Pith. sign in

REVIEW 4 major objections 6 minor 59 references

GaRe: Relightable 3D Gaussian Splatting for Outdoor Scenes from Unconstrained Photo Collections

T0 review · 4 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read GaRe decomposes unconstrained outdoor photo collections into reflectance plus sun, sky, and indirect shading, and recombines them with ray-traced visibility to relight scenes in real time.

desk verdict A credible engineering advance for outdoor 3DGS relighting, but the physical-decomposition claim outruns the evidence; worth revision, not rejection. read the letter →

arxiv 2507.20512 v1 pith:GDHRYC66 submitted 2025-07-28 cs.CV

Haiyang Bai , Jiaqi Zhu , Songru Jiang , Wei Huang , Tao Lu , Yuanqi Li , Jie Guo , Runze Fu
show 2 more authors
Yanwen Guo Lijun Chen
This is my paper · ORCID
classification cs.CV
keywords relighting3DGaussiansplattingintrinsicimagedecompositionoutdoorscenessunvisibilityray-tracedshadowsunconstrainedphotocollectionsreal-timerendering
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 aims to establish that the global illumination in unconstrained outdoor photo collections—casual, pose-calibrated images taken under varying weather and times of day—can be decomposed into three physically meaningful shading layers (direct sun, sky, and indirect bounce light) and a single reflectance layer, and that these layers can be recombined to relight the scene. The reward would be practical: instead of compressing each image's lighting into one opaque appearance vector, an artist could edit sun position, sky tone, or bounce intensity separately, including casting dynamic shadows, while rendering in real time. The paper argues this is achievable by supervising the decomposition with masks derived from image residuals and ray-traced visibility, and demonstrates relit novel views and shadow changes on four landmark photo collections.

What carries the argument

The load-bearing object is the multiplicative decomposition identity I = (V·S_sun + S_sky + S_ind)·R together with the sun visibility map V. Around it, three mechanisms carry the argument: (1) residual-based visibility extraction—an ambient-only Gaussian model is trained on masked pixels, and the residual between full and ambient renderings is binary-clustered into sunlit and shadowed regions; (2) region-based supervision with a structural consistency loss—Eqs. (9)–(11) constrain each rendered shading component against ground truth in its own mask region, and the SCL term (12) propagates structure from the well-constrained reflectance to all shading components; (3) ray-traced visibility with sky and front filters, followed by baking into per-Gaussian visibility features so shadows for arbitrary light directions render in real time.

What would settle it

Render a synthetic outdoor scene with known sun position, known sky radiance, and constant albedo; run GaRe on the renderings and check whether the recovered sun visibility matches the true shadow mask and whether recovered reflectance stays constant when the sun direction is changed. Any systematic deviation between the recovered visibility and the rendered shadow mask, or reflectance drift with sun angle, would falsify the claimed decomposition.

Watch

Extended reading notes

Core claim

GaRe models every observed image as I = (V·S_sun + S_sky + S_ind)·R, where R is Lambertian surface reflectance, S_sun, S_sky, and S_ind are shading from direct sunlight, sky radiance, and indirect light, and V is a sun visibility mask that is 1 in sunlit pixels and 0 elsewhere. The paper's central discovery is that this decomposition can be learned from unconstrained photo collections without explicit 3D modeling of shadows. A residual-based procedure first trains an ambient-only Gaussian model, computes the residual between the fully lit image and the ambient render, and binary-clusters that residual to obtain a refined sun visibility map. Then per-Gaussian features and per-image embeddings are decoded into the shading components, supervised region-wise by the visibility and sky masks and globally by a structural consistency loss. Finally, ray tracing through the Gaussians produces visibility from arbitrary sun directions, and this visibility is baked into per-Gaussian features so all shading components and shadows are computed in one rasterization pass.

Load-bearing premise

The decomposition stands or falls on hand-set preprocessing thresholds that decide which pixels count as sky and which count as sunlit: if those thresholds mislabel shadows or sky, every shading component derived from them is wrong.

Editorial extensions

If this is right

  • Users can relight an outdoor scene by editing or interpolating the sun, sky, and indirect shading embeddings independently, rather than changing a single global appearance code.
  • Dynamic shadows follow from choosing a new sun direction at inference time, because ray-traced visibility is baked into Gaussian features and rendered in the same pass as the shading.
  • Novel-view synthesis on unconstrained landmark datasets matches or exceeds existing baselines while keeping sharp light-shadow boundaries, as shown in the paper's quantitative and qualitative evaluations.
  • The method extends naturally to cloudy images by setting sun visibility to zero, reducing the global model to sky plus indirect shading plus reflectance.
  • Because all shading and shadow components are computed in a single rasterization pass, relighting runs in real time rather than requiring the expensive per-image optimization of radiance-field relighting.

