Pith. sign in

REVIEW 3 major objections 7 minor 80 references

Learning Heterogeneous Mixture of Scene Experts for Large-scale Neural Radiance Fields

T0 review · 3 major / 7 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read A learnable hash gating network that splits a city-scale scene into scale-appropriate hash experts improves rendering quality while training 8× faster and rendering 16× faster than predecessor methods.

desk verdict Well-engineered extension that delivers real efficiency gains and a useful heterogeneous hash-expert design, but the unqualified SOTA claim and the homogeneous-only main benchmark overstate the accuracy case. read the letter →

arxiv 2505.02005 v2 pith:4LVPKGZD submitted 2025-05-04 cs.CV

Zhenxing Mi , Ping Yin , Xue Xiao , Dan Xu This is my paper · ORCID
classification cs.CV
keywords neuralradiancefieldslarge-scalescenemodelingmixtureofexpertslearnabledecompositionhashgridencodingheterogeneousrepresentationnovelviewsynthesisscalablerendering
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 establish that scene decomposition for large-scale neural radiance fields should be learned, not hand-designed, and that the learned experts should differ from one another in grid resolution. It proposes a single end-to-end framework in which a hash-based gating network routes each 3D point to one of several hash-grid experts, each covering a different resolution range. The paper claims this heterogeneous mixture captures the varied geometry and appearance of real scenes, and that it reaches higher rendering accuracy than prior large-scale NeRFs on datasets up to 6.5 km² while cutting training time by 8× and rendering time by 16× relative to its predecessor.

What carries the argument

The load-bearing object is the resolution pyramid of hash experts: eight multi-resolution hash encodings whose base resolutions are logarithmically spaced from 16 to 512 and whose desired resolutions from 2048 to 16384, so experts range from coarse to fine. The mechanism that makes routing learnable is the gate-value multiplication $\hat E(x)=G(x)_i E_i(x)$, which sends gradient through the discrete selection, supported by a gating balance loss $L_b = n\sum_i f_i g_i$ that encourages balanced use of experts. Around this core sit a hash-based gating network that computes gate values by looking up a hash feature and passing it through a small MLP, and a fused hash-dispatching implementation that avoids explicit input reorganization and capacity padding.

What would settle it

Train the heterogeneous and homogeneous versions on a large scene whose detail is spread uniformly across scales, or whose finest structures sit below the smallest expert resolution; if the heterogeneous advantage vanishes or the gate shows no resolution-gradient correlation, the fixed resolution pyramid, not learnable decomposition, is carrying the result.

Watch

Extended reading notes

Core claim

The central claim, stated in the paper's own terms, is that a large-scale scene is best represented by a Heterogeneous Mixture of Hash Experts (HMoHE): a set of hash-based NeRF experts whose grid resolutions form a pyramid from coarse to fine, together with a trainable hash gate that learns which resolution each scene region needs. The paper argues that the discrete Top-1 expert choice becomes optimizable because the selected expert's output is multiplied by the gate value, so gradients flow back into the gating network, and a balance loss keeps experts from collapsing. It presents evidence that on urban scenes the gate assignments correlate with local density gradients: low-resolution experts receive low-gradient regions and high-resolution experts receive high-gradient regions, which the homogeneous baseline does not show. The paper's quantitative claim is state-of-the-art rendering accuracy on Mega-NeRF, Block-NeRF, UrbanBIS, and MatrixCity scenes, with 8× faster training and 16× faster rendering than the previous Switch-NeRF.

Load-bearing premise

The load-bearing premise is that a hand-set ladder of coarse-to-fine hash-grid resolutions is the right axis for separating a scene's parts, and that image-reconstruction error plus one balance term can teach the gate that separation without any geometric guide.

Editorial extensions

If this is right

  • Large-scale scenes can be reconstructed end-to-end with no manual partitioning, so the same training pipeline transfers to new cities or districts without hand-tuned decomposition rules.
  • Because only one expert is active per 3D point, adding experts raises capacity without proportionally raising per-point computation, making the method more scalable by expert count.
  • The resolution pyramid lets flat ground and detailed facades be represented by the same network at different grid scales, which is why the heterogeneous version beats the homogeneous one on every reported metric.
  • At 6.65 s per 1,152 × 864 image with less than half the memory of prior methods, the model becomes practical for interactive or repeated rendering of very large scenes.
  • Training for 50k steps (5.5 h) already reaches state-of-the-art accuracy on most Mega-NeRF metrics, so the method offers a direct compute-accuracy trade-off for users with limited GPU budgets.

Reading between the lines

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

  • The fixed resolution ladder is a strong prior rather than a learned quantity; scenes whose feature-scale distribution falls outside the chosen 16-to-16384 range would likely need the ladder adapted per scene, a testable extension the paper does not explore.
  • Since the gate learns only from final rendered color, the decomposition may track appearance as much as geometry; adding depth or normal supervision could produce partitions that are more geometrically coherent and might help in texture-poor regions.
  • The same sparsely gated design with heterogeneous hash resolutions could be carried over to other grid-based 3D representations such as Gaussian splatting, offering those methods a learnable alternative to manual cell division.
  • The reported correlation between expert resolution and local density gradient suggests a 'scale field' could be predicted directly, possibly reducing the number of experts or improving initialization.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 7 minor

Summary. This paper proposes Switch-NeRF++, a mixture-of-experts framework for large-scale NeRF. It replaces the MLP gating network and MLP experts of the authors' previous Switch-NeRF with a hash-based gating network and a set of hash-encoding experts with different resolution ranges ('heterogeneous mixture of hash experts', HMoHE). The gating network is trained end-to-end with a photometric loss and a balance loss, and each 3D point is dispatched to a single expert. Experiments are reported on Mega-NeRF, Block-NeRF, an adapted UrbanBIS scene covering more than 6.5 km^2, and MatrixCity. The paper claims state-of-the-art rendering accuracy, an 8x training acceleration, and a 16x rendering acceleration over Switch-NeRF.

Significance. The proposed architecture addresses a real limitation of prior large-scale NeRFs: hand-crafted decomposition rules and homogeneous expert capacity. If the heterogeneous design is validated on the main benchmarks, the paper will make a useful contribution. The hash-based gating and the efficient hash-dispatching implementation are thoughtfully engineered, and the ablations (gating design, expert count, Top-1 vs. Top-2, hash capacity, Gumbel selection) are extensive. The adaptation of a 6.5 km^2 UrbanBIS scene is also valuable for benchmarking scalability. However, as written, the manuscript does not support its central claim: the heterogeneous experts are evaluated only on Block-NeRF and UrbanBIS, while the headline Mega-NeRF and MatrixCity results are obtained with homogeneous experts. In addition, the unqualified 'state-of-the-art' accuracy claim is contradicted by the paper's own VastGaussian comparison in Table 14. These issues are fixable with re-evaluation or careful qualification, so the paper merits a major revision rather than rejection.

major comments (3)
  1. [Section 4.2, Tables 1 and 3] The implementation details state that 'For the Mega-NeRF and Matrixcity datasets, we use the same hash experts.' The headline accuracy and efficiency results in Tables 1 and 3 therefore come from the homogeneous variant, not from the proposed HMoHE. The central claim that Switch-NeRF++ 'learns heterogeneous decomposition and heterogeneous NeRFs' is not tested on the primary Mega-NeRF benchmark; the heterogeneous results appear only in Table 2 (Block-NeRF and UrbanBIS). Either the experiments need to be rerun with heterogeneous experts on Mega-NeRF and MatrixCity, or the paper must explicitly state that the headline benchmark uses a homogeneous variant and reframe the core claim accordingly.
  2. [Abstract, Section 4.4, Table 14] The unqualified claim of 'state-of-the-art scene rendering accuracy' is inconsistent with the paper's own VastGaussian-aligned results. In Table 14, VastGaussian has higher PSNR than Switch-NeRF++ on Building (23.50 vs. 22.41), Rubble (26.92 vs. 26.02), Residence (24.25 vs. 24.00), and Campus (26.00 vs. 24.37), with Switch-NeRF++ higher only on Sci-Art (27.40 vs. 26.81). The text correctly concedes that 'Switch-NeRF++ underperforms Vast-Gaussian on some scenes,' contradicting the abstract and contribution statements. The accuracy claim should be qualified, for example as state-of-the-art among NeRF/ray-based large-scale methods or as competitive with large-scale Gaussian-splatting methods.
  3. [Table 1, Section 4.3] The claimed 8x training acceleration is based on Switch-NeRF++* trained for 50k steps (5.5h) compared with Switch-NeRF trained for 500k steps (43h). This is not a matched-quality comparison: on PSNR, the 50k model is lower than Switch-NeRF on Building (21.27 vs. 21.54), Sci-Art (25.65 vs. 26.52), and Campus (23.54 vs. 23.62). The 'state-of-the-art results with 5.5h' statement in the table caption is therefore not supported on all scenes and metrics. To substantiate the acceleration claim, the authors should report iso-quality or matched-budget comparisons, such as Switch-NeRF at the same number of steps or the time required to match Switch-NeRF's final PSNR.
minor comments (7)
  1. [Section 3.3, Eq. (5)] The balance loss is written as 'Lb = nPfigi' with undefined symbols and an unexplained 'minimized as 1' statement; please state the standard form n * sum(f_i * p_i) and define f_i and p_i clearly.
  2. [Algorithm 1, line 18] The pseudocode writes 'G = Softmax(fg)', which is inconsistent with Eq. (2), where the gate values are obtained by applying Softmax to MLP(Hash(x)). Please fix the pseudocode to match the main text.
  3. [Table 3] The table title contains the typo 'PNSR'; it should be 'PSNR'.
  4. [Section 3.2] In the pyramid heterogeneous expert design paragraph, 'Swhtch-NeRF++' should be 'Switch-NeRF++'.
  5. [Table 2] The abbreviations 'Homo.' and 'Heter.' are used in the table but are not defined in the caption or table header; please define them.
  6. [Section 4.4, Table 14 discussion] The text attributes the SSIM/LPIPS gap to the SSIM loss used by Gaussian-splatting methods and to rasterization differences. This is plausible, but it should be supported by an ablation in which an SSIM loss is added to Switch-NeRF++, or explicitly stated as a hypothesis rather than as an explanation.
  7. [Figures 10 and 11] The point-cloud decomposition visualizations are compelling but qualitative. A numeric summary of the correlation between density gradients and expert resolutions would strengthen the claim that the heterogeneous experts learn scale-adaptive decompositions.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the architecture is validated against external benchmarks and no claim reduces to a fitted input or self-citation chain.

