid =:2304.13445 , primaryclass =

    doi arXiv

  89. [105]

    arXiv.org , urldate =

    Sun, Hanxiao and Gao, YuPeng and Xie, Jin and Yang, Jian and Wang, Beibei , year =. arXiv.org , urldate =

  90. [106]

    Su, Rui and Dong, Honghao and Jin, Haojie and Chen, Yisong and Wang, Guoping and Li, Sheng , year =. Vertex. doi:10.48550/arXiv.2508.07852 , urldate =. arXiv , keywords =:2508.07852 , primaryclass =

    doi arXiv

  91. [107]

    Sztrajman, Alejandro and Rainer, Gilles and Ritschel, Tobias and Weyrich, Tim , year =. Neural. Computer Graphics Forum , volume =. doi:10.1111/cgf.14335 , urldate =

    doi

  92. [108]

    and Mildenhall, Ben and

    Tancik, Matthew and Srinivasan, Pratul P. and Mildenhall, Ben and. Fourier. 2020 , month = jun, number =. arXiv , keywords =:2006.10739 , primaryclass =

    2020 arXiv

  93. [109]

    Tetzlaff, Michael , year =. High-. IEEE Transactions on Visualization and Computer Graphics , volume =. doi:10.1109/TVCG.2023.3235277 , urldate =

    2023

  94. [111]

    and Lukovi

    Truong, Anh and Mahmoud, Ahmed H. and Lukovi. Low-. 2025 , month = apr, number =. doi:10.48550/arXiv.2504.15933 , urldate =. arXiv , langid =:2504.15933 , primaryclass =

    2025 doi

  95. [112]

    and Lenz, R.K

    Tsai, R.Y. and Lenz, R.K. , year =. A New Technique for Fully Autonomous and Efficient. IEEE Transactions on Robotics and Automation , volume =. doi:10.1109/70.34770 , urldate =

    doi

  96. [113]

    and Kaiser, Lukasz and Polosukhin, Illia , year =

    Vaswani, Ashish and Shazeer, Noam and Parmar, Niki and Uszkoreit, Jakob and Jones, Llion and Gomez, Aidan N. and Kaiser, Lukasz and Polosukhin, Illia , year =. Attention. doi:10.48550/arXiv.1706.03762 , urldate =. arXiv , keywords =:1706.03762 , primaryclass =

    doi arXiv

  97. [114]

    and Srinivasan, Pratul P

    Verbin, Dor and Hedman, Peter and Mildenhall, Ben and Zickler, Todd and Barron, Jonathan T. and Srinivasan, Pratul P. , year =. Ref-. doi:10.48550/arXiv.2112.03907 , urldate =. arXiv , keywords =:2112.03907 , primaryclass =

    doi arXiv

  98. [115]

    2022 , month = jul, journal =

    Differentiable Signed Distance Function Rendering , author =. 2022 , month = jul, journal =. doi:10.1145/3528223.3530139 , urldate =

    2022

  99. [116]

    and Li, Hongsong and Torrance, Kenneth E

    Walter, Bruce and Marschner, Stephen R. and Li, Hongsong and Torrance, Kenneth E. , year =. Microfacet Models for Refraction through Rough Surfaces , booktitle =

  100. [117]

    Adaptive

    Wang, Zian and Shen, Tianchang and. Adaptive. 2023 , month = dec, journal =. doi:10.1145/3618390 , urldate =

    2023 doi

  101. [118]

    doi:10.48550/arXiv.2503.11651 , urldate =

    Wang, Jianyuan and Chen, Minghao and Karaev, Nikita and Vedaldi, Andrea and Rupprecht, Christian and Novotny, David , year =. doi:10.48550/arXiv.2503.11651 , urldate =. arXiv , keywords =:2503.11651 , primaryclass =

    doi arXiv

  102. [119]

    Arbitrarily Layered Micro-Facet Surfaces , booktitle =

    Weidlich, Andrea and Wilkie, Alexander , year =. Arbitrarily Layered Micro-Facet Surfaces , booktitle =. doi:10.1145/1321261.1321292 , urldate =