Reading between the lines

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

  • If the decomposition generalizes, the per-Gaussian sun/sky/indirect features and visibility map could serve as pseudo ground truth for training single-image outdoor relighting networks, a step the paper leaves to future work.
  • Replacing the fixed preprocessing thresholds for sky and sunlit pixels with learned or adaptive masks could extend the method to scenes with unusual albedo or partial overcast; the paper tests only fixed thresholds.
  • The Lambertian assumption means specular surfaces such as water, glass, or wet roads are folded into reflectance or shading artifacts; adding a specular term to Eq. (1) is a natural testable extension.
  • Because visibility is baked from sampled sun directions, extremely low sun angles or a sun that moves during a sequence may require re-baking; rendering a continuous sun trajectory is an inference-time extension not demonstrated in the paper.
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

4 major / 6 minor

Summary. This paper presents GaRe, a 3D Gaussian splatting framework for relighting outdoor scenes captured from unconstrained photo collections. The method decomposes each image into reflectance and three shading components—sun, sky, and indirect—following the Lambertian model I = (V·S_sun + S_sky + S_ind)·R, where V is a sun visibility map. The framework has three main contributions: (1) residual-based sun visibility extraction via binary clustering on the difference between a globally illuminated image and an ambient-only rendered counterpart; (2) a region-based supervision framework with structural consistency loss for illumination decomposition; and (3) ray-tracing-based shadow modeling with visibility baking for real-time rendering. The authors evaluate novel-view synthesis on four outdoor datasets (Staats Theater, Brandenburg Gate, Trevi Fountain, Sacre Coeur) and provide qualitative results for decomposition, relighting, and shadow effects. The paper also includes ablations for the visibility extraction, structural consistency loss, and two-step filtering strategy.

Significance. If the central claim holds, GaRe is a meaningful advance over methods that encode outdoor illumination as a single latent vector: it enables per-component manipulation of sun, sky, and indirect light, and supports dynamic shadow synthesis via baked ray-traced visibility. The quantitative NVS results are competitive with state-of-the-art methods on four standard datasets, and the ablations support the individual contributions of residual-based visibility, SCL, and the filtering strategy. The paper is clearly written and the method is reproducible in structure. However, the physical interpretability claim rests on a decomposition that is supervised only through heuristic masks and never directly validated quantitatively; this is the core weakness that needs to be addressed before the central claim can be accepted.