full rationale

Switch-NeRF++ is an empirical systems paper; its central claims (accuracy, 8x training acceleration, 16x rendering acceleration) are validated against external datasets (Mega-NeRF, Block-NeRF, UrbanBIS, MatrixCity) and external baselines (Instant-NGP, Mega-NeRF, 3D-GS, VastGaussian, CityGaussian, Hierarchical-3DGaussians, Zip-NeRF). The resolution pyramid in Algorithm 1 is a fixed architectural choice, not a parameter fitted to the reported target metrics; the gating balance weight is held constant (lambda = 5e-4) and is ablated. The comparisons to Switch-NeRF [4] are measurements of the authors' own prior system used as a baseline, not a self-citation invoked as evidence for correctness of the new design. No equation reduces to its own inputs: Eq. (2)-(3) define the gating/expert multiplication, Eq. (4)-(5) define the losses, and the reported PSNR/SSIM/LPIPS numbers come from volume rendering of trained models on held-out evaluation images. The one passage that requires qualification is Table 14, where the paper concedes 'While Switch-NeRF++ underperforms Vast-Gaussian on some scenes (e.g., Building, Rubble, and Residence)'; this weakens the unqualified 'state-of-the-art' phrasing in the abstract, but that is a claim-support or correctness issue, not a circular derivation. No circularity found.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

The paper introduces no new physical or mathematical entities; its experts and gating network are compositions of existing hash grids and MLPs. The central assumptions are domain-level: that resolution range is the right axis of heterogeneity and that a photometric loss plus balance loss can learn a stable decomposition.

free parameters (4)
  • Hash expert resolution ranges (base and desired, 8 experts) = base resolutions log-spaced in [16,512], desired resolutions log-spaced in [2048,16384]
    Chosen by hand to cover coarse to fine scene scales; not derived from scene data. Central to the heterogeneous expert claim in Section 3.2 and Algorithm 1.
  • Number of hash experts = 8
    Ablated in Table 4 (2 vs 8), but the default is a design choice; capacity and compute scale with this number.
  • Gating balance loss weight lambda = 5e-4
    Set as a constant for all experiments; only one value is tested, and removing it causes training collapse in Table 8.
  • Maximum hash map size in gating and experts = 2^19
    Set to match the network capacity of the INGP baseline, which uses 2^22; not otherwise justified by data.
assumptions (3)
  • domain assumption Large-scale scenes can be decomposed into regions whose appearance and geometry complexity is well described by a single grid resolution scale.
    The entire heterogeneous expert design assumes scale is the dominant axis of heterogeneity; no spectral or statistical analysis justifies this beyond the density-gradient visualizations in Fig. 11.
  • domain assumption The gating network, trained with photometric loss and balance loss, will converge to a stable decomposition that generalizes to held-out views.
    Relies on MoE training dynamics from [11] and [64]; the paper shows it works empirically but provides no proof.
  • standard math The volume rendering equation and the hash encoding from Instant-NGP are correct background.
    Standard NeRF background in Section 3.3 and hash encoding from [5] are assumed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning Heterogeneous Mixture of Scene Experts for Large-scale Neural Radiance Fields." pith.science (2026). https://pith.science/paper/4LVPKGZD

@misc{pith2026250502005,
  author       = {Pith},
  title        = {Pith review of: Learning Heterogeneous Mixture of Scene Experts for Large-scale Neural Radiance Fields},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4LVPKGZD}},
  note         = {Machine review of arXiv:2505.02005}
}
abstract

Recent NeRF methods on large-scale scenes have underlined the importance of scene decomposition for scalable NeRFs. Although achieving reasonable scalability, there are several critical problems remaining unexplored, i.e., learnable decomposition, modeling scene heterogeneity, and modeling efficiency. In this paper, we introduce Switch-NeRF++, a Heterogeneous Mixture of Hash Experts (HMoHE) network that addresses these challenges within a unified framework. It is a highly scalable NeRF that learns heterogeneous decomposition and heterogeneous NeRFs efficiently for large-scale scenes in an end-to-end manner. In our framework, a gating network learns to decompose scenes and allocates 3D points to specialized NeRF experts. This gating network is co-optimized with the experts by our proposed Sparsely Gated Mixture of Experts (MoE) NeRF framework. We incorporate a hash-based gating network and distinct heterogeneous hash experts. The hash-based gating efficiently learns the decomposition of the large-scale scene. The distinct heterogeneous hash experts consist of hash grids of different resolution ranges, enabling effective learning of the heterogeneous representation of different scene parts. These design choices make our framework an end-to-end and highly scalable NeRF solution for real-world large-scale scene modeling to achieve both quality and efficiency. We evaluate our accuracy and scalability on existing large-scale NeRF datasets and a new dataset with very large-scale scenes ($>6.5km^2$) from UrbanBIS. Extensive experiments demonstrate that our approach can be easily scaled to various large-scale scenes and achieve state-of-the-art scene rendering accuracy. Furthermore, our method exhibits significant efficiency, with an 8x acceleration in training and a 16x acceleration in rendering compared to Switch-NeRF. Codes will be released at https://github.com/MiZhenxing/Switch-NeRF.