  103. [120]

    Practical

    Weier, Philippe and Riviere, J. Practical

  104. [121]

    and Kraus, M

    Weiler, M. and Kraus, M. and Ertl, T. , year =. Hardware-Based View-Independent Cell Projection , booktitle =. doi:10.1109/SWG.2002.1226505 , urldate =

    2002 arXiv

  105. [122]

    doi:10.48550/arXiv.2508.09830 , urldate =

    Wei, Shenxing and Li, Jinxi and Yang, Yafei and Zhou, Siyuan and Yang, Bo , year =. doi:10.48550/arXiv.2508.09830 , urldate =. arXiv , keywords =:2508.09830 , primaryclass =

    doi arXiv

  106. [123]

    Principles of

    Weyrich, Tim and Lawrence, Jason and Lensch, Hendrik and Rusinkiewicz, Szymon and Zickler, Todd , langid =. Principles of

  107. [124]

    Uncertainty for

    Wiersma, Ruben and Philip, Julien and Ha. Uncertainty for. 2025 , langid =

    2025

  108. [125]

    Learning

    Wortsman, Mitchell and Horton, Maxwell and Guestrin, Carlos and Farhadi, Ali and Rastegari, Mohammad , year =. Learning. doi:10.48550/arXiv.2102.10472 , urldate =. arXiv , keywords =:2102.10472 , primaryclass =

    doi arXiv

  109. [126]

    Wu, Qi and Esturo, Janick Martinez and Mirzaei, Ashkan and

  110. [129]

    Xue, Bowen and Zhao, Shuang and Jensen, Henrik Wann and Montazeri, Zahra , year =. An. arXiv , keywords =:2307.10135 , primaryclass =

    arXiv

  111. [130]

    Computer Graphics Forum , volume =

    Xue, Bowen and Zhao, Shuang and Jensen, Henrik Wann and Montazeri, Zahra , title =. Computer Graphics Forum , volume =. doi:https://doi.org/10.1111/cgf.15116 , eprint =

    doi

  112. [131]

    Xu, Bing and Wu, Liwen and Hasan, Milos and Luan, Fujun and Georgiev, Iliyan and Xu, Zexiang and Ramamoorthi, Ravi , year =. Special. doi:10.1145/3588432.3591524 , urldate =

  113. [132]

    doi:10.48550/arXiv.2508.04366 , urldate =

    Yang, Songyun and Han, Yufei and Zhang, Jilong and Liang, Kongming and Yu, Peng and Qu, Zhaowei and Guo, Heng , year =. doi:10.48550/arXiv.2508.04366 , urldate =. arXiv , keywords =:2508.04366 , primaryclass =

    doi arXiv

  114. [133]

    and Szeliski, Richard and Barron, Jonathan T

    Yariv, Lior and Hedman, Peter and Reiser, Christian and Verbin, Dor and Srinivasan, Pratul P. and Szeliski, Richard and Barron, Jonathan T. and Mildenhall, Ben , year =. doi:10.48550/arXiv.2302.14859 , urldate =. arXiv , langid =:2302.14859 , primaryclass =

    doi arXiv

  115. [134]

    Multiview

    Yariv, Lior and Kasten, Yoni and Moran, Dror and Galun, Meirav and Atzmon, Matan and Basri, Ronen and Lipman, Yaron , year =. Multiview. doi:10.48550/arXiv.2003.09852 , urldate =. arXiv , keywords =:2003.09852 , primaryclass =

    2003 doi arXiv

  116. [135]

    Deep Sets , booktitle =

    Zaheer, Manzil and Kottur, Satwik and Ravanbakhsh, Siamak and Poczos, Barnabas and Salakhutdinov, Russ R and Smola, Alexander J , editor =. Deep Sets , booktitle =. 2017 , volume =

    2017

  117. [136]

    doi:10.48550/arXiv.2311.16671 , urldate =

    Zarzar, Jesus and Ghanem, Bernard , year =. doi:10.48550/arXiv.2311.16671 , urldate =. arXiv , langid =:2311.16671 , primaryclass =

    doi arXiv

  118. [137]

    Real-Time

    Zeltner, Tizian and Rousselle, Fabrice and Weidlich, Andrea and Clarberg, Petrik and Nov. Real-Time. 2024 , month = jun, journal =. doi:10.1145/3659577 , urldate =

    2024 doi

  119. [138]

    Zhang, Qijian and Hou, Junhui and Wang, Wenping and He, Ying , year =. Flatten. doi:10.48550/arXiv.2405.14633 , urldate =. arXiv , keywords =:2405.14633 , publisher =

    doi arXiv

  120. [139]

    2020 , month = aug, journal =

    Path-Space Differentiable Rendering , author =. 2020 , month = aug, journal =. doi:10.1145/3386569.3392383 , urldate =

    2020

  121. [140]

    Hybrelighter:

    Zhao, Hanwen and Akers, John and Elmieh, Baback and. Hybrelighter:. 2025 , month = aug, number =. doi:10.48550/arXiv.2508.14930 , urldate =. arXiv , keywords =:2508.14930 , primaryclass =

    2025 doi arXiv

  122. [141]

    Zheng, Chuankun and Zheng, Ruzhang and Wang, Rui and Zhao, Shuang and Bao, Hujun , year =. A. ACM Transactions on Graphics , volume =. doi:10.1145/3490385 , urldate =

    doi

  123. [142]

    IEEE Transactions on Visualization and Computer Graphics , volume =

    Zhong, Hongliang and Zhang, Jingbo and Liao, Jing , year =. IEEE Transactions on Visualization and Computer Graphics , volume =. doi:10.1109/TVCG.2023.3330518 , urldate =. arXiv , keywords =:2310.11864 , primaryclass =

    2023

  124. [143]

    doi:10.48550/arXiv.2507.00476 , urldate =

    Zhou, Chenliang and Hu, Zheyuan and Oztireli, Cengiz , year =. doi:10.48550/arXiv.2507.00476 , urldate =. arXiv , keywords =:2507.00476 , primaryclass =

    doi arXiv

  125. [144]

    Gaussian

    Zhu, Zuo-Liang and Yang, Jian and Wang, Beibei , year =. Gaussian. doi:10.48550/arXiv.2507.15629 , urldate =. arXiv , langid =:2507.15629 , primaryclass =

    doi arXiv

  126. [145]

    doi:10.48550/arXiv.2406.18544 , urldate =

    Zhu, Zuo-Liang and Wang, Beibei and Yang, Jian , year =. doi:10.48550/arXiv.2406.18544 , urldate =. arXiv , langid =:2406.18544 , primaryclass =

    doi arXiv

  127. [146]

    doi:10.48550/arXiv.2508.04508 , urldate =

    Zhu, Haodong and Li, Changbai and Ren, Yangyang and Feng, Zichao and Liu, Xuhui and Chen, Hanlin and Zhen, Xiantong and Zhang, Baochang , year =. doi:10.48550/arXiv.2508.04508 , urldate =. arXiv , keywords =:2508.04508 , primaryclass =

    doi arXiv

  128. [147]

    Filip, J. and V. Template-Based Sampling of Anisotropic. Computer Graphics Forum , volume =. 2014 , month =

    2014

  129. [148]

    Transactions on Graphics (Proceedings of SIGGRAPH Asia) , volume =

    Jonathan Dupuy and Wenzel Jakob , title =. Transactions on Graphics (Proceedings of SIGGRAPH Asia) , volume =. 2018 , month = nov, doi =

    2018

  130. [149]

    Computer Graphics Forum , volume=

    Real-Time Neural Materials using Block-Compressed Features , author=. Computer Graphics Forum , volume=. 2024 , organization=

    2024

  131. [150]

    ACM Transactions on Graphics (TOG) , volume=

    A practical extension to microfacet theory for the modeling of varying iridescence , author=. ACM Transactions on Graphics (TOG) , volume=. 2017 , publisher=

    2017

  132. [151]

    SIGGRAPH Course: Physically Based Shading in Theory and Practice , volume=

    Moving frostbite to physically based rendering 3.0 , author=. SIGGRAPH Course: Physically Based Shading in Theory and Practice , volume=

  133. [152]

    SIGGRAPH Course, Physically Based Shading , volume=

    Revisiting physically based shading at imageworks , author=. SIGGRAPH Course, Physically Based Shading , volume=

  134. [153]

    14th Central European Seminar on Computer Graphics , volume=

    Layered materials in real-time rendering , author=. 14th Central European Seminar on Computer Graphics , volume=

  135. [154]

    New York: Academic , volume=

    Multiple light scattering , author=. New York: Academic , volume=. 1981 , publisher=

    1981

  136. [155]

    ACM Transactions on Graphics (ToG) , volume=

    A comprehensive framework for rendering layered materials , author=. ACM Transactions on Graphics (ToG) , volume=. 2014 , publisher=

    2014

  137. [156]

    ACM Transactions on Graphics (TOG) , volume=

    The layer laboratory: a calculus for additive and subtractive composition of anisotropic surface reflectance , author=. ACM Transactions on Graphics (TOG) , volume=. 2018 , publisher=

    2018

  138. [157]

    SIGGRAPH Asia 2019 Technical Briefs , pages=

    Real-time rendering of layered materials with anisotropic normal distributions , author=. SIGGRAPH Asia 2019 Technical Briefs , pages=

    2019

  139. [158]

    ACM Transactions on Graphics , volume=

    Efficient rendering of layered materials using an atomic decomposition with statistical operators , author=. ACM Transactions on Graphics , volume=

  140. [159]

    Journal of Computer Graphics Techniques (JCGT) , volume=

    Rendering layered materials with anisotropic interfaces , author=. Journal of Computer Graphics Techniques (JCGT) , volume=

  141. [160]

    ACM Transactions on Graphics (TOG) , volume=

    Transfer matrix based layered materials rendering , author=. ACM Transactions on Graphics (TOG) , volume=. 2021 , publisher=

    2021

  142. [161]

    IEEE transactions on visualization and computer graphics , volume=

    Rendering thin transparent layers with extended normal distribution functions , author=. IEEE transactions on visualization and computer graphics , volume=. 2016 , publisher=

    2016

  143. [162]

    Journal of Computer Graphics Techniques (JCGT) , volume =

    Eric Heitz , title =. Journal of Computer Graphics Techniques (JCGT) , volume =. 2014 , month =

    2014

  144. [163]

    2019 , eprint=

    PyTorch: An Imperative Style, High-Performance Deep Learning Library , author=. 2019 , eprint=

    2019

  145. [164]

    CoRR , year=

    Adam: A Method for Stochastic Optimization , author=. CoRR , year=

  146. [165]

    International Conference on Learning Representations , year=

    Decoupled Weight Decay Regularization , author=. International Conference on Learning Representations , year=

  147. [166]

    arXiv preprint arXiv:2512.23696 , year=

    OpenPBR: Novel Features and Implementation Details , author=. arXiv preprint arXiv:2512.23696 , year=

  148. [167]

    1963 , publisher =

    The scattering of electromagnetic waves from rough surfaces , author=. 1963 , publisher =

    1963

  149. [168]

    ACM SIGGRAPH 2003 Papers , pages =

    Matusik, Wojciech and Pfister, Hanspeter and Brand, Matt and McMillan, Leonard , title =. ACM SIGGRAPH 2003 Papers , pages =. 2003 , isbn =. doi:10.1145/1201775.882343 , url =

    2003

  150. [169]

    Lafortune, Eric P. F. and Foo, Sing-Choong and Torrance, Kenneth E. and Greenberg, Donald P. , title =. 1997 , isbn =. doi:10.1145/258734.258801 , booktitle =

    1997

  151. [170]

    Experimental analysis of BRDF models , year =

    Ngan, Addy and Durand, Fr\'. Experimental analysis of BRDF models , year =. Proceedings of the Sixteenth Eurographics Conference on Rendering Techniques , pages =

  152. [171]

    2023 , isbn =

    Lucas, Simon and Ribardiere, Mickael and Pacanowski, Romain and Barla, Pascal , title =. 2023 , isbn =. doi:10.1145/3610548.3618241 , booktitle =

    2023

  153. [172]

    A Fully-correlated Anisotropic Micrograin BSDF Model , year =

    Lucas, Simon and Ribardi\`. A Fully-correlated Anisotropic Micrograin BSDF Model , year =. doi:10.1145/3658224 , journal =

    doi

  154. [173]

    2018 , issue_date =

    Xie, Feng and Hanrahan, Pat , title =. 2018 , issue_date =. doi:10.1145/3272127.3275078 , journal =

    2018

  155. [174]

    2016 , issue_date =

    Heitz, Eric and Hanika, Johannes and d'Eon, Eugene and Dachsbacher, Carsten , title =. 2016 , issue_date =. doi:10.1145/2897824.2925943 , journal =

    2016

  156. [175]

    2023 , issue_date =

    Yu, Yunchen and Xia, Mengqi and Walter, Bruce and Michielssen, Eric and Marschner, Steve , title =. 2023 , issue_date =. doi:10.1145/3592414 , journal =

    2023 doi

  157. [176]

    2025 , MONTH = Jun, KEYWORDS =

    Fourneau, Gary and Barla, Pascal and Pacanowski, Romain , BOOKTITLE =. 2025 , MONTH = Jun, KEYWORDS =

    2025

  158. [177]

    2017 , issue_date =

    Holzschuch, Nicolas and Pacanowski, Romain , title =. 2017 , issue_date =. doi:10.1145/3072959.3073621 , journal =

    2017

  159. [178]

    2024 , MONTH = Apr, DOI =

    Fourneau, Gary and Pacanowski, Romain and Barla, Pascal , JOURNAL =. 2024 , MONTH = Apr, DOI =

    2024

  160. [179]

    ACM Transactions on Graphics (TOG) , volume=

    Efficient and accurate spherical kernel integrals using isotropic decomposition , author=. ACM Transactions on Graphics (TOG) , volume=. 2015 , publisher=

    2015

  161. [180]

    Computer Graphics Forum , volume =

    Mahajan, Dhruv and Tseng, Yu-Ting and Ramamoorthi, Ravi , title =. Computer Graphics Forum , volume =. doi:https://doi.org/10.1111/j.1467-8659.2008.01251.x , year =

    2008

  162. [181]

    ACM Trans

    Lawrence, Jason and Rusinkiewicz, Szymon and Ramamoorthi, Ravi , title =. ACM Trans. Graph. , month = aug, pages =. 2004 , issue_date =. doi:10.1145/1015706.1015751 , abstract =

    2004

  163. [182]

    Spherical wavelets: efficiently representing functions on the sphere , year =

    Schr\". Spherical wavelets: efficiently representing functions on the sphere , year =. doi:10.1145/218380.218439 , booktitle =

  164. [183]

    ACM Trans

    Wang, Jiaping and Ren, Peiran and Gong, Minmin and Snyder, John and Guo, Baining , title =. ACM Trans. Graph. , month = dec, pages =. 2009 , issue_date =. doi:10.1145/1618452.1618479 , abstract =

    2009

  165. [184]

    Rational BRDF , year=

    Pacanowski, Romain and Salazar Celis, Oliver and Schlick, Christophe and Granier, Xavier and Poulin, Pierre and Cuyt, Annie , journal=. Rational BRDF , year=

  166. [185]

    Computer Graphics Forum , volume =

    Bilgili, Ahmet and Oztürk, Aydn and Kurt, Murat , title =. Computer Graphics Forum , volume =. doi:https://doi.org/10.1111/j.1467-8659.2011.02072.x , abstract =

    2011

  167. [186]

    ACM Trans

    Xu, Kun and Sun, Wei-Lun and Dong, Zhao and Zhao, Dan-Yong and Wu, Run-Dong and Hu, Shi-Min , title =. ACM Trans. Graph. , month = nov, articleno =. 2013 , issue_date =. doi:10.1145/2508363.2508386 , abstract =

    2013

  168. [187]

    Wavelet encoding of BRDFs for real-time rendering , year =

    Claustres, Luc and Barthe, Lo\". Wavelet encoding of BRDFs for real-time rendering , year =. Proceedings of Graphics Interface 2007 , pages =. doi:10.1145/1268517.1268546 , abstract =

    2007

  169. [188]

    and Doorn, Andrea J

    Koenderink, Jan J. and Doorn, Andrea J. van and Stavridi, Marigo , title =. Proceedings of the 4th European Conference on Computer Vision-Volume II - Volume II , pages =. 1996 , isbn =

    1996

  170. [189]

    2005 , isbn =

    Zickler, Todd and Enrique, Sebastian and Ramamoorthi, Ravi and Belhumeur, Peter , title =. 2005 , isbn =

    2005

  171. [190]

    , author=

    Fast arbitrary brdf shading for low-frequency lighting using spherical harmonics. , author=. Rendering Techniques , volume=

  172. [191]

    IEEE Transactions on Visualization and Computer Graphics , volume=

    Estimating homogeneous data-driven BRDF parameters from a reflectance map under known natural lighting , author=. IEEE Transactions on Visualization and Computer Graphics , volume=. 2021 , publisher=

    2021

  173. [192]

    Real-Time Rendering 4th Edition , year =

    Tomas Akenine-M\". Real-Time Rendering 4th Edition , year =

  174. [193]

    ACM Trans

    Tongbuasirilai, Tanaboon and Unger, Jonas and Guillemot, Christine and Miandji, Ehsan , title =. ACM Trans. Graph. , month = oct, articleno =. 2022 , issue_date =. doi:10.1145/3533427 , abstract =

    2022 doi

  175. [194]

    Eurographics Workshop on Rendering Techniques , pages=

    A new change of variables for efficient BRDF representation , author=. Eurographics Workshop on Rendering Techniques , pages=. 1998 , organization=

    1998

  176. [195]

    ACM SIGGRAPH 2004 Papers , pages =

    Lawrence, Jason and Rusinkiewicz, Szymon and Ramamoorthi, Ravi , title =. ACM SIGGRAPH 2004 Papers , pages =. 2004 , isbn =. doi:10.1145/1186562.1015751 , abstract =

    2004

  177. [196]

    BRDF Explorer , year =

  178. [197]

    Journal of Computer Graphics Techniques (JCGT) , volume = 7, number = 4, pages =

    Eric Heitz , title =. Journal of Computer Graphics Techniques (JCGT) , volume = 7, number = 4, pages =

  179. [198]

    2023 , booktitle =

    Vaidyanathan, Karthik and Salvi, Marco and Wronski, Bartlomiej and Akenine-Möller, Tomas and Ebelin, Pontus and Lefohn, Aaron , title = ". 2023 , booktitle =

    2023

  180. [199]

    Computer Graphics Forum , volume =

    Bai, Yaoyi and Wu, Songyin and Zeng, Zheng and Yan, Lingqi and Wang, Beibei , title =. Computer Graphics Forum , volume =

  181. [201]

    Computer Graphics Forum , volume =

    Xu, Zilin and Zhou, Yang and Litman, Yehonathan and Chiang, Matt Jen-Yuan and Yan, Lingqi and Michels, Anton , title =. Computer Graphics Forum , volume =

  182. [202]

    2025 , isbn =

    Xu, Zilin and Chen, Xiang and Liu, Chen and Wang, Beibei and Wang, Lu and Montazeri, Zahra and Yan, Ling-Qi , title =. 2025 , isbn =. doi:10.1145/3721238.3730626 , booktitle =

    2025

Pith tools

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