major comments (4)
  1. [Sec. 8.1, Sec. 8.2, Eq. (3), Eq. (4)] The decomposition is fundamentally supervised by the coarse sun visibility mask V_c and the sky mask M_sky, both derived from fixed thresholds (brightness threshold tau_V=0.3 after the gamma transform in Eq. 17, and disparity threshold tau_D=0.1 in Sec. 8.1). Brightness alone cannot separate illumination from albedo: a low-albedo sunlit surface can be darker than a high-albedo shadowed surface. Similarly, relative disparity predictions from Depth Anything V2 are not scale-calibrated across images, so a fixed threshold tau_D=0.1 may misclassify sky pixels for some scenes. Since the ambient-only model is trained with Eq. 3 using V_c, the residual map in Eq. 4 inherits these errors, and the refined visibility V propagates them into Eq. 9 and the region-based losses. The manuscript provides no quantitative evidence that the masks are accurate, and the paper's own ablation in Fig. 7 shows only one example. Please provide a quantitative evaluation of the visibility and sky masks (e.g., against hand-annotated labels on a subset of images), or show that the final decomposition is insensitive to the thresholds.
  2. [Sec. 3.4, Eq. (14)] The transmittance recurrence in Eq. (14) is not derived and appears dimensionally suspect: the term 'n_{j-1} d_i^T' is introduced without defining the product or explaining how a normal vector and a light-direction vector enter the transmittance of a ray in a point-based traversal. If n and d are unit vectors, their dot product is a scalar between -1 and 1, which would multiply the scalar transmittance term (1-alpha_{j-1}) T_{j-1} without any physical basis in the standard 3DGS compositing equations. If n d^T is intended as an outer product, the equation would be a matrix, which contradicts the scalar nature of transmittance. Please provide the derivation of Eq. (14) and clarify the dimensions of each term, or replace it with the standard transmittance update T_j = (1 - alpha_{j-1}) T_{j-1}.
  3. [Sec. 3.3, Eqs. (9)–(11), Sec. 5.2, Table 1] The physical interpretability of the decomposition is validated only through qualitative figures; Table 1 reports only novel-view synthesis metrics. Equations (9)–(11) compare sums of components against the same input image, so no loss ever observes a single component against an independent target. To support the claim that the components are physically interpretable, please add quantitative evaluation of the decomposition itself: e.g., relighting agreement with held-out images, decomposition consistency across views in the same illumination condition, or a comparison against a baseline that removes the decomposition (e.g., using a single shading component) on the relighting tasks. At minimum, include a quantitative study of how the components change under interpolation in Fig. 5, and report standard intrinsic-image metrics if any labels are available.
  4. [Sec. 3.2, Eq. (3), Sec. 8.2] The coarse visibility mask V_c is derived from the same images used to train the ambient-only model, and the residual clustering of Eq. 4 then uses the residuals of that model to produce the refined visibility V. This is a self-referential pipeline: if V_c is incorrect, the ambient-only model learns to fit the wrong pixels, and the residuals will reflect the model's error rather than the true illumination. Please provide a quantitative analysis of the sensitivity of the final visibility to the choice of taus and V_c, including the effect on relighting quality and decomposition consistency.
minor comments (6)
  1. [Sec. 8.1] The description of the sky mask threshold 'tau_D = 0.1' is unclear: it is not stated whether the condition is disparity < tau_D or disparity > tau_D, and the text says 'near-zero disparity values' but the threshold is 0.1; please clarify the assignment rule and how the threshold was chosen.
  2. [Sec. 3.2, Eq. (3)] The notation in Eq. (3) uses a comma inside the norm: '||(I_amb - I) · (1 - V_c)||_1' — the comma appears to be a typo and should be removed.
  3. [Sec. 3.4, Eq. (14)] The index convention in Eq. (14) is inconsistent with the surrounding text: T_j is defined in terms of T_{j-1}, but the text says 'After traversing, the remanent transmittance T_j corresponds to the ray-traced visibility of Gaussian G_k.' It should be clarified which index corresponds to which Gaussian along the ray.
  4. [Sec. 3.4, Eq. (15)] The per-Gaussian visibility prediction v_k in Eq. (15) outputs values in [-1, 1], but the ray-traced visibility v_rt is stated to be in R_+ (positive). The loss L_vis in Eq. (16) compares them directly; please clarify how the signed output is mapped to a visibility in [0,1], or why the range mismatch is acceptable.
  5. [Sec. 5.2, Fig. 4–6] The qualitative figures would be more convincing if they included a comparison to the state-of-the-art relighting baselines (e.g., NeRF-OSR, GS-W) for the decomposition and relighting results, not just NVS.
  6. [Sec. 2.3] The related-work section is thorough, but the description of 'unstructured photo collections' would benefit from a clear distinction between the fixed-illumination multi-view datasets and the unconstrained photo collections used here, especially in Table 1 where NeRF-W and 3DGS baselines are evaluated on the same data.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the derivation chain is self-contained, though the sun-sky-indirect decomposition is heuristic and underdetermined.

full rationale

GaRe's central quantitative result, novel-view synthesis (Table 1), is trained and evaluated against external baselines, so it is not a circular prediction. The sun-visibility pipeline (Sec. 3.2, Eqs. 3-4) is a bootstrap: an ambient-only model is trained with a coarse mask V_c from HSV-threshold preprocessing, and the refined visibility V is computed by clustering the residual |I - I_amb|. V is influenced by V_c but is not identical to it by construction; this is a heuristic estimation procedure, not a definitional equivalence or a fitted parameter renamed as a prediction. The region-based losses (Eqs. 9-11) constrain sums of shading components against the same input image using masks built from the same heuristics, so the individual sun/sky/indirect factorization is underdetermined and lacks direct quantitative validation; however, underdetermination is a correctness and validation concern, not circularity. The paper's own Sec. 11 acknowledges dependence on scene geometry and imperfect surfaces. The only self-citation, [51] for sunny/cloudy pre-classification, is not load-bearing for the main relighting claim. No step in the derivation reduces, by the paper's own equations, to its own input.

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

The central decomposition rests on several hand-set thresholds and the Lambertian/intrinsic-image model. No new physical entities are introduced; the per-Gaussian features and per-image embeddings are learned parameters rather than postulates. The main circularity risk is that the supervision for the decomposition comes from the same heuristic masks used to train the ambient-only model, so the 'predicted' visibility and shading are partly forced by preprocessing choices.

free parameters (4)
  • Sky mask disparity threshold tau_D = 0.1
    Used to classify sky pixels in Sec. 8.1; the mask Msky is central to the region-based supervision and visibility settings.
  • Coarse sun visibility brightness threshold tau_V = 0.3
    Used to classify sunlit pixels in Sec. 8.2; the coarse visibility Vc is used to train the ambient-only model and to derive the refined visibility.
  • Gamma transform coefficients (beta, epsilon, gamma) = beta=1/255, epsilon=0.1, gamma=1.5
    Eq. 17 defines the brightness transform used for coarse sun visibility; these hand-set values directly influence the masks.
  • Loss weights lambda_sun1, lambda_sky1, lambda_ind1, lambda_sun_sc, lambda_sky_sc, lambda_ind_sc = 1.0, 10.0, 10.0, 0.1, 5.0, 5.0
    Chosen in Sec. 4 to balance region supervision and structural consistency; no sensitivity study is provided, and they affect decomposition quality.
assumptions (5)
  • domain assumption Lambertian surface assumption: observed appearance is the product of incident illumination and surface reflectance.
    Stated in Sec. 3.1; ignores specularity and view-dependent materials, which limits the physical accuracy of the decomposition for real outdoor scenes.
  • domain assumption The sky is at infinite distance and appears as near-zero disparity in monocular depth estimates.
    Used in Sec. 8.1 to generate sky masks from Depth Anything V2 disparity; fails for sky occluded by mountains or buildings, or for reflective surfaces.
  • domain assumption Intrinsic image decomposition holds: I = (V*S_sun + S_sky + S_ind)*R.
    Eq. 1 is the core model; it assumes a clean multiplicative separation that real scenes with interreflections and non-Lambertian effects do not satisfy exactly.
  • ad hoc to paper Binary clustering of residuals separates sunlit from shadowed pixels.
    Eq. 4 assumes that the difference between the global image and the ambient-only render is dominated by sun effects; strong texture or colored materials can create large residuals that break this assumption.
  • ad hoc to paper Ray-tracing transmittance recurrence: T_j = (1-alpha_{j-1}) T_{j-1} n_{j-1} d_i^T.
    Eq. 14 is presented without derivation and appears dimensionally inconsistent; it embeds a normal dot product into a transmittance factor, which is not standard alpha compositing.

how reviews work

0 comments
Cite this review

Pith. "Pith review of GaRe: Relightable 3D Gaussian Splatting for Outdoor Scenes from Unconstrained Photo Collections." pith.science (2026). https://pith.science/paper/GDHRYC66

@misc{pith2026250720512,
  author       = {Pith},
  title        = {Pith review of: GaRe: Relightable 3D Gaussian Splatting for Outdoor Scenes from Unconstrained Photo Collections},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GDHRYC66}},
  note         = {Machine review of arXiv:2507.20512}
}
read the original abstract

We propose a 3D Gaussian splatting-based framework for outdoor relighting that leverages intrinsic image decomposition to precisely integrate sunlight, sky radiance, and indirect lighting from unconstrained photo collections. Unlike prior methods that compress the per-image global illumination into a single latent vector, our approach enables simultaneously diverse shading manipulation and the generation of dynamic shadow effects. This is achieved through three key innovations: (1) a residual-based sun visibility extraction method to accurately separate direct sunlight effects, (2) a region-based supervision framework with a structural consistency loss for physically interpretable and coherent illumination decomposition, and (3) a ray-tracing-based technique for realistic shadow simulation. Extensive experiments demonstrate that our framework synthesizes novel views with competitive fidelity against state-of-the-art relighting solutions and produces more natural and multifaceted illumination and shadow effects.

Figures

Figures reproduced from arXiv: 2507.20512 by the authors.

Figure 1
Figure 1. Building on the principles of intrinsic image decomposition, our relighting framework, termed [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of illumination decomposition. The GaRe framework accurately decomposes global illumination into reflectance [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Comparison with SOTA methods in outdoor NVS. Our method delivers high-quality texture reconstruction for landmarks and [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Visualization of the decomposed components and synthesis results. Sun shading captures the effects of direct sunlight, sky shad [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Relighting through illumination interpolation. Visualization shows how scene appearance evolves when interpolating between [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Visualization of shadow effect. Left: (Top) Decomposed [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: Comparison of visibility extraction methods. (a) The [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: Comparison of the effect of structural consistency loss [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 10
Figure 10. Figure 10: Visualization of the illumination decomposition results under cloudy conditions, where direct sunlight is absent. The scene is [PITH_FULL_IMAGE:figures/full_fig_p012_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 53 canonical work pages

  1. [1]

    Mip-nerf: A multiscale representation for anti-aliasing neu- ral radiance fields

    Jonathan T Barron, Ben Mildenhall, Matthew Tancik, Peter Hedman, Ricardo Martin-Brualla, and Pratul P Srinivasan. Mip-nerf: A multiscale representation for anti-aliasing neu- ral radiance fields. In ICCV, pages 5855–5864, 2021. 2

    2021

  2. [2]

    Intrinsic images in the wild

    Sean Bell, Kavita Bala, and Noah Snavely. Intrinsic images in the wild. ACM TOG, 33(4):159:1–159:12, 2014. 2

    work page 2014

  3. [3]

    Gs3: Efficient relighting with triple gaussian splatting

    Zoubin Bi, Yixin Zeng, Chong Zeng, Fan Pei, Xiang Feng, Kun Zhou, and Hongzhi Wu. Gs3: Efficient relighting with triple gaussian splatting. In ACM TOG, pages 1–12, 2024. 2, 3

    work page 2024

  4. [4]

    Bar- ron, Ce Liu, and Hendrik P.A

    Mark Boss, Raphael Braun, Varun Jampani, Jonathan T. Bar- ron, Ce Liu, and Hendrik P.A. Lensch. Nerd: Neural re- flectance decomposition from image collections. In ICCV,

  5. [5]

    Hallucinated neural radiance fields in the wild

    Xingyu Chen, Qi Zhang, Xiaoyu Li, Yue Chen, Ying Feng, Xuan Wang, and Jue Wang. Hallucinated neural radiance fields in the wild. In CVPR, pages 12943–12952, 2022. 2, 3, 7

    work page 2022

  6. [6]

    Swag: Splatting in the wild images with appearance-conditioned gaussians

    Hiba Dahmani, Moussab Bennehar, Nathan Piasco, Luis Roldao, and Dzmitry Tsishkou. Swag: Splatting in the wild images with appearance-conditioned gaussians. In ECCV, pages 325–340. Springer, 2025. 7, 11

    work page 2025

  7. [7]

    Multi-view intrinsic images of outdoors scenes with an application to relighting

    Sylvain Duch ˆene, Clement Riant, Gaurav Chaurasia, Jorge Lopez-Moreno, Pierre-Yves Laffont, Stefan Popov, Adrien Bousseau, and George Drettakis. Multi-view intrinsic images of outdoors scenes with an application to relighting. ACM TOG, page 16, 2015. 2, 3

    work page 2015

  8. [8]

    K-planes: Ex- plicit radiance fields in space, time, and appearance

    Sara Fridovich-Keil, Giacomo Meanti, Frederik Rahbæk War- burg, Benjamin Recht, and Angjoo Kanazawa. K-planes: Ex- plicit radiance fields in space, time, and appearance. InCVPR, pages 12479–12488, 2023. 2

    work page 2023

Show all 59 references
  1. [9]

    Relightable 3d gaussian: Real-time point cloud relighting with brdf decomposition and ray trac- ing

    Jian Gao, Chun Gu, Youtian Lin, Hao Zhu, Xun Cao, Li Zhang, and Yao Yao. Relightable 3d gaussian: Real-time point cloud relighting with brdf decomposition and ray trac- ing. arXiv preprint arXiv:2311.16043, 2023. 2, 3

    2023 arXiv

  2. [10]

    Outcast: Outdoor single-image relighting with cast shadows

    David Griffiths, Tobias Ritschel, and Julien Philip. Outcast: Outdoor single-image relighting with cast shadows. In Com- puter Graphics Forum, pages 179–193. Wiley Online Library,

  3. [11]

    Nerfren: Neural radiance fields with reflections

    Yuan-Chen Guo, Di Kang, Linchao Bao, Yu He, and Song- Hai Zhang. Nerfren: Neural radiance fields with reflections. In CVPR, pages 18409–18418, 2022. 2

    2022

  4. [12]

    Progressive photon mapping

    Toshiya Hachisuka, Shinji Ogaki, and Henrik Wann Jensen. Progressive photon mapping. ACM Transactions on Graphics (TOG), 27(5):130:1–130:8, 2008. 2

    2008

  5. [13]

    A rapid hierarchical radiosity algorithm

    Pat Hanrahan, David Salzman, and Larry Aupperle. A rapid hierarchical radiosity algorithm. In Proceedings of the 28th annual conference on Computer graphics and interactive techniques, pages 197–206, 1991. 2

    1991

  6. [14]

    Real-time, photo-realistic, physically based rendering of fine scale hu- man skin structure

    Antonio Haro, Brian Guenter, and Irfan Essa. Real-time, photo-realistic, physically based rendering of fine scale hu- man skin structure. In Rendering Techniques 2001: Proceed- ings of the Eurographics Workshop in London, United King- dom, June 25–27, 2001 12, pages 53–62. Spr...

    2001

  7. [15]

    Gaussianshader: 3d gaussian splatting with shading functions for reflective sur- faces

    Yingwenqi Jiang, Jiadong Tu, Yuan Liu, Xifeng Gao, Xiaox- iao Long, Wenping Wang, and Yuexin Ma. Gaussianshader: 3d gaussian splatting with shading functions for reflective sur- faces. In CVPR, pages 5322–5332, 2024. 2

    2024

  8. [16]

    Tensoir: Tensorial inverse rendering

    Haian Jin, Isabella Liu, Peijia Xu, Xiaoshuai Zhang, Songfang Han, Sai Bi, Xiaowei Zhou, Zexiang Xu, and Hao Su. Tensoir: Tensorial inverse rendering. In CVPR, 2023. 3

    2023

  9. [17]

    3d gaussian splatting for real-time radiance field rendering

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimk ¨uhler, and George Drettakis. 3d gaussian splatting for real-time radiance field rendering. ACM TOG, 42(4):139–1, 2023. 2, 3, 5, 7, 11

    2023

  10. [18]

    Precomputed local radiance transfer for real- time lighting design

    Anders Wang Kristensen, Tomas Akenine-M ¨oller, and Hen- rik Wann Jensen. Precomputed local radiance transfer for real- time lighting design. ACM TOG, 24(3):1208–1215, 2005. 2

    2005

  11. [19]

    Neroic: Neu- ral rendering of objects from online image collections

    Zhengfei Kuang, Kyle Olszewski, Menglei Chai, Zeng Huang, Panos Achlioptas, and Sergey Tulyakov. Neroic: Neu- ral rendering of objects from online image collections. ACM TOG, 41(4):1–12, 2022. 2

    2022

  12. [20]

    Wildgaussians: 3d gaussian splatting in the wild

    Jonas Kulhanek, Songyou Peng, Zuzana Kukelova, Marc Pollefeys, and Torsten Sattler. Wildgaussians: 3d gaussian splatting in the wild. arXiv preprint arXiv:2407.08447, 2024. 2, 3, 5, 7, 11

    2024 arXiv

  13. [21]

    Coherent intrinsic images from photo collections

    Pierre-Yves Laffont, Adrien Bousseau, Sylvain Paris, Fr ´edo Durand, and George Drettakis. Coherent intrinsic images from photo collections. ACM TOG , 31(6):202:1–202:11,

  14. [22]

    Rich intrinsic image decomposition of outdoor scenes from multiple views

    Pierre-Yves Laffont, Adrien Bousseau, and George Drettakis. Rich intrinsic image decomposition of outdoor scenes from multiple views. IEEE TVCG, 19(2):210–224, 2012. 2, 3

    2012

  15. [23]

    Neulighting: Neural lighting for free viewpoint outdoor scene relighting with unconstrained photo collections

    Quewei Li, Jie Guo, Yang Fei, Feichao Li, and Yanwen Guo. Neulighting: Neural lighting for free viewpoint outdoor scene relighting with unconstrained photo collections. In ACM TOG, pages 1–9, 2022. 3

    2022

  16. [24]

    Inverse rendering for 9 complex indoor scenes: Shape, spatially-varying lighting and SVBRDF from a single image

    Zhengqin Li, Mohammad Shafiei, Ravi Ramamoorthi, Kalyan Sunkavalli, and Manmohan Chandraker. Inverse rendering for 9 complex indoor scenes: Shape, spatially-varying lighting and SVBRDF from a single image. In CVPR, pages 2472–2481. Computer Vision Foundation / IEEE, 2020. 2

    2020

  17. [25]

    Multi-view inverse rendering for large-scale real-world indoor scenes

    Zhen Li, Lingli Wang, Mofang Cheng, Cihui Pan, and Jiaqi Yang. Multi-view inverse rendering for large-scale real-world indoor scenes. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 12499– 12509, 2023. 3

    2023

  18. [26]

    Gs-ir: 3d gaussian splatting for inverse rendering

    Zhihao Liang, Qi Zhang, Ying Feng, Ying Shan, and Kui Jia. Gs-ir: 3d gaussian splatting for inverse rendering. In CVPR, pages 21644–21653, 2024. 2, 3

    2024

  19. [27]

    Scaffold-gs: Structured 3d gaussians for view-adaptive rendering

    Tao Lu, Mulin Yu, Linning Xu, Yuanbo Xiangli, Limin Wang, Dahua Lin, and Bo Dai. Scaffold-gs: Structured 3d gaussians for view-adaptive rendering. In CVPR, pages 20654–20664,

  20. [28]

    Dynamic dif- fuse global illumination with ray-traced irradiance fields

    Zander Majercik, Jean-Philippe Guertin, Derek Nowrouzezahrai, and Morgan McGuire. Dynamic dif- fuse global illumination with ray-traced irradiance fields. ACM TOG, 8(2), 2019. 2

    2019

  21. [29]

    Inverse rendering for computer graphics

    Stephen Robert Marschner. Inverse rendering for computer graphics. Cornell University, 1998. 2

    1998

  22. [30]

    Nerf in the wild: Neural radiance fields for uncon- strained photo collections

    Ricardo Martin-Brualla, Noha Radwan, Mehdi SM Sajjadi, Jonathan T Barron, Alexey Dosovitskiy, and Daniel Duck- worth. Nerf in the wild: Neural radiance fields for uncon- strained photo collections. In ECCV, pages 7210–7219, 2021. 2, 3, 7

    2021

  23. [31]

    Nerf: Representing scenes as neural radiance fields for view synthe- sis

    Ben Mildenhall, Pratul P Srinivasan, Matthew Tancik, Jonathan T Barron, Ravi Ramamoorthi, and Ren Ng. Nerf: Representing scenes as neural radiance fields for view synthe- sis. Communications of the ACM, 65(1):99–106, 2020. 2

    2020

  24. [32]

    Giljoo Nam, Joo Ho Lee, Diego Gutierrez, and Min H. Kim. Practical SVBRDF acquisition of 3d objects with unstructured flash photography. ACM TOG, 37(6):267, 2018. 2

    2018

  25. [33]

    Shape from interreflections

    Shree K Nayar, Katsushi Ikeuchi, and Takeo Kanade. Shape from interreflections. IJCV, 6:173–195, 1991. 2

    1991

  26. [34]

    All-frequency shadows using non-linear wavelet lighting approximation

    Ren Ng, Ravi Ramamoorthi, and Pat Hanrahan. All-frequency shadows using non-linear wavelet lighting approximation. ACM TOG, 22(3):376–381, 2003. 2

    2003

  27. [35]

    Diffuse reflectance from rough surfaces

    Michael Oren and Shree K Nayar. Diffuse reflectance from rough surfaces. In CVPR, pages 763–764. IEEE, 1993. 3

    1993

  28. [36]

    Physically based rendering: From theory to implementation

    Matt Pharr, Wenzel Jakob, and Greg Humphreys. Physically based rendering: From theory to implementation. MIT Press,

  29. [37]

    Multi-view relighting using a geometry-aware network

    Julien Philip, Micha ¨el Gharbi, Tinghui Zhou, Alexei A Efros, and George Drettakis. Multi-view relighting using a geometry-aware network. ACM TOG, 38(4):78–1, 2019. 3

    2019

  30. [38]

    Nerf for outdoor scene relighting

    Viktor Rudnev, Mohamed Elgharib, William Smith, Lingjie Liu, Vladislav Golyanik, and Christian Theobalt. Nerf for outdoor scene relighting. In ECCV, pages 615–631. Springer,

  31. [39]

    Gir: 3d gaussian inverse rendering for relightable scene factorization

    Yahao Shi, Yanmin Wu, Chenming Wu, Xing Liu, Chen Zhao, Haocheng Feng, Jingtuo Liu, Liangjun Zhang, Jian Zhang, Bin Zhou, Errui Ding, and Jingdong Wang. Gir: 3d gaussian inverse rendering for relightable scene factorization. Arxiv,

  32. [40]

    Precomputed radiance transfer for real-time rendering in dynamic, low- frequency lighting environments

    Peter-Pike Sloan, Jan Kautz, and John Snyder. Precomputed radiance transfer for real-time rendering in dynamic, low- frequency lighting environments. In Seminal Graphics Pa- pers: Pushing the Boundaries, Volume 2, pages 339–348. AK Peters/CRC Press, 2023. 2

    2023

  33. [41]

    Srinivasan, Boyang Deng, Xiuming Zhang, Matthew Tancik, Ben Mildenhall, and Jonathan T

    Pratul P. Srinivasan, Boyang Deng, Xiuming Zhang, Matthew Tancik, Ben Mildenhall, and Jonathan T. Barron. Nerv: Neu- ral reflectance and visibility fields for relighting and view syn- thesis. In CVPR, 2021. 2

    2021

  34. [42]

    Relight my nerf: A dataset for novel view synthesis and re- lighting of real world objects

    Marco Toschi, Riccardo De Matteo, Riccardo Spezialetti, Daniele De Gregorio, Luigi Di Stefano, and Samuele Salti. Relight my nerf: A dataset for novel view synthesis and re- lighting of real world objects. In CVPR, pages 20762–20772,

  35. [43]

    Procedural modeling and physically based rendering for synthetic data generation in automotive applica- tions

    Apostolia Tsirikoglou, Joel Kronander, Magnus Wrenninge, and Jonas Unger. Procedural modeling and physically based rendering for synthetic data generation in automotive applica- tions. arXiv preprint arXiv:1710.06270, 2017. 1

    2017 arXiv

  36. [44]

    Realtime ray tracing and its use for interactive global illumination

    Ingo Wald, Timothy J Purcell, J ¨org Schmittler, Carsten Ben- thin, and Philipp Slusallek. Realtime ray tracing and its use for interactive global illumination. In Eurographics (State of the Art Reports), pages 85–122, 2003. 2

    2003

  37. [45]

    Gpu-based soft shadow rendering using non-linear pre-computed radiance transfer approximation

    Lili Wang, Jing Wang, and Qinping Zhao. Gpu-based soft shadow rendering using non-linear pre-computed radiance transfer approximation. In VSMM, pages 81–92. Springer,

  38. [46]

    An improved illumination model for shaded display

    Turner Whitted. An improved illumination model for shaded display. ACM Transactions on Graphics (TOG), 24(3):4–13,

  39. [47]

    Simultane- ous localization and appearance estimation with a consumer RGB-D camera

    Hongzhi Wu, Zhaotian Wang, and Kun Zhou. Simultane- ous localization and appearance estimation with a consumer RGB-D camera. IEEE Trans. Vis. Comput. Graph. , 22(8): 2012–2023, 2016. 2

    2012

  40. [48]

    Recover- ing shape and spatially-varying surface reflectance under un- known illumination

    Rui Xia, Yue Dong, Pieter Peers, and Xin Tong. Recover- ing shape and spatially-varying surface reflectance under un- known illumination. ACM TOG, 35(6):187:1–187:12, 2016. 2

    2016

  41. [49]

    Depth anything v2

    Lihe Yang, Bingyi Kang, Zilong Huang, Zhen Zhao, Xiao- gang Xu, Jiashi Feng, and Hengshuang Zhao. Depth anything v2. arXiv preprint arXiv:2406.09414, 2024. 11

    2024 arXiv

  42. [50]

    Intrinsicnerf: Learning intrinsic neural radiance fields for editable novel view synthesis

    Weicai Ye, Shuo Chen, Chong Bao, Hujun Bao, Marc Polle- feys, Zhaopeng Cui, and Guofeng Zhang. Intrinsicnerf: Learning intrinsic neural radiance fields for editable novel view synthesis. In ICCV, 2023. 2

    2023

  43. [51]

    Hierarchical disentan- gled representation learning for outdoor illumination estima- tion and editing

    Piaopiao Yu, Jie Guo, Fan Huang, Cheng Zhou, Hongwei Che, Xiao Ling, and Yanwen Guo. Hierarchical disentan- gled representation learning for outdoor illumination estima- tion and editing. In ICCV, pages 15313–15322, 2021. 11

    2021

  44. [52]

    Relighting neural radiance fields with shadow and highlight hints

    Chong Zeng, Guojun Chen, Yue Dong, Pieter Peers, Hongzhi Wu, and Xin Tong. Relighting neural radiance fields with shadow and highlight hints. In ACM TOG, pages 1–11, 2023. 2

    2023

  45. [53]

    Gaussian in the wild: 3d gaussian splatting for unconstrained image collections

    Dongbin Zhang, Chuming Wang, Weitao Wang, Peihao Li, Minghan Qin, and Haoqian Wang. Gaussian in the wild: 3d gaussian splatting for unconstrained image collections. In ECCV, pages 341–359. Springer, 2025. 2, 3, 5, 7, 11

    2025

  46. [54]

    Nerfac- tor: Neural factorization of shape and reflectance under an unknown illumination

    Xiuming Zhang, Pratul P Srinivasan, Boyang Deng, Paul De- bevec, William T Freeman, and Jonathan T Barron. Nerfac- tor: Neural factorization of shape and reflectance under an unknown illumination. ACM TOG, 40(6):1–18, 2021. 2 10

    2021

  47. [55]

    Modeling indirect illumination for inverse rendering

    Yuanqing Zhang, Jiaming Sun, Xingyi He, Huan Fu, Rongfei Jia, and Xiaowei Zhou. Modeling indirect illumination for inverse rendering. In CVPR, 2022. 3

    2022

  48. [56]

    Addi- tionally, we enhance the preprocessing step by extracting sky masksMsky and coarse sun visibilityVc for all training im- ages, which are used in Sec

    Dataset Preprocess Similar to the original Gaussian model, our method utilizes a dataset of pose-calibrated outdoor images and sparse point clouds generated via Structure-from-Motion (SfM). Addi- tionally, we enhance the preprocessing step by extracting sky masksMsky and coars...

  49. [57]

    However, it introduces inaccuracies in the visi- bility of the sky region

    Sky Semantics The two-step filtering strategy effectively mitigates the mis- leading influence of floating Gaussians in the sky on surface visibility. However, it introduces inaccuracies in the visi- bility of the sky region. To resolve this, we define a sky semantic ok∈ R+ fo...

  50. [58]

    In these cases, the effect of sunlight is disregarded, and the global illumination for each viewpoint is decomposed into sky shading and indirect shading, as shown in Fig

    Cloudy Day In our experimental dataset, there are several instances rep- resenting scenarios without direct sunlight,i.e., cloudy days. In these cases, the effect of sunlight is disregarded, and the global illumination for each viewpoint is decomposed into sky shading and indi...

  51. [59]

    Limitations and Future Work Our method is not without limitations. Firstly, due to the phased training process in our framework, the overall train- ing duration is relatively longer compared to state-of-the-art methods such as 3DGS [17] and SW AG [6]. However, it remains super...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.