Figures

Figures reproduced from arXiv: 2505.02005 by the authors.

Figure 1
Figure 1. The bird’s-eye view of a scene in UrbanBIS [ [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Large-scale NeRFs with different decomposition methods. Dotted lines indicate non-differentiable operations, while [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. The framework overview of Switch-NeRF and Switch-NeRF++ for large-scale scene modeling. In Switch-NeRF++, [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (10 more)
Figure 4
Figure 4. Figure 4: (a) A detailed illustration of the MLP-based gating network with an MLP expert in Switch-NeRF. (b) A detailed [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: A detailed illustration of our pyramid design of [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: The comparison between the uniform dispatch and [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Block diagram illustrating the integrated architectural [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 8
Figure 8. Figure 8: A qualitative comparison of rendered images of the INGP [ [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: (a) Rendered images of INGP [5], Switch-NeRF [4], and our Switch-NeRF++ on UrbanBIS. Ours can render images with notably higher quality compared to other methods. (b) A visualization of our decomposition of the radiance field of the UrbanBIS dataset. Different experts …
Figure 11
Figure 11. Figure 11: The visualization of the distribution of density [PITH_FULL_IMAGE:figures/full_fig_p013_11.png]
Figure 10
Figure 10. Figure 10: Visualization results of our decomposition of the ra [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]
Figure 13
Figure 13. Figure 13: The accuracy of our Switch-NeRF++ and other meth [PITH_FULL_IMAGE:figures/full_fig_p014_13.png]
Figure 12
Figure 12. Figure 12: Comparison with large-scale Gaussian methods and [PITH_FULL_IMAGE:figures/full_fig_p014_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

80 extracted references · 62 canonical work pages

  1. [1]

    Nerf: Representing scenes as neural radiance fields for view synthesis,

    B. Mildenhall, P . P . Srinivasan, M. Tancik, J. T. Barron, R. Ra- mamoorthi, and R. Ng, “Nerf: Representing scenes as neural radiance fields for view synthesis,” in ECCV, 2020

    2020

  2. [2]

    Mega-nerf: Scal- able construction of large-scale nerfs for virtual fly-throughs,

    H. Turki, D. Ramanan, and M. Satyanarayanan, “Mega-nerf: Scal- able construction of large-scale nerfs for virtual fly-throughs,” in CVPR, 2022

    work page 2022

  3. [3]

    Block-nerf: Scalable large scene neural view synthesis,

    M. Tancik, V . Casser, X. Yan, S. Pradhan, B. Mildenhall, P . P . Srinivasan, J. T. Barron, and H. Kretzschmar, “Block-nerf: Scalable large scene neural view synthesis,” in CVPR, 2022

    work page 2022

  4. [4]

    Switch-neRF: Learning scene decomposition with mixture of experts for large-scale neural radiance fields,

    Z. MI and D. Xu, “Switch-neRF: Learning scene decomposition with mixture of experts for large-scale neural radiance fields,” in ICLR, 2023

    work page 2023

  5. [5]

    Instant neural graphics primitives with a multiresolution hash encoding,

    T. Müller, A. Evans, C. Schied, and A. Keller, “Instant neural graphics primitives with a multiresolution hash encoding,” ACM T rans. Graph., vol. 41, no. 4, pp. 102:1–102:15, 2022

    work page 2022

  6. [6]

    Direct voxel grid optimization: Super-fast convergence for radiance fields reconstruction,

    C. Sun, M. Sun, and H. Chen, “Direct voxel grid optimization: Super-fast convergence for radiance fields reconstruction,” in CVPR, 2022

    work page 2022

  7. [7]

    Plenoxels: Radiance fields without neural net- works,

    S. Fridovich-Keil, A. Yu, M. Tancik, Q. Chen, B. Recht, and A. Kanazawa, “Plenoxels: Radiance fields without neural net- works,” in CVPR, 2022

    work page 2022

  8. [8]

    Tensorf: Tensorial radiance fields,

    A. Chen, Z. Xu, A. Geiger, J. Yu, and H. Su, “Tensorf: Tensorial radiance fields,” in ECCV, 2022

    work page 2022

Show all 80 references
  1. [9]

    Grid-guided neural radiance fields for large urban scenes,

    L. Xu, Y. Xiangli, S. Peng, X. Pan, N. Zhao, C. Theobalt, B. Dai, and D. Lin, “Grid-guided neural radiance fields for large urban scenes,” in CVPR, 2023

    2023

  2. [10]

    Ur- banbis: a large-scale benchmark for fine-grained urban building instance segmentation,

    G. Yang, F. Xue, Q. Zhang, K. Xie, C.-W. Fu, and H. Huang, “Ur- banbis: a large-scale benchmark for fine-grained urban building instance segmentation,” in SIGGRAPH, 2023

    2023

  3. [11]

    Outrageously large neural networks: The sparsely-gated mixture-of-experts layer,

    N. Shazeer, A. Mirhoseini, K. Maziarz, A. Davis, Q. V . Le, G. E. Hinton, and J. Dean, “Outrageously large neural networks: The sparsely-gated mixture-of-experts layer,” in ICLR, 2017

    2017

  4. [12]

    Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity,

    W. Fedus, B. Zoph, and N. Shazeer, “Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity,” JMLR, 2022

    2022

  5. [13]

    Object recognition from local scale-invariant fea- tures,

    D. G. Lowe, “Object recognition from local scale-invariant fea- tures,” in ICCV, vol. 2, pp. 1150–1157, IEEE, 1999

    1999

  6. [14]

    Efficient sequential corre- spondence selection by cosegmentation,

    J. Cech, J. Matas, and M. Perdoch, “Efficient sequential corre- spondence selection by cosegmentation,” TP AMI, vol. 32, no. 9, pp. 1568–1581, 2010

    2010

  7. [15]

    A vote-and-verify strategy for fast spatial verification in image retrieval,

    J. L. Schönberger, T. Price, T. Sattler, J.-M. Frahm, and M. Pollefeys, “A vote-and-verify strategy for fast spatial verification in image retrieval,” in ACCV, pp. 321–337, Springer, 2017

    2017

  8. [16]

    Su- perglue: Learning feature matching with graph neural networks,

    P .-E. Sarlin, D. DeTone, T. Malisiewicz, and A. Rabinovich, “Su- perglue: Learning feature matching with graph neural networks,” in CVPR, pp. 4938–4947, 2020

    2020

  9. [17]

    Photo tourism: exploring photo collections in 3d,

    N. Snavely, S. M. Seitz, and R. Szeliski, “Photo tourism: exploring photo collections in 3d,” in SIGGRAPH, pp. 835–846, 2006

    2006

  10. [18]

    Discrete-continuous optimization for large-scale structure from motion,

    D. Crandall, A. Owens, N. Snavely, and D. Huttenlocher, “Discrete-continuous optimization for large-scale structure from motion,” in CVPR, pp. 3001–3008, IEEE, 2011

    2011

  11. [19]

    Structure-from-motion revis- ited,

    J. L. Schonberger and J.-M. Frahm, “Structure-from-motion revis- ited,” in CVPR, pp. 4104–4113, 2016

    2016

  12. [20]

    Accurate, dense, and robust multiview stereopsis,

    Y. Furukawa and J. Ponce, “Accurate, dense, and robust multiview stereopsis,” IEEE TP AMI, vol. 32, no. 8, pp. 1362–1376, 2009

    2009

  13. [21]

    Pixel- wise view selection for unstructured multi-view stereo,

    J. L. Schönberger, E. Zheng, J.-M. Frahm, and M. Pollefeys, “Pixel- wise view selection for unstructured multi-view stereo,” in ECCV, pp. 501–518, Springer, 2016

    2016

  14. [22]

    Multi-scale geometric consistency guided multi-view stereo,

    Q. Xu and W. Tao, “Multi-scale geometric consistency guided multi-view stereo,” in CVPR, pp. 5483–5492, 2019

    2019

  15. [23]

    Generalized binary search network for highly-efficient multi-view stereo,

    Z. Mi, C. Di, and D. Xu, “Generalized binary search network for highly-efficient multi-view stereo,” in CVPR, pp. 12991–13000, 2022

    2022

  16. [24]

    Mvsnet: Depth inference for unstructured multi-view stereo,

    Y. Yao, Z. Luo, S. Li, T. Fang, and L. Quan, “Mvsnet: Depth inference for unstructured multi-view stereo,” in ECCV, pp. 767– 783, 2018

    2018

  17. [25]

    Point-based multi-view stereo network,

    R. Chen, S. Han, J. Xu, and H. Su, “Point-based multi-view stereo network,” in ICCV, pp. 1538–1547, 2019

    2019

  18. [26]

    Cascade cost vol- ume for high-resolution multi-view stereo and stereo matching,

    X. Gu, Z. Fan, S. Zhu, Z. Dai, F. Tan, and P . Tan, “Cascade cost vol- ume for high-resolution multi-view stereo and stereo matching,” in CVPR, pp. 2495–2504, 2020

    2020

  19. [27]

    Learning inverse depth regression for multi- view stereo with correlation cost volume,

    Q. Xu and W. Tao, “Learning inverse depth regression for multi- view stereo with correlation cost volume,” in AAAI, vol. 34, pp. 12508–12515, 2020

    2020

  20. [28]

    Nerf++: An- alyzing and improving neural radiance fields,

    K. Zhang, G. Riegler, N. Snavely, and V . Koltun, “Nerf++: An- alyzing and improving neural radiance fields,” arXiv preprint arXiv:2010.07492, 2020

    2010 arXiv

  21. [29]

    Nerf in the wild: Neural radiance fields for unconstrained photo collections,

    R. Martin-Brualla, N. Radwan, M. S. Sajjadi, J. T. Barron, A. Doso- vitskiy, and D. Duckworth, “Nerf in the wild: Neural radiance fields for unconstrained photo collections,” in CVPR, 2021

    2021

  22. [30]

    Mip-nerf: A multiscale represen- tation for anti-aliasing neural radiance fields,

    J. T. Barron, B. Mildenhall, M. Tancik, P . Hedman, R. Martin- Brualla, and P . P . Srinivasan, “Mip-nerf: A multiscale represen- tation for anti-aliasing neural radiance fields,” in ICCV, 2021

    2021

  23. [31]

    Mip-nerf 360: Unbounded anti-aliased neural radiance fields,

    J. T. Barron, B. Mildenhall, D. Verbin, P . P . Srinivasan, and P . Hed- man, “Mip-nerf 360: Unbounded anti-aliased neural radiance fields,” in CVPR, pp. 5470–5479, 2022

    2022

  24. [32]

    Depth-supervised NeRF: Fewer views and faster training for free,

    K. Deng, A. Liu, J.-Y. Zhu, and D. Ramanan, “Depth-supervised NeRF: Fewer views and faster training for free,” in CVPR, June 2022

    2022

  25. [33]

    Sparf: Neural radiance fields from sparse and noisy poses,

    P . Truong, M.-J. Rakotosaona, F. Manhardt, and F. Tombari, “Sparf: Neural radiance fields from sparse and noisy poses,” CVPR, 2023

    2023

  26. [34]

    Cvt-xrf: Contrastive in-voxel transformer for 3d consistent radiance fields from sparse inputs,

    Y. Zhong, L. Hong, Z. Li, and D. Xu, “Cvt-xrf: Contrastive in-voxel transformer for 3d consistent radiance fields from sparse inputs,” in CVPR, 2024

    2024

  27. [35]

    Neus: Learning neural implicit surfaces by volume rendering for multi-view reconstruction,

    P . Wang, L. Liu, Y. Liu, C. Theobalt, T. Komura, and W. Wang, “Neus: Learning neural implicit surfaces by volume rendering for multi-view reconstruction,” NeurIPS, 2021

    2021

  28. [36]

    Neuralangelo: High-fidelity neural surface reconstruction,

    Z. Li, T. Müller, A. Evans, R. H. Taylor, M. Unberath, M.-Y. Liu, and C.-H. Lin, “Neuralangelo: High-fidelity neural surface reconstruction,” in CVPR, 2023

    2023

  29. [37]

    Structure-from-motion revis- ited,

    J. L. Schönberger and J.-M. Frahm, “Structure-from-motion revis- ited,” in CVPR, 2016

    2016

  30. [38]

    Nerf–: Neural radiance fields without known camera parameters,

    Z. Wang, S. Wu, W. Xie, M. Chen, and V . A. Prisacariu, “Nerf–: Neural radiance fields without known camera parameters,” arXiv preprint arXiv:2102.07064, 2021

    2021 arXiv

  31. [39]

    Barf: Bundle- adjusting neural radiance fields,

    C.-H. Lin, W.-C. Ma, A. Torralba, and S. Lucey, “Barf: Bundle- adjusting neural radiance fields,” in IEEE International Conference on Computer Vision (ICCV) , 2021

    2021

  32. [40]

    PlenOc- trees for real-time rendering of neural radiance fields,

    A. Yu, R. Li, M. Tancik, H. Li, R. Ng, and A. Kanazawa, “PlenOc- trees for real-time rendering of neural radiance fields,” in ICCV, 2021

    2021

  33. [41]

    Fastnerf: High-fidelity neural rendering at 200fps,

    S. J. Garbin, M. Kowalski, M. Johnson, J. Shotton, and J. Valentin, “Fastnerf: High-fidelity neural rendering at 200fps,” inICCV, 2021

    2021

  34. [42]

    Kilonerf: Speeding up neural radiance fields with thousands of tiny mlps,

    C. Reiser, S. Peng, Y. Liao, and A. Geiger, “Kilonerf: Speeding up neural radiance fields with thousands of tiny mlps,” inICCV, 2021

    2021

  35. [43]

    Plenoxels: Radiance fields without neural net- works,

    S. Fridovich-Keil, A. Yu, M. Tancik, Q. Chen, B. Recht, and A. Kanazawa, “Plenoxels: Radiance fields without neural net- works,” in CVPR, pp. 5501–5510, June 2022

    2022

  36. [44]

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

    B. Kerbl, G. Kopanas, T. Leimkühler, and G. Drettakis, “3d gaussian splatting for real-time radiance field rendering,” ACM T ransactions on Graphics, 2023

    2023

  37. [45]

    pixelnerf: Neural radiance fields from one or few images,

    A. Yu, V . Ye, M. Tancik, and A. Kanazawa, “pixelnerf: Neural radiance fields from one or few images,” in CVPR, pp. 4578–4587, 2021

    2021

  38. [46]

    Mvsnerf: Fast generalizable radiance field reconstruction from multi-view stereo,

    A. Chen, Z. Xu, F. Zhao, X. Zhang, F. Xiang, J. Yu, and H. Su, “Mvsnerf: Fast generalizable radiance field reconstruction from multi-view stereo,” in Proceedings of the IEEE/CVF international conference on computer vision , pp. 14124–14133, 2021

    2021

  39. [47]

    Nerfingmvs: Guided optimization of neural radiance fields for indoor multi- view stereo,

    Y. Wei, S. Liu, Y. Rao, W. Zhao, J. Lu, and J. Zhou, “Nerfingmvs: Guided optimization of neural radiance fields for indoor multi- view stereo,” in ICCV, 2021

    2021

  40. [48]

    Ibrnet: Learning multi-view image-based rendering,

    Q. Wang, Z. Wang, K. Genova, P . P . Srinivasan, H. Zhou, J. T. Barron, R. Martin-Brualla, N. Snavely, and T. Funkhouser, “Ibrnet: Learning multi-view image-based rendering,” in CVPR, 2021

    2021

  41. [49]

    Efficient geometry-aware 3d generative adversarial networks,

    E. R. Chan, C. Z. Lin, M. A. Chan, K. Nagano, B. Pan, S. De Mello, O. Gallo, L. J. Guibas, J. Tremblay, S. Khamis, et al. , “Efficient geometry-aware 3d generative adversarial networks,” in CVPR, pp. 16123–16133, 2022

    2022

  42. [50]

    Lrm: Large reconstruction model for single image to 3d,

    Y. Hong, K. Zhang, J. Gu, S. Bi, Y. Zhou, D. Liu, F. Liu, K. Sunkavalli, T. Bui, and H. Tan, “Lrm: Large reconstruction model for single image to 3d,” in ICLR, 2024

    2024

  43. [51]

    In- teractive3d: Create what you want by interactive 3d generation,

    S. Dong, L. Ding, Z. Huang, Z. Wang, T. Xue, and D. Xu, “In- teractive3d: Create what you want by interactive 3d generation,” JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2015 18 in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. ...

    2015

  44. [52]

    SPIn-NeRF: Multiview segmentation and perceptual inpainting with neural radiance fields,

    A. Mirzaei, T. Aumentado-Armstrong, K. G. Derpanis, J. Kelly, M. A. Brubaker, I. Gilitschenski, and A. Levinshtein, “SPIn-NeRF: Multiview segmentation and perceptual inpainting with neural radiance fields,” in CVPR, 2023

    2023

  45. [53]

    Gscream: Learning 3d geometry and feature consistent gaussian splatting for object removal,

    Y. Wang, Q. Wu, G. Zhang, and D. Xu, “Gscream: Learning 3d geometry and feature consistent gaussian splatting for object removal,” in ECCV, 2024

    2024

  46. [54]

    Efficient large-scale scene rep- resentation with a hybrid of high-resolution grid and plane fea- tures,

    Y. Zhang, G. Chen, and S. Cui, “Efficient large-scale scene rep- resentation with a hybrid of high-resolution grid and plane fea- tures,” arXiv preprint arXiv:2303.03003 , 2023

    2023 arXiv

  47. [55]

    3d gaus- sian splatting for real-time radiance field rendering,

    B. Kerbl, G. Kopanas, T. Leimkühler, and G. Drettakis, “3d gaus- sian splatting for real-time radiance field rendering,” T oG, pp. 1– 14, 2023

    2023

  48. [56]

    A hierarchical 3d gaussian representation for real- time rendering of very large datasets,

    B. Kerbl, A. Meuleman, G. Kopanas, M. Wimmer, A. Lanvin, and G. Drettakis, “A hierarchical 3d gaussian representation for real- time rendering of very large datasets,” ACM T oG, vol. 43, no. 4, 2024

    2024

  49. [57]

    Citygaussian: Real-time high-quality large-scale scene rendering with gaussians,

    Y. Liu, H. Guan, C. Luo, L. Fan, N. Wang, J. Peng, and Z. Zhang, “Citygaussian: Real-time high-quality large-scale scene rendering with gaussians,” arXiv preprint arXiv:2404.01133 , 2024

    2024 arXiv

  50. [58]

    Dogaussian: Distributed-oriented gaus- sian splatting for large-scale 3d reconstruction via gaussian con- sensus,

    Y. Chen and G. H. Lee, “Dogaussian: Distributed-oriented gaus- sian splatting for large-scale 3d reconstruction via gaussian con- sensus,” arXiv preprint arXiv:2405.13943 , 2024

    2024 arXiv

  51. [59]

    Vastgaussian: Vast 3d gaussians for large scene reconstruction,

    J. Lin, Z. Li, X. Tang, J. Liu, S. Liu, J. Liu, Y. Lu, X. Wu, S. Xu, Y. Yan, and W. Yang, “Vastgaussian: Vast 3d gaussians for large scene reconstruction,” in CVPR, 2024

    2024

  52. [60]

    Pygs: Large-scale scene representation with pyramidal 3d gaussian splatting,

    Z. Wang and D. Xu, “Pygs: Large-scale scene representation with pyramidal 3d gaussian splatting,” arXiv preprint arXiv:2405.16829 , 2024

    2024 arXiv

  53. [61]

    Occlugaussian: Occlusion-aware gaussian splatting for large scene reconstruction and rendering,

    S. Liu, X. Tang, Z. Li, Y. He, C. Ye, J. Liu, B. Huang, S. Zhou, and X. Wu, “Occlugaussian: Occlusion-aware gaussian splatting for large scene reconstruction and rendering,” arXiv preprint arXiv:2503.16177, 2025

    2025

  54. [62]

    Adaptive mixtures of local experts,

    R. A. Jacobs, M. I. Jordan, S. J. Nowlan, and G. E. Hinton, “Adaptive mixtures of local experts,” Neural computation , vol. 3, no. 1, pp. 79–87, 1991

    1991

  55. [63]

    Hierarchical mixtures of experts and the em algorithm,

    M. I. Jordan and R. A. Jacobs, “Hierarchical mixtures of experts and the em algorithm,” Neural computation, vol. 6, no. 2, pp. 181– 214, 1994

    1994

  56. [64]

    Gshard: Scaling giant models with conditional computation and automatic sharding,

    D. Lepikhin, H. Lee, Y. Xu, D. Chen, O. Firat, Y. Huang, M. Krikun, N. Shazeer, and Z. Chen, “Gshard: Scaling giant models with conditional computation and automatic sharding,” in ICLR, 2021

    2021

  57. [65]

    Scaling vision with sparse mixture of experts,

    C. Riquelme, J. Puigcerver, B. Mustafa, M. Neumann, R. Jenatton, A. S. Pinto, D. Keysers, and N. Houlsby, “Scaling vision with sparse mixture of experts,” in NeurIPS, 2021

    2021

  58. [66]

    Taskexpert: Dynamically assembling multi-task representations with memorial mixture-of-experts,

    H. Ye and D. Xu, “Taskexpert: Dynamically assembling multi-task representations with memorial mixture-of-experts,” in ICCV, 2023

    2023

  59. [67]

    Einops: Clear and reliable tensor manipulations with einstein-like notation,

    A. Rogozhnikov, “Einops: Clear and reliable tensor manipulations with einstein-like notation,” in ICLR, 2021

    2021

  60. [68]

    Hash layers for large sparse models,

    S. Roller, S. Sukhbaatar, J. Weston, et al. , “Hash layers for large sparse models,” NeurIPS, 2021

    2021

  61. [69]

    Base layers: Simplifying training of large, sparse models,

    M. Lewis, S. Bhosale, T. Dettmers, N. Goyal, and L. Zettlemoyer, “Base layers: Simplifying training of large, sparse models,” in ICML, pp. 6265–6274, PMLR, 2021

    2021

  62. [70]

    Mixture-of-experts with expert choice routing,

    Y. Zhou, T. Lei, H. Liu, N. Du, Y. Huang, V . Zhao, A. M. Dai, Q. V . Le, J. Laudon, et al. , “Mixture-of-experts with expert choice routing,” Advances in Neural Information Processing Systems , vol. 35, pp. 7103–7114, 2022

    2022

  63. [71]

    Mesh- tensorflow: Deep learning for supercomputers,

    N. Shazeer, Y. Cheng, N. Parmar, D. Tran, A. Vaswani, P . Koanan- takool, P . Hawkins, H. Lee, M. Hong, C. Young, et al. , “Mesh- tensorflow: Deep learning for supercomputers,” NeurIPS, 2018

    2018

  64. [72]

    Deepspeed-moe: Advancing mixture- of-experts inference and training to power next-generation AI scale,

    S. Rajbhandari, C. Li, Z. Yao, M. Zhang, R. Y. Aminabadi, A. A. Awan, J. Rasley, and Y. He, “Deepspeed-moe: Advancing mixture- of-experts inference and training to power next-generation AI scale,” in ICML, 2022

    2022

  65. [73]

    Tutel: Adaptive mixture-of-experts at scale,

    C. Hwang, W. Cui, Y. Xiong, Z. Yang, Z. Liu, H. Hu, Z. Wang, R. Salas, J. Jose, P . Ram, et al., “Tutel: Adaptive mixture-of-experts at scale,” arXiv preprint arXiv:2206.03382 , 2022

    2022 arXiv

  66. [74]

    Mip-nerf 360: Unbounded anti-aliased neural radiance fields,

    J. T. Barron, B. Mildenhall, D. Verbin, P . P . Srinivasan, and P . Hed- man, “Mip-nerf 360: Unbounded anti-aliased neural radiance fields,” in CVPR, 2022

    2022

  67. [75]

    Deepseek-v3 technical report,

    A. Liu, B. Feng, B. Xue, B. Wang, B. Wu, C. Lu, C. Zhao, C. Deng, C. Zhang, C. Ruan, et al. , “Deepseek-v3 technical report,” arXiv preprint arXiv:2412.19437, 2024

    2024 arXiv

  68. [76]

    Ma- trixcity: A large-scale city dataset for city-scale neural rendering and beyond,

    Y. Li, L. Jiang, L. Xu, Y. Xiangli, Z. Wang, D. Lin, and B. Dai, “Ma- trixcity: A large-scale city dataset for city-scale neural rendering and beyond,” arXiv e-prints, pp. arXiv–2308, 2023

    2023

  69. [77]

    Image quality assessment: from error visibility to structural similarity,

    Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P . Simoncelli, “Image quality assessment: from error visibility to structural similarity,” TIP, 2004

    2004

  70. [78]

    The unreasonable effectiveness of deep features as a perceptual metric,

    R. Zhang, P . Isola, A. A. Efros, E. Shechtman, and O. Wang, “The unreasonable effectiveness of deep features as a perceptual metric,” in CVPR, 2018

    2018

  71. [79]

    3d is here: Point cloud library (pcl),

    R. B. Rusu and S. Cousins, “3d is here: Point cloud library (pcl),” in ICRA, IEEE, 2011

    2011

  72. [80]

    Zip-nerf: Anti-aliased grid-based neural radiance fields,

    J. T. Barron, B. Mildenhall, D. Verbin, P . P . Srinivasan, and P . Hed- man, “Zip-nerf: Anti-aliased grid-based neural radiance fields,” in ICCV, pp. 19697–19705, 2023. Zhenxing Mi is a Ph.D. student in the De- partment of Computer Science and Engineer- ing at Hong Kong Unive...

    2023

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.