Pith. sign in

REVIEW 3 major objections 5 minor 49 references

GIFStream: 4D Gaussian-based Immersive Video with Feature Stream

T0 review · 3 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 By attaching time-dependent feature streams to canonical anchors, GIFStream compresses dynamic 6-DoF scenes to a few megabytes per group of pictures and renders them in real time, with the best rate–distortion curve among 4D Gaussian…

desk verdict Solid system-level contribution to 4D Gaussian compression, but the acknowledged GOP-boundary flicker undercuts the continuous-video claim; worth a careful review with requests for temporal consistency metrics and code. read the letter →

arxiv 2505.07539 v1 pith:AI32WNSU submitted 2025-05-12 cs.CV

Hao Li , Sicheng Li , Xiang Gao , Abudouaihati Batuer , Lu Yu , Yiyi Liao This is my paper · ORCID
classification cs.CV
keywords 4DGaussiansplattingimmersivevideofeaturestreamdeformationfieldneuralcompressionentropycodingrate-distortionoptimizationnovelviewsynthesis
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

GIFStream makes the case that a 4D Gaussian representation and its compression should be designed together. The paper adds time-dependent feature streams to a deformation-based 4D Gaussian model, giving it enough capacity to represent fast motion while keeping the representation temporally aligned for compression. Because the streams are motion-adaptive, static regions drop them almost entirely, and the remaining features can be sorted into two video-like arrays and encoded end-to-end with quantization-aware training and an autoregressive entropy model. On three dynamic multi-view datasets, GIFStream reports the best rate–distortion performance among 4D Gaussian compression baselines, with about 5–10 MB per 65-frame group, real-time rendering above 60 FPS, and a stream rate near 30 Mbps.

What carries the argument

The load-bearing object is the anchor-based feature stream. Each canonical anchor, following Scaffold-GS, carries a time-independent feature $f\in\mathbb{R}^{C}$ and a time-dependent feature stream $f_t\in\mathbb{R}^{P}$; two small MLPs decode them into per-timestamp Gaussian attributes and anchor motion, with $M_{\mathrm{de}}$ scaling the stream so static anchors can drop it. The streams are motion-adaptive: they are pruned where no motion exists, which is what keeps the added parameters small. The compression side is carried by a PCA-guided 3D-to-2D sorting that rearranges anchors by position and the principal components of $f$ into two smooth video arrays, $V_{\mathrm{TI}}$ for time-independent parameters and $V_{\mathrm{GF}}$ for feature streams, so that an autoregressive entropy network can predict each frame's distribution from previous frames and rANS can entropy-code the quantized arrays. The same trained representation can be fed to conventional codecs like HEVC, at lower compression efficiency but with hardware acceleration.

What would settle it

Measure the conditional entropy of the PCA-sorted feature-stream video against the same features in unsorted or random order: if the sorted layout does not yield a substantial per-anchor bit saving under the same autoregressive context, the rate–distortion advantage over 4D Gaussian baselines would not transfer to scenes where similar anchors cannot be brought together.

Watch

Extended reading notes

Core claim

The paper's central claim is that deformation-based 4D Gaussian representations fail on fast motion because their deformation fields have too little capacity, while windowed 4D Gaussian representations waste storage because their primitives lack temporal correspondence; GIFStream resolves both problems by attaching a sparse, time-dependent feature stream $f_t$ to each canonical anchor, alongside the time-independent feature $f$. These streams are decoded into Gaussian attributes and $\mathrm{SE}(3)$ motion per timestamp, giving deformation-based methods the extra capacity to model fast motion while keeping the representation temporally aligned. A learnable mask $M_{\mathrm{de}}$ zeros the streams for static anchors, so in simple scenes only about 0.3% of anchors need streams, while hard scenes keep about 30%. The aligned structure lets both the static parameters and the feature streams be reorganized into two video-like arrays and compressed end-to-end with quantization-aware training and an autoregressive entropy model followed by rANS coding. On the Neur3D, Panoptic Sports, and MPEG datasets, GIFStream reports the best rate–distortion performance among the compared 4D Gaussian compression methods, with stored sizes near 5–10 MB per 65-frame group and rendering above 60 FPS on an RTX 4090.

Load-bearing premise

The whole compression pipeline assumes that the PCA-sorted layout turns the pruned, sparse feature streams and the static parameters into smooth, video-like arrays, so that a simple autoregressive entropy model can achieve the reported bitrates without the overhead of the pruning mask or the irregular layout eroding the gains.

Editorial extensions

If this is right

  • 6-DoF immersive video becomes streamable at bitrates comparable to 4K 2D video (about 30 Mbps) while preserving real-time rendering on consumer GPUs.
  • Deformation-based 4D Gaussian methods can model fast motion without switching to memory-hungry windowed 4D Gaussian primitives.
  • Because the representation is reorganized into video-like arrays, standard codecs such as HEVC can compress it too, making the format compatible with existing hardware decoders at some rate–distortion cost.
  • Storage scales automatically with scene dynamics: static scenes keep almost no feature streams, so the same framework adapts from near-static to fast-moving content.

Reading between the lines

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

  • If the central claim holds, the anchor-plus-feature-stream design could be grafted onto other primitive-based dynamic representations, such as per-frame Scaffold-GS, to obtain temporally consistent compression without a learned deformation field.
  • The separation into time-independent and time-dependent streams suggests a progressive-decode path: a decoder could fetch only part of the feature-stream channels and trade a little fidelity for bandwidth, which the paper does not demonstrate.
  • One testable extension is to replace the Gaussian-likelihood entropy model with a transformer-based context, which might close the remaining gap to dedicated video codecs on the smooth, video-like arrays.
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 / 5 minor

Summary. GIFStream proposes a 4D Gaussian representation for dynamic free-viewpoint video, augmenting Scaffold-GS anchors with time-dependent feature streams in addition to time-independent features. The representation decodes per-timestamp Gaussian attributes and SE(3) motions with small MLPs, applies motion-adaptive pruning, and is compressed by sorting anchors into a time-independent video and a time-dependent feature video followed by quantization-aware training, autoregressive entropy modeling, and rANS entropy coding. Experiments on Neur3D, Panoptic Sports, and MPEG compare against 4DGS, STG, E-D3DGS, and CSTG; the paper reports improved rate-distortion at low storage, rendering above 60 FPS on an RTX 4090, and sub-second decoding for a two-second GOP. The central claim is that this achieves high-quality immersive video at roughly 30 Mbps.

Significance. If the results hold, this is a meaningful representation-compression co-design for 4D Gaussian immersive video. The strengths are the clean formulation of feature streams with motion-adaptive sparsity, the concrete end-to-end compression pipeline with a real rANS codec, the per-component ablation, and the multi-rate evaluations. The bitrate numbers are not circular: the entropy model is trained to match the representation and the final sizes are measured by actual entropy coding. The main gap is that the abstract-level claim concerns continuous immersive video while the evaluation is per-frame, and the supplementary explicitly concedes GOP-boundary flicker without measuring it. The PCA-sorting assumption is validated only on bounded indoor scenes, but I treat that as a generalization caveat rather than a fatal flaw because the evidence on the tested datasets is consistent.

major comments (3)
  1. [Supp. Sec. 3–4; Table 1] The supplementary explicitly acknowledges flicker across GOPs in the static background, citing "flame salmon 1", and lists background inconsistencies between GOPs as a limitation. No experiment quantifies this artifact, and all reported quality metrics in Table 1 and Table 5 are per-frame averages that do not penalize inter-GOP discontinuities. Because the abstract promises "high-quality immersive video" in a continuous playback setting, the current evidence supports quality within a GOP but not the streamed-video claim. Please add a temporal-consistency evaluation, for example inter-GOP PSNR/SSIM at boundaries, t-LPIPS, or a user study, and report it with and without the previous-GOP initialization described in the supplementary.
  2. [Sec. 4.1 and Table 1; Supp. Sec. 3] All quantitative results are single runs without error bars or multiple seeds, and the supplementary states that baseline hyper-parameters on MPEG were selected by running multiple combinations and keeping the best. With one run per method and per-dataset tuning, the reader cannot assess whether the reported gains over CSTG (for example 30.72 versus 29.48 PSNR on MPEG at comparable storage) are significant or within run-to-run variation. Please report variance across at least three training runs for the main comparison and describe the baseline hyper-parameter selection procedure more quantitatively.
  3. [Abstract and Sec. 4.3; Table 1] The abstract and introduction state that the bit rate is "30 Mbps" and "comparable to that of 4K 2D videos", but Table 1 reports storage in MB per GOP and no comparison with any 2D codec bit rate is given. For the MPEG data, 7 MB per 65-frame GOP at 30 fps corresponds to about 25.8 Mbps, so the headline number is plausible, but the paper should state the exact conversion, define at which rate point the 30 Mbps figure is achieved, and either substantiate or remove the comparison to 4K 2D video.
minor comments (5)
  1. [Sec. 3.2] The sorted-2D-layout assumption is validated only on three bounded indoor datasets; a sentence discussing potential failure modes for unbounded or larger scenes would improve the generality discussion.
  2. [Sec. 4.3] The decoding-speed paragraph reports rANS rates as 5 FPS for time-independent parameters and 200 FPS for feature streams; please clarify whether these are per-GOP or per-frame rates and provide an end-to-end decoding latency budget for one GOP.
  3. [Table 5] Per-scene Neur3D results are reported only for GIFStream; adding the corresponding baseline per-scene numbers would help readers verify that the average advantage holds across all scenes.
  4. [Eq. (2) and Eq. (11)] There is a typo in Eq. (2) ("denots"), and Eq. (11) is split awkwardly across lines; please fix the formatting.
  5. [References] References [9] and [10] are the same K-Planes paper; one of the entries should be removed or given a distinct citation context.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the representation, compression, and rate-distortion evaluation are self-contained and validated against external baselines.

full rationale

The paper's derivation chain is not circular. The 4D Gaussian representation (canonical anchors plus time-dependent feature streams) is decoded by MLPs into Gaussian attributes and motion, and compression is achieved by reorganizing features into 2D videos, training an autoregressive entropy model, and performing actual rANS entropy coding. Bitrates are measured from the coded representation, not assumed from the entropy loss. The claimed result—high-quality immersive video at 30 Mbps—is supported by comparisons against external baselines (CSTG, 4DGS, STG, 4DGaussian, E-D3DGS) on standard datasets (Neur3D, Panoptic Sports, MPEG). The entropy model is trained to match the representation, which is normal learned compression practice, not a circular prediction: the reported storage sizes come from quantized and entropy-coded parameters. The only self-citation is NeRFCodec [21] in the related-work discussion, and it is not load-bearing for any central claim. The acknowledged GOP-boundary flicker (Supp. Sec. 3-4) is a correctness/limitation concern about temporal consistency, not a circularity in the derivation. No fitted parameter is renamed as a prediction, no uniqueness theorem is imported from the authors, and no ansatz is smuggled in via self-citation. The derivation is self-contained against external benchmarks.

Assumptions & free parameters 6 free parameters · 6 assumptions · 1 invented entities

The central claim rests on a standard learned-rendering and compression stack: differentiable Gaussian splatting, anchor-based canonical space, SE(3) deformation, quantization-aware training, and an autoregressive entropy model. Beyond the usual hyperparameters, the main assumptions are that the sorted 2D feature layout remains codec-friendly and that the learned entropy model accurately predicts the rANS bitrate. No new physical entities are postulated; the time-dependent feature stream is a model component, not an externally verifiable phenomenon.

free parameters (6)
  • feature stream channel count P = 4 (Neur3D), 8 (MPEG)
    Set by dataset in Sec. 4.2; larger P is used for faster-changing MPEG scenes, affecting model capacity and bitrate.
  • time-independent feature channel count C = 48 for three Neur3D scenes, 24 for others
    Chosen per scene for better quality (Sec. 4.2).
  • number of Gaussian primitives per anchor K = 5
    Fixed in all experiments (Sec. 4.2).
  • entropy loss weight lambda_e = 0.012 to 0.00025
    Adjusted to sweep rate points (Sec. 4.2).
  • regularization weight lambda_r = 0.0005
    Set in Sec. 4.2.
  • gradient accumulation weight alpha (Eq. 12) = not specified
    Weight in the densification gradient combination; value not reported in the paper, so replication requires guessing.
assumptions (6)
  • standard math 3D Gaussian Splatting provides a differentiable rendering forward model with view-dependent colors.
    Used as the rendering backbone (Sec. 3.1, from Kerbl et al.).
  • domain assumption Scaffold-GS anchor-based representation is a valid canonical space for dynamic scenes.
    The canonical space and anchor decoding follow Scaffold-GS (Sec. 3.1).
  • domain assumption SE(3) transformations with quaternion parameterization can represent the motion of Gaussian anchors.
    Motion head predicts rotation and translation in SE(3) (Eq. 6-8).
  • ad hoc to paper The sorted 2D layout of anchor features has sufficient spatial and temporal correlation for video codecs to compress effectively.
    Core compression claim; validated only empirically in Sec. 3.2 and Table 4.
  • domain assumption Quantization-aware training with a Gaussian entropy model yields bitrates close to the true rANS-coded bitstream.
    Standard in learned compression; used to compute entropy loss (Eq. 11).
  • domain assumption The appearance of the scene is captured by per-Gaussian color with zero-degree SH (or learned color), so static reconstruction from COLMAP points is a sufficient starting point.
    Initialization from first-frame COLMAP sparse points (Sec. 4.2).
invented entities (1)
  • time-dependent feature stream f_t on each anchor
    purpose: Model complex fast motion beyond a low-capacity deformation field
    A learned latent sequence attached to anchors; its benefit is shown only through in-paper ablations, with no external falsifiable prediction. The stream itself is not a physically measurable entity.

how reviews work

0 comments
Cite this review

Pith. "Pith review of GIFStream: 4D Gaussian-based Immersive Video with Feature Stream." pith.science (2026). https://pith.science/paper/AI32WNSU

@misc{pith2026250507539,
  author       = {Pith},
  title        = {Pith review of: GIFStream: 4D Gaussian-based Immersive Video with Feature Stream},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AI32WNSU}},
  note         = {Machine review of arXiv:2505.07539}
}
read the original abstract

Immersive video offers a 6-Dof-free viewing experience, potentially playing a key role in future video technology. Recently, 4D Gaussian Splatting has gained attention as an effective approach for immersive video due to its high rendering efficiency and quality, though maintaining quality with manageable storage remains challenging. To address this, we introduce GIFStream, a novel 4D Gaussian representation using a canonical space and a deformation field enhanced with time-dependent feature streams. These feature streams enable complex motion modeling and allow efficient compression by leveraging temporal correspondence and motion-aware pruning. Additionally, we incorporate both temporal and spatial compression networks for end-to-end compression. Experimental results show that GIFStream delivers high-quality immersive video at 30 Mbps, with real-time rendering and fast decoding on an RTX 4090. Project page: https://xdimlab.github.io/GIFStream

Figures

Figures reproduced from arXiv: 2505.07539 by the authors.

Figure 1
Figure 1. GIFStream achieves high quality and small storage size on dynamic scenes containing fast motion. We present the rendering results on a challenging scene on the left and the Rate-SSIM curve on the right. Abstract Immersive video offers a 6-Dof-free viewing experience, po￾tentially playing a key role in future video technology. Re￾cently, 4D Gaussian Splatting has gained attention as an effective approach for immersiv… view at source ↗
Figure 2
Figure 2. Comparison of 4D Representations. 1) Deformation￾based representation stores a 3D Gaussian in a canonical space and its deformation along a long time horizon. and 2) 4D Gaussian representation which models a windowed spacetime region. We propose 3) GIFStream by adding time-dependent feature streams on top of deformation-based representation, improving its capacity while maintaining temporal alignment for efficient c… view at source ↗
Figure 3
Figure 3. [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Motion Illustration. We predict the rotation and trans￾lations on the local coordinate system of the anchor. Gaussian Motion Prediction Head: In most scenarios, the motion of the Gaussians exhibits local smoothness. To capture this local smoothness prior, we aggregate …
Figure 5
Figure 5. Figure 5: RD Curve Comparison on MPEG dataset. We visual￾ize the RD Curve results in the GOP 65 setting. 4D Gaussian Splatting methods, while 4DGaussian and E￾D3DGS represents deformation-based methods. Datasets: We conduct experiments on three datasets: the widely used Neur3D […
Figure 6
Figure 6. Figure 6: Qualitative Comparison. We implement the 4DGaussian, CSTG and compare their reconstructed quality with ours on these scenes. Our methods achieve better quality in scenes with fast motion, see the basketball player at the bottom right corner of the 3th row. straint. Add…
Figure 7
Figure 7. Figure 7: Ablation Study about KNN Aggregation. We compare the visual quality of dynamic details in this figure. ture streams by using only the time-independent feature and positional encoding of time as the input of MLPs (w/o fea￾ture stream ft). The results show a clear degrad…
Figure 8
Figure 8. Figure 8: RD Curve Comparision on MPEG dataset. We visualize the RD Curve results in the GOP 65 setting. results for the Neur3D dataset are provided in [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: More Qualitative Evaluation. We present complete frames from the MPEG dataset along with a comparison of local details in this figure. tialization. This is due to the instability of densification. The adaptive sampling strategy from STG [23] or the con￾tinual training …
Figure 10
Figure 10. Figure 10: More Qualitative Evaluation. We present complete frames from the Neur3D dataset along with a comparison of local details in this figure [PITH_FULL_IMAGE:figures/full_fig_p014_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 31 canonical work pages

  1. [1]

    Per-gaussian embedding-based deformation for deformable 3d gaussian splatting.arXiv preprint arXiv:2404.03613, 2024

    Jeongmin Bae, Seoha Kim, Youngsik Yun, Hahyun Lee, Gun Bang, and Youngjung Uh. Per-gaussian embedding-based deformation for deformable 3d gaussian splatting.arXiv preprint arXiv:2404.03613, 2024. 1, 2, 4, 6, 7

    arXiv 2024

  2. [2]

    Variational image compression with a scale hyperprior.arXiv preprint arXiv:1802.01436,

    Johannes Ball ´e, David Minnen, Saurabh Singh, Sung Jin Hwang, and Nick Johnston. Variational image compression with a scale hyperprior.arXiv preprint arXiv:1802.01436,

    arXiv

  3. [3]

    Overview of the versatile video coding (vvc) standard and its applications

    Benjamin Bross, Ye-Kui Wang, Yan Ye, Shan Liu, Jianle Chen, Gary J Sullivan, and Jens-Rainer Ohm. Overview of the versatile video coding (vvc) standard and its applications. IEEE Transactions on Circuits and Systems for Video Tech- nology, 31(10):3736–3764, 2021. 5

    work page 2021

  4. [4]

    Hexplane: A fast representa- tion for dynamic scenes

    Ang Cao and Justin Johnson. Hexplane: A fast representa- tion for dynamic scenes. pages 130–141, 2023. 2

    work page 2023

  5. [5]

    Hac: Hash-grid assisted context for 3d gaussian splatting compression

    Yihang Chen, Qianyi Wu, Weiyao Lin, Mehrtash Harandi, and Jianfei Cai. Hac: Hash-grid assisted context for 3d gaussian splatting compression. InEuropean Conference on Computer Vision, pages 422–438. Springer, 2025. 3, 5, 1

    work page 2025

  6. [6]

    4d-rotor gaussian splatting: towards efficient novel view synthesis for dynamic scenes

    Yuanxing Duan, Fangyin Wei, Qiyu Dai, Yuhang He, Wen- zheng Chen, and Baoquan Chen. 4d-rotor gaussian splatting: towards efficient novel view synthesis for dynamic scenes. InACM SIGGRAPH 2024 Conference Papers, pages 1–11,

    2024

  7. [7]

    Asymmetric numeral systems: entropy coding combining speed of huffman coding with compression rate of arithmetic coding.arXiv preprint arXiv:1311.2540, 2013

    Jarek Duda. Asymmetric numeral systems: entropy coding combining speed of huffman coding with compression rate of arithmetic coding.arXiv preprint arXiv:1311.2540, 2013. 5

    arXiv 2013

  8. [8]

    Lightgaussian: Unbounded 3d gaussian compression with 15x reduction and 200+ fps

    Zhiwen Fan, Kevin Wang, Kairun Wen, Zehao Zhu, De- jia Xu, and Zhangyang Wang. Lightgaussian: Unbounded 3d gaussian compression with 15x reduction and 200+ fps. arXiv preprint arXiv:2311.17245, 2023. 3, 1

    arXiv 2023

Show all 49 references
  1. [9]

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

    Sara Fridovich-Keil, Giacomo Meanti, Frederik Rahbæk Warburg, Benjamin Recht, and Angjoo Kanazawa. K-planes: Explicit radiance fields in space, time, and appearance. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 12479–12488, 2023. 2

    2023

  2. [10]

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

    Sara Fridovich-Keil, Giacomo Meanti, Frederik Rahbæk Warburg, Benjamin Recht, and Angjoo Kanazawa. K-planes: Explicit radiance fields in space, time, and appearance. pages 12479–12488, 2023. 7

    2023

  3. [11]

    Gaussianflow: Splatting gaussian dynamics for 4d content creation.arXiv preprint arXiv:2403.12365, 2024

    Quankai Gao, Qiangeng Xu, Zhe Cao, Ben Mildenhall, Wen- chao Ma, Le Chen, Danhang Tang, and Ulrich Neumann. Gaussianflow: Splatting gaussian dynamics for 4d content creation.arXiv preprint arXiv:2403.12365, 2024. 1, 2, 4

    2024 arXiv

  4. [12]

    Ea- gles: Efficient accelerated 3d gaussians with lightweight en- codings.arXiv preprint arXiv:2312.04564, 2023

    Sharath Girish, Kamal Gupta, and Abhinav Shrivastava. Ea- gles: Efficient accelerated 3d gaussians with lightweight en- codings.arXiv preprint arXiv:2312.04564, 2023. 3, 1

    2023 arXiv

  5. [13]

    Sc-gs: Sparse-controlled gaussian splatting for editable dynamic scenes

    Yi-Hua Huang, Yang-Tian Sun, Ziyi Yang, Xiaoyang Lyu, Yan-Pei Cao, and Xiaojuan Qi. Sc-gs: Sparse-controlled gaussian splatting for editable dynamic scenes. InProceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4220–4230, 2024. 1, 2, 4

    2024

  6. [14]

    Panoptic studio: A massively multiview system for social motion capture

    Hanbyul Joo, Hao Liu, Lei Tan, Lin Gui, Bart Nabbe, Iain Matthews, Takeo Kanade, Shohei Nobuhara, and Yaser Sheikh. Panoptic studio: A massively multiview system for social motion capture. InProceedings of the IEEE inter- national conference on computer vision, pages 3334–3342,

  7. [15]

    3d gaussian splatting for real-time radiance field rendering.ACM Trans

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

  8. [16]

    Dynmf: Neural motion factorization for real-time dynamic view synthesis with 3d gaussian splatting

    Agelos Kratimenos, Jiahui Lei, and Kostas Daniilidis. Dynmf: Neural motion factorization for real-time dynamic view synthesis with 3d gaussian splatting. InEuropean Con- ference on Computer Vision, pages 252–269. Springer, 2025. 1, 2, 4

    2025

  9. [17]

    Compact 3d gaussian representation for radiance field

    Joo Chan Lee, Daniel Rho, Xiangyu Sun, Jong Hwan Ko, and Eunbyung Park. Compact 3d gaussian representation for radiance field. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 21719– 21728, 2024. 3, 1

    2024

  10. [18]

    Compact 3d gaussian splat- ting for static and dynamic radiance fields.arXiv preprint arXiv:2408.03822, 2024

    Joo Chan Lee, Daniel Rho, Xiangyu Sun, Jong Hwan Ko, and Eunbyung Park. Compact 3d gaussian splat- ting for static and dynamic radiance fields.arXiv preprint arXiv:2408.03822, 2024. 3, 6, 7, 1

    2024 arXiv

  11. [19]

    Deep contextual video com- pression.Advances in Neural Information Processing Sys- tems, 34:18114–18125, 2021

    Jiahao Li, Bin Li, and Yan Lu. Deep contextual video com- pression.Advances in Neural Information Processing Sys- tems, 34:18114–18125, 2021. 5

    2021

  12. [20]

    Compressing volumetric radiance fields to 1 mb

    Lingzhi Li, Zhen Shen, Zhongshu Wang, Li Shen, and Liefeng Bo. Compressing volumetric radiance fields to 1 mb. pages 4222–4231, 2023. 3

    2023

  13. [21]

    Nerfcodec: Neural feature compression meets neural radiance fields for memory-efficient scene representation

    Sicheng Li, Hao Li, Yiyi Liao, and Lu Yu. Nerfcodec: Neural feature compression meets neural radiance fields for memory-efficient scene representation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 21274–21283, 2024. 3

    2024

  14. [22]

    Neural 3d video synthesis from multi-view video

    Tianye Li, Mira Slavcheva, Michael Zollhoefer, Simon Green, Christoph Lassner, Changil Kim, Tanner Schmidt, Steven Lovegrove, Michael Goesele, Richard Newcombe, et al. Neural 3d video synthesis from multi-view video. In Proceedings of the IEEE/CVF Conference on Computer Vi- si...

    2022

  15. [23]

    Spacetime gaus- sian feature splatting for real-time dynamic view synthesis

    Zhan Li, Zhang Chen, Zhong Li, and Yi Xu. Spacetime gaus- sian feature splatting for real-time dynamic view synthesis. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 8508–8520, 2024. 1, 2, 6, 7, 3

    2024

  16. [24]

    Dvc: An end-to-end deep video com- pression framework

    Guo Lu, Wanli Ouyang, Dong Xu, Xiaoyun Zhang, Chunlei Cai, and Zhiyong Gao. Dvc: An end-to-end deep video com- pression framework. InProceedings of the IEEE/CVF con- ference on computer vision and pattern recognition, pages 11006–11015, 2019. 5

    2019

  17. [25]

    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. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 20654–20664, 2024. 2, 5

    2024

  18. [26]

    Dynamic 3d gaussians: Tracking by per- sistent dynamic view synthesis

    Jonathon Luiten, Georgios Kopanas, Bastian Leibe, and Deva Ramanan. Dynamic 3d gaussians: Tracking by per- sistent dynamic view synthesis. In2024 International Con- ference on 3D Vision (3DV), pages 800–809. IEEE, 2024. 7

    2024

  19. [27]

    Vct: A video compression transformer.arXiv preprint arXiv:2206.07307, 2022

    Fabian Mentzer, George Toderici, David Minnen, Sung-Jin Hwang, Sergi Caelles, Mario Lucic, and Eirikur Agusts- son. Vct: A video compression transformer.arXiv preprint arXiv:2206.07307, 2022. 5

    2022 arXiv

  20. [28]

    Nerf: Representing scenes as neural radiance fields for view syn- thesis

    Ben Mildenhall, Pratul P Srinivasan, Matthew Tancik, Jonathan T Barron, Ravi Ramamoorthi, and Ren Ng. Nerf: Representing scenes as neural radiance fields for view syn- thesis. 2020. 2

    2020

  21. [29]

    Compact 3d scene representation via self- organizing gaussian grids.arXiv preprint arXiv:2312.13299,

    Wieland Morgenstern, Florian Barthel, Anna Hilsmann, and Peter Eisert. Compact 3d scene representation via self- organizing gaussian grids.arXiv preprint arXiv:2312.13299,

    arXiv

  22. [30]

    Compact3d: Com- pressing gaussian splat radiance field models with vector quantization.arXiv preprint arXiv:2311.18159, 2023

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

    2023 arXiv

  23. [31]

    Compressed 3d gaussian splatting for accelerated novel view synthesis

    Simon Niedermayr, Josef Stumpfegger, and R ¨udiger West- ermann. Compressed 3d gaussian splatting for accelerated novