Pith. sign in

REVIEW 4 major objections 5 minor 64 references

3D Gaussian Splatting Data Compression with Mixture of Priors

T0 review · 4 major / 5 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 The paper claims that replacing a single hyperprior MLP with a gated mixture of lightweight prior MLPs, and quantizing each attribute element with its own gradient-guided step, yields top rate-distortion performance for 3D Gaussian…

desk verdict Credible incremental gains in 3DGS compression from a gated mixture-of-priors and gradient-weighted quantization, but the decoder side of the quantization matrix is under-specified and the ablation text contradicts its own table. read the letter →

arxiv 2505.03310 v2 pith:SFJYTZRX submitted 2025-05-06 cs.CV

Lei Liu , Zhenghao Chen , Dong Xu This is my paper · ORCID
classification cs.CV
keywords 3DGaussianSplattingdatacompressionMixtureofPriorsCoarse-to-FineQuantizationhyperpriorentropymodelanchor-based3DGSelement-wiserate-distortionoptimizationGumbel-Softmax
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

The paper claims that two standard bottlenecks in 3D Gaussian Splatting compression—a weak hyperprior for entropy coding and a single shared quantization step for lossy coding—can be fixed by one mechanism. It replaces the single hyperprior MLP used in earlier anchor-based codecs with a Mixture of Priors (MoP): several lightweight MLPs produce diverse prior features, and a gating network blends them into a unified MoP feature. That feature serves both as the hyperprior for arithmetic coding and as guidance for a Coarse-to-Fine Quantization (C2FQ) stage that expands a scalar quantization step into an element-wise matrix. On Mip-NeRF360, BungeeNeRF, DeepBlending, and Tanks&Temples, the reported rate-distortion results beat the HAC and Context-GS baselines, with BDBR savings of roughly 16–42% depending on dataset. If these results hold, one shared module can improve both the lossless entropy-coding stage and the lossy quantization stage of anchor-based 3DGS codecs.

What carries the argument

The central object is the Mixture of Priors (MoP) feature, $G = \sum_{i=1}^n w_i \mathbf{p}_i$, where each $\mathbf{p}_i$ is the output of a distinct lightweight MLP applied to interpolated anchor locations and $w_i$ comes from a small gating network. This feature carries the argument because it is the shared source for both entropy modeling and quantization guidance. The second mechanism, Coarse-to-Fine Quantization (C2FQ), turns one scalar step $Q_0$ into a quantization matrix by selecting a scale via Gumbel-Softmax/Max, expanding the rescaled value into a per-anchor vector through an MLP, and then expanding that vector into a per-element matrix using averaged gradients across camera views, avoiding extra network parameters.

What would settle it

Take the trained model and re-encode the test scenes with the per-element gradient weights used to build $\mathbf{Q}_4$ randomly permuted, leaving all other parameters and the bitstream pipeline unchanged. If the size-PSNR curve barely moves, the claimed gradient-based element-wise mechanism is not the source of the C2FQ improvement; if it degrades sharply, the mechanism is confirmed.

Watch

Extended reading notes

Core claim

The paper's central claim is that a hyperprior built from several diverse lightweight priors, fused by a learned gate, gives a more accurate conditional distribution for entropy coding than the single shallow MLPs used in HAC and Context-GS, and that the same feature can drive per-element quantization. Quantization proceeds coarse-to-fine: a predefined step is first rescaled by a Gumbel-Softmax-selected scale, expanded by an MLP into a per-anchor vector, then multiplied by averaged per-element gradients to form a quantization matrix $\mathbf{Q}_4$. Attributes are quantized as $\hat{A} = \mathrm{Round}(A \times \mathbf{Q}_4)/\mathbf{Q}_4$, so each element of every anchor receives its own step size. The paper reports that this design reduces storage by more than 97% relative to unquantized 3DGS and by over 88% relative to Scaffold-GS, while outperforming HAC and Context-GS in both size and fidelity across four benchmarks.

Load-bearing premise

The load-bearing premise is that an element's average gradient magnitude across camera views measures how much its quantization step matters for the final rate-distortion trade-off; if that link fails, the reported C2FQ gains could come from the coarse scale selection or the per-anchor vector stage rather than from element-wise granularity.

Editorial extensions

If this is right

  • Because the MoP feature serves both the entropy model and the quantizer, the lossless and lossy branches of an anchor-based 3DGS codec can be improved by a single change rather than two separate modules.
  • Element-wise quantization lets the rate-distortion trade-off be tuned at the level of individual anchor attributes, which is finer than the anchor-level or attribute-level control in HAC and Context-GS.
  • Keeping the expansion network-free via gradients means the per-element quantization matrix adds no stored model parameters, so its storage overhead is confined to the small MoP MLPs.
  • The reported BDBR reductions of roughly 16–42% against HAC and Context-GS on the tested datasets imply that substantial storage savings are available without sacrificing rendering PSNR.
  • The same anchor-based pipeline with MoP and C2FQ can be trained end-to-end, because the Gumbel-Softmax approximation keeps the scale selection differentiable during training.

Reading between the lines

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

  • The paper does not isolate how much of the C2FQ gain comes from the gradient-based matrix versus the coarser scale and vector stages; its Tank&Temples ablation suggests the matrix contributes less than the vector, so the element-wise gradient weighting is the least-tested link.
  • If the gradient-importance premise transfers, the same quantization-matrix recipe could be attached to other hyperprior-based 3DGS codecs, or to 4D Gaussian Splatting compression, without retraining their entropy models.
  • The gating weights visualized across BungeeNeRF scenes hint that the MoP feature specializes per scene; a direct test would be whether scene-adaptive gating alone, with fixed experts, explains most of the entropy-model gain.
  • A useful stress test is to compare the five-expert MoP against a single MLP with the same total parameter count; if that match performs equally well, diversity rather than capacity is the active ingredient.
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 / 5 minor

Summary. The manuscript proposes an anchor-based 3DGS compression pipeline with two novel components: a Mixture of Priors (MoP) hyperprior and a Coarse-to-Fine Quantization (C2FQ) module. Multiple lightweight MLPs produce diverse prior features that are gated into a single MoP feature; this feature conditions the entropy model for lossless coding and guides an element-wise quantization step matrix for lossy coding. The method is evaluated on Mip-NeRF360, BungeeNeRF, DeepBlending, and Tank&Temples against a broad set of baselines, reporting state-of-the-art rate-distortion numbers in Table 1 and BDBR gains of 16.6-41.8 percent over HAC and Context-GS on the two datasets in Table 3.

Significance. The contribution is potentially useful: the size/quality table and BDBR comparisons, if reproducible, indicate a meaningful improvement over strong anchor-based compression baselines. Strengths of the paper are the broad benchmark coverage, the direct comparison to HAC and Context-GS, the storage breakdown in Table 5, and the parameter-size comparison in Table 4. However, the central codec description omits how the quantization matrix is made available to the decoder and whether its bit cost is counted, and the ablation narrative in Section 4.4 is not consistent with Table 2. These issues must be resolved before the state-of-the-art claim can be accepted.

major comments (4)
  1. [Sec. 3.2/3.4 and Fig. 2/Table 5] The quantization matrix Q4 is computed from averaged element-wise gradients of the training loss across multiple camera views (Sec. 3.4) and is needed to dequantize \hat A = Round(A*Q4)/Q4. The paper does not state that Q4 is transmitted as side information, nor that the decoder can recompute it. At decode time only the bitstream and compressed locations are available; the original training views and ground-truth rendering loss are not. Without Q4 the arithmetic decoder cannot determine the quantization bin boundaries or invert the quantization. If Q4 is side information, its bit cost must be included in the storage breakdown in Table 5 and in the BDBR numbers; if it is recomputed, the required reference data must be specified. As written, the lossless decoding stage in Sec. 3.2 is not an implementable procedure.
  2. [Sec. 4.4, Table 2] The text says that 'Ours w/o C2FQ & MoP' compared with 'Ours w/o C2FQ' yields a further decrease of 0.05 PSNR, 0.005 SSIM, and a 0.57 MB increase in storage. Table 2 shows differences of 0.17 dB, 0.003, and 0.03 MB, respectively. The additive decomposition also does not follow from the table: removing MoP alone costs 0.22 dB and 0.10 MB, removing C2FQ alone costs 0.06 dB and 0.13 MB, but the combined ablation costs 0.23 dB and 0.16 MB relative to the complete model. Please correct the narrative and discuss the interaction between the two components rather than claiming independent gains.
  3. [Sec. 4.3, Table 1] The statement that compared with Scaffold-GS 'our approach achieves over 88% storage savings and consistently delivers better reconstruction quality' is not supported by Table 1 for the low-rate configuration: on DeepBlending, Ours (low-rate) has PSNR 30.20 vs. Scaffold-GS 30.21 and LPIPS 0.260 vs. 0.254. Please restrict the claim to the high-rate configuration or qualify which operating point is meant.
  4. [Sec. 3.4, Table 6] The core assumption that gradient magnitude measures an element's importance to rate-distortion performance is not directly tested. Table 6 only removes the quantization matrix and/or vector; it does not compare the gradient-based weighting against equal weights, random weights, or a learned per-element matrix at matched storage. Please add such a control experiment and report the normalization/clamping of the averaged gradients and the contents of the scale list. Without this, the reported C2FQ gains cannot be attributed to the gradient-guided granularity rather than to the scale selection or the MoP guidance.
minor comments (5)
  1. [Sec. 4.3, Table 3] BDBR is reported for only Mip-NeRF360 and DeepBlending, although the table caption and the abstract imply all four benchmarks; please state this explicitly or add the missing datasets.
  2. [Sec. 4.2] Key hyperparameters are not reported: the loss weight lambda in Eq. (2), the predefined quantization step Q0, the scale list, and the MLP/gating architecture dimensions. Without these values the experiments are not reproducible.
  3. [Sec. 3.3, Fig. 4] The claim that random initialization yields diverse priors is only illustrated by a weight visualization; a quantitative diversity measure or a same-capacity single-MLP baseline would be more convincing.
  4. [Throughout] There are several minor typos, including 'interploated' (Sec. 3.3), 'lightweigth' (Sec. 4.4), 'Gubmel' (Sec. 3.4), and a missing comma in 'Different with previous 3DGS compression networks [4,51] that employ a single MLP'.
  5. [Fig. 3] Figure 3 omits Tank&Temples despite Table 1 including it; either add that dataset to the RD curves or explain why it is excluded.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported gains are empirical RD comparisons on external benchmarks, and MoP/C2FQ are defined independently of the measured outputs.

full rationale

Walking the derivation chain, the central claims are empirical rate-distortion comparisons against HAC and Context-GS on four public benchmarks. The MoP feature is defined in Eq. (1) as a gated sum of lightweight MLP outputs over interpolated compressed locations; it is not defined in terms of the entropy-coded bitrate or the rendered quality. The lossless entropy model uses p(A|G) as a standard conditional prior, and Lanchor in Eq. (2) is the estimated storage cost from that model, which is a training objective rather than a relabeled evaluation metric. C2FQ constructs Q4 from the predefined step Q0, a scale-list selection, the MoP feature, and averaged attribute gradients; although the gradient collection procedure in Sec. 4.2 is iterative, it is an encoder-side adaptive quantization heuristic, not a fitted parameter renamed as a prediction. Table 5 reports actual component sizes, and the BDBR numbers are computed against external baselines. The paper does not invoke any uniqueness theorem, and while it cites prior work by the same authors, those citations are background compression methods rather than the load-bearing justification for the RD gains. The one substantive concern, that Sec. 3.4 does not state how Q4 is transmitted or recomputed at the decoder, is an implementation or under-specification issue, not circularity: Q4 is not defined in terms of the outputs it predicts. No equation reduces to its own input, so the circularity score is 0.

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

The central claim rests on standard information-theoretic assumptions, inherited 3DGS and anchor pipeline assumptions, and two method-specific heuristics: gradient magnitude as an importance proxy and initialization-based diversity. The key unstated choices are the quantization scale list, the base step Q0, the loss weight lambda, and the number of expert MLPs, all of which influence the reported rate-distortion points.

free parameters (4)
  • Number of MoP expert MLPs = 5
    Empirically set in Section 4.2 to balance storage overhead and prior diversity; the reported gains depend on this choice.
  • Loss balancing weight lambda = not reported
    Appears in Eq. (2) as the weight on estimated anchor bitrate; no value or schedule is given, and it controls the rate-distortion operating point.
  • Predefined quantization step Q0 = not reported
    Starting step rescaled by scale s in C2FQ; value determines quantization granularity and is not specified.
  • C2FQ scale list = not reported
    Gumbel-Softmax selects a scale from this list; the list contents are not stated, making the coarse-to-fine stage underspecified.
assumptions (4)
  • standard math Cross-entropy H(q,p) lower-bounds the bitrate of entropy coding given estimated distribution p.
    Section 3.1 Preliminaries; standard information-theoretic result used to justify training the entropy model.
  • domain assumption Anchor-based 3DGS representation and HAC-style interpolation are a valid base for compression.
    Sections 3.2 and 3.3; the framework inherits Scaffold-GS anchors and HAC's interpolation without re-validation.
  • ad hoc to paper Gradient magnitude of an attribute element measures its importance to rate-distortion loss.
    Section 3.4 states this relationship to justify the gradient-weighted quantization matrix; no supporting analysis or ablation isolating this assumption is provided.
  • ad hoc to paper Randomly different initialization yields diverse prior features.
    Section 3.3 relies on distinct initialization to create diversity; only qualitative weight visualizations in Figure 4 are offered as evidence.

how reviews work

0 comments
Cite this review

Pith. "Pith review of 3D Gaussian Splatting Data Compression with Mixture of Priors." pith.science (2026). https://pith.science/paper/SFJYTZRX

@misc{pith2026250503310,
  author       = {Pith},
  title        = {Pith review of: 3D Gaussian Splatting Data Compression with Mixture of Priors},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SFJYTZRX}},
  note         = {Machine review of arXiv:2505.03310}
}
read the original abstract

3D Gaussian Splatting (3DGS) data compression is crucial for enabling efficient storage and transmission in 3D scene modeling. However, its development remains limited due to inadequate entropy models and suboptimal quantization strategies for both lossless and lossy compression scenarios, where existing methods have yet to 1) fully leverage hyperprior information to construct robust conditional entropy models, and 2) apply fine-grained, element-wise quantization strategies for improved compression granularity. In this work, we propose a novel Mixture of Priors (MoP) strategy to simultaneously address these two challenges. Specifically, inspired by the Mixture-of-Experts (MoE) paradigm, our MoP approach processes hyperprior information through multiple lightweight MLPs to generate diverse prior features, which are subsequently integrated into the MoP feature via a gating mechanism. To enhance lossless compression, the resulting MoP feature is utilized as a hyperprior to improve conditional entropy modeling. Meanwhile, for lossy compression, we employ the MoP feature as guidance information in an element-wise quantization procedure, leveraging a prior-guided Coarse-to-Fine Quantization (C2FQ) strategy with a predefined quantization step value. Specifically, we expand the quantization step value into a matrix and adaptively refine it from coarse to fine granularity, guided by the MoP feature, thereby obtaining a quantization step matrix that facilitates element-wise quantization. Extensive experiments demonstrate that our proposed 3DGS data compression framework achieves state-of-the-art performance across multiple benchmarks, including Mip-NeRF360, BungeeNeRF, DeepBlending, and Tank&Temples.

Figures

Figures reproduced from arXiv: 2505.03310 by the authors.

Figure 1
Figure 1. (a) The overview of our 3DGS compression framework, which integrates the proposed Mixture-of-Priors (MoP) [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Details of our Coarse-to-Fine Quantization mod [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. The Rate-Distortion (RD) curves on three benchmarks, including Mip-NeRF360, BungeeNeRF, and DeepBlending. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Visualization of “amsterdam”, “pompidou”, and “room” scenes from BungeeNeRF dataset, and the corresponding [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Visualization comparison between HAC and our [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

64 extracted references · 43 canonical work pages

  1. [1]

    VVC test model (VTM)

    2022. VVC test model (VTM). https://jvet.hhi.fraunhofer.de/, accessed:2024

    work page 2022

  2. [2]

    Johannes Ballé, David Minnen, Saurabh Singh, Sung Jin Hwang, and Nick John- ston. 2018. Variational image compression with a scale hyperprior. ICLR (2018)

    work page 2018

  3. [3]

    Jonathan T Barron, Ben Mildenhall, Dor Verbin, Pratul P Srinivasan, and Peter Hedman. 2022. Mip-nerf 360: Unbounded anti-aliased neural radiance fields. In CVPR. 5470–5479

    work page 2022

  4. [4]

    Yihang Chen, Qianyi Wu, Weiyao Lin, Mehrtash Harandi, and Jianfei Cai. 2024. Hac: Hash-grid assisted context for 3d gaussian splatting compression. In ECCV. Springer, 422–438

    work page 2024

  5. [5]

    Zicong Chen, Zhenghao Chen, Wei Jiang, Wei Wang, Lei Liu, and Dong Xu. 2025. 4DGS-CC: A Contextual Coding Framework for 4D Gaussian Splatting Data Compression. arXiv preprint arXiv:2504.18925 (2025)

    arXiv 2025

  6. [6]

    Zhenghao Chen, Shuhang Gu, Guo Lu, and Dong Xu. 2022. Exploiting intra- slice and inter-slice redundancy for learning-based lossless volumetric image compression. IEEE TIP 31 (2022), 1697–1707

    work page 2022

  7. [7]

    Zhenghao Chen, Guo Lu, Zhihao Hu, Shan Liu, Wei Jiang, and Dong Xu. 2022. LSVC: A Learning-Based Stereo Video Compression Framework. In CVPR. 6073– 6082

    work page 2022

  8. [8]

    Zhenghao Chen, Lucas Relic, Roberto Azevedo, Yang Zhang, Markus Gross, Dong Xu, Luping Zhou, and Christopher Schroers. 2023. Neural video compression with spatio-temporal cross-covariance transformers. In ACM MM. 8543–8551

    work page 2023

Show all 64 references
  1. [9]

    Zhenghao Chen, Luping Zhou, Zhihao Hu, and Dong Xu. 2024. Group-aware parameter-efficient updating for content-adaptive neural video compression. In ACM MM. 11022–11031

    2024

  2. [10]

    Zhengxue Cheng, Heming Sun, Masaru Takeuchi, and Jiro Katto. 2020. Learned image compression with discretized gaussian mixture likelihoods and attention modules. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 7939–7948

    2020

  3. [11]

    Jinyoung Choi and Bohyung Han. 2020. Task-aware quantization network for jpeg image compression. In ECCV. Springer, 309–324

    2020

  4. [12]

    Thomas M Cover. 1999. Elements of information theory . John Wiley & Sons

    1999

  5. [13]

    Nan Du, Yanping Huang, Andrew M Dai, Simon Tong, Dmitry Lepikhin, Yuanzhong Xu, Maxim Krikun, Yanqi Zhou, Adams Wei Yu, Orhan Firat, et al

  6. [14]

    Zhiwen Fan, Kevin Wang, Kairun Wen, Zehao Zhu, Dejia Xu, and Zhangyang Wang. 2023. Lightgaussian: Unbounded 3d gaussian compression with 15x reduction and 200+ fps. arXiv preprint arXiv:2311.17245 (2023)

    2023 arXiv

  7. [15]

    William Fedus, Barret Zoph, and Noam Shazeer. 2022. Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity. Journal of Machine Learning Research 23, 120 (2022), 1–39

    2022

  8. [16]

    Haisheng Fu, Jie Liang, Zhenman Fang, Jingning Han, Feng Liang, and Guohe Zhang. 2024. WeConvene: Learned Image Compression with Wavelet-Domain Convolution and Entropy Model. In ECCV. Springer, 37–53

    2024

  9. [17]

    Sharath Girish, Kamal Gupta, and Abhinav Shrivastava. 2023. Eagles: Effi- cient accelerated 3d gaussians with lightweight encodings. arXiv preprint arXiv:2312.04564 (2023)

    2023 arXiv

  10. [18]

    Tao Han, Zhenghao Chen, Song Guo, Wanghan Xu, and Lei Bai. 2024. Cra5: Extreme compression of era5 for portable global climate and weather research via an efficient variational transformer. arXiv preprint arXiv:2405.03376 (2024)

    2024 arXiv

  11. [19]

    Peter Hedman, Julien Philip, True Price, Jan-Michael Frahm, George Drettakis, and Gabriel Brostow. 2018. Deep blending for free-viewpoint image-based ren- dering. ACM Trans. Graph. 37, 6 (2018), 1–15

    2018

  12. [20]

    Zhihao Hu, Zhenghao Chen, Dong Xu, Guo Lu, Wanli Ouyang, and Shuhang Gu

  13. [21]

    Zhihao Hu, Guo Lu, and Dong Xu. 2021. FVC: A new framework towards deep video compression in feature space. In CVPR. 1502–1511

    2021

  14. [22]

    Robert A Jacobs, Michael I Jordan, Steven J Nowlan, and Geoffrey E Hinton. 1991. Adaptive mixtures of local experts. Neural computation 3, 1 (1991), 79–87

    1991

  15. [23]

    Eric Jang, Shixiang Gu, and Ben Poole. 2017. Categorical reparametrization with gumble-softmax. In ICLR

    2017

  16. [24]

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimkühler, and George Drettakis

  17. [25]

    Diederik P Kingma. 2015. Adam: A method for stochastic optimization. ICLR (2015)

    2015

  18. [26]

    Arno Knapitsch, Jaesik Park, Qian-Yi Zhou, and Vladlen Koltun. 2017. Tanks and temples: Benchmarking large-scale scene reconstruction. ACM Trans. Graph. 36, 4 (2017), 1–13

    2017

  19. [27]

    Jooyoung Lee, Seyoon Jeong, and Munchurl Kim. 2022. Selective compression learning of latent representations for variable-rate image compression. NeurIPS 35 (2022), 13146–13157

    2022

  20. [28]

    Joo Chan Lee, Daniel Rho, Xiangyu Sun, Jong Hwan Ko, and Eunbyung Park. 2024. Compact 3d gaussian representation for radiance field. In CVPR. 21719–21728

    2024

  21. [29]

    Zhongyue Lei, Xuemin Hong, Jianghong Shi, Minxian Su, Chaoheng Lin, and Wei Xia. 2023. Quantization-Based Adaptive Deep Image Compression Using Semantic Information. IEEE Access 11 (2023), 118061–118077

    2023

  22. [30]

    Dmitry Lepikhin, HyoukJoong Lee, Yuanzhong Xu, Dehao Chen, Orhan Firat, Yanping Huang, Maxim Krikun, Noam Shazeer, and Zhifeng Chen. 2021. Gshard: Scaling giant models with conditional computation and automatic sharding.ICLR (2021)

    2021

  23. [31]

    Jiahao Li, Bin Li, and Yan Lu. 2021. Deep Contextual Video Compression.Advances in Neural Information Processing Systems 34 (2021)

    2021

  24. [32]

    Jiahao Li, Bin Li, and Yan Lu. 2023. Neural Video Compression with Diverse Contexts. In IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2023, Vancouver, Canada, June 18-22, 2023

    2023

  25. [33]

    Jiahao Li, Bin Li, and Yan Lu. 2024. Neural Video Compression with Feature Modulation. In IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2024, Seattle, W A, USA, June 17-21, 2024

    2024

  26. [34]

    Lei Liu, Zhenghao Chen, Zhihao Hu, and Dong Xu. 2025. An Efficient Adaptive Compression Method for Human Perception and Machine Vision Tasks. arXiv preprint arXiv:2501.04329 (2025)

    2025 arXiv

  27. [35]

    Lei Liu, Zhihao Hu, and Zhenghao Chen. 2024. Towards point cloud compression for machine perception: A simple and strong baseline by learning the octree depth level predictor. In International Joint Conference on Artificial Intelligence WorkShop. Springer, 3–17

    2024

  28. [36]

    Lei Liu, Zhihao Hu, Zhenghao Chen, and Dong Xu. 2023. Icmh-net: Neural image compression towards both machine vision and human vision. InACM MM. 8047–8056

    2023

  29. [37]

    Lei Liu, Zhihao Hu, and Jing Zhang. 2023. PCHM-Net: A New Point Cloud Compression Framework for Both Human Vision and Machine Vision. In ICME. 1997–2002

    2023

  30. [38]

    Xiangrui Liu, Xinju Wu, Pingping Zhang, Shiqi Wang, Zhu Li, and Sam Kwong

  31. [39]

    Guo Lu, Wanli Ouyang, Dong Xu, Xiaoyun Zhang, Chunlei Cai, and Zhiyong Gao. 2019. Dvc: An end-to-end deep video compression framework. In CVPR. 11006–11015

    2019

  32. [40]

    Tao Lu, Mulin Yu, Linning Xu, Yuanbo Xiangli, Limin Wang, Dahua Lin, and Bo Dai. 2024. Scaffold-gs: Structured 3d gaussians for view-adaptive rendering. In CVPR. 20654–20664

    2024

  33. [41]

    Jixiang Luo, Yan Wang, and Hongwei Qin. 2024. Super-high-fidelity image compression via hierarchical-roi and adaptive quantization. arXiv preprint arXiv:2403.13030 (2024)

    2024 arXiv

  34. [42]

    David Minnen, Johannes Ballé, and George D Toderici. 2018. Joint autoregressive and hierarchical priors for learned image compression. NeurIPS 31 (2018)

    2018

  35. [43]

    Wieland Morgenstern, Florian Barthel, Anna Hilsmann, and Peter Eisert. 2024. Compact 3d scene representation via self-organizing gaussian grids. ECCV (2024)

    2024

  36. [44]

    KL Navaneet, Kossar Pourahmadi Meibodi, Soroush Abbasi Koohpayegani, and Hamed Pirsiavash. 2023. Compact3d: Compressing gaussian splat radiance field models with vector quantization. arXiv preprint arXiv:2311.18159 (2023)

    2023 arXiv

  37. [45]

    Simon Niedermayr, Josef Stumpfegger, and Rüdiger Westermann. 2024. Com- pressed 3d gaussian splatting for accelerated novel view synthesis. In CVPR. 10349–10358

    2024

  38. [46]

    Panagiotis Papantonakis, Georgios Kopanas, Bernhard Kerbl, Alexandre Lanvin, and George Drettakis. 2024. Reducing the Memory Footprint of 3D Gaussian Splatting. Proceedings of the ACM on Computer Graphics and Interactive Techniques 7, 1 (2024), 1–17

    2024

  39. [47]

    Noam Shazeer, Azalia Mirhoseini, Krzysztof Maziarz, Andy Davis, Quoc Le, Geoffrey Hinton, and Jeff Dean. 2017. Outrageously large neural networks: The sparsely-gated mixture-of-experts layer. ICLR (2017)

    2017

  40. [48]

    Gary J Sullivan, Jens-Rainer Ohm, Woo-Jin Han, and Thomas Wiegand. 2012. Overview of the high efficiency video coding (HEVC) standard. IEEE TCSVT 22, 12 (2012), 1649–1668

    2012

  41. [49]

    Kedeng Tong, Yaojun Wu, Yue Li, Kai Zhang, Li Zhang, and Xin Jin. 2023. QVRF: A Quantization-Error-Aware Variable Rate Framework for Learned Image Com- pression. In ICIP. 1310–1314

    2023

  42. [50]

    Henan Wang, Hanxin Zhu, Tianyu He, Runsen Feng, Jiajun Deng, Jiang Bian, and Zhibo Chen. 2024. End-to-end rate-distortion optimized 3d gaussian repre- sentation. In ECCV. Springer, 76–92

    2024

  43. [51]

    Yufei Wang, Zhihao Li, Lanqing Guo, Wenhan Yang, Alex Kot, and Bihan Wen

  44. [52]

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Simoncelli. 2004. Image quality assessment: from error visibility to structural similarity. IEEE TIP 13, 4 (2004), 600–612

    2004

  45. [53]

    Guoqing Xiang, Huizhu Jia, Mingyuan Yang, Yuan Li, and Xiaodong Xie. 2018. A novel adaptive quantization method for video coding. Multimedia Tools and Applications 77 (2018), 14817–14840. MM ’25, October 27–31, 2025, Dublin, Ireland Lei Liu, Zhenghao Chen, & Dong Xu

    2018

  46. [54]

    Yuanbo Xiangli, Linning Xu, Xingang Pan, Nanxuan Zhao, Anyi Rao, Christian Theobalt, Bo Dai, and Dahua Lin. 2022. Bungeenerf: Progressive neural radiance field for extreme multi-scale scene rendering. In ECCV. Springer, 106–122

    2022

  47. [55]

    Shuzhao Xie, Weixiang Zhang, Chen Tang, Yunpeng Bai, Rongwei Lu, Shijia Ge, and Zhi Wang. 2024. Mesongs: Post-training compression of 3d gaussians via efficient attribute transformation. In ECCV. Springer, 434–452

    2024

  48. [56]

    In The Thirty-eighth Annual Conference on Neural Information Processing Systems

    ContextGS : Compact 3D Gaussian Splatting with Anchor Level Context Model. In The Thirty-eighth Annual Conference on Neural Information Processing Systems. https://openreview.net/forum?id=W2qGSMl2Uu

  49. [57]

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shechtman, and Oliver Wang

  50. [58]

    Simiao Zuo, Qingru Zhang, Chen Liang, Pengcheng He, Tuo Zhao, and Weizhu Chen. 2022. MoEBERT: from BERT to Mixture-of-Experts via Importance-Guided Adaptation. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics:...

    2022

  51. [61]

    Fuzhao Xue, Ziji Shi, Futao Wei, Yuxuan Lou, Yong Liu, and Yang You. 2022. Go wider instead of deeper. In AAAI, Vol. 36. 8779–8787

    2022

  52. [2018]

    The unreasonable effectiveness of deep features as a perceptual metric. In CVPR. 586–595

  53. [2020]

    Improving deep video compression by resolution-adaptive flow coding. In ECCV. Springer, 193–209

  54. [2022]

    PMLR, 5547–5569

    Glam: Efficient scaling of language models with mixture-of-experts. PMLR, 5547–5569

  55. [2023]

    ACM Trans

    3D Gaussian Splatting for Real-Time Radiance Field Rendering. ACM Trans. Graph. 42, 4 (2023), 139–1

    2023

  56. [2024]

    In ACM MM

    CompGS: Efficient 3D Scene Representation via Compressed Gaussian Splatting. In ACM MM

Pith tools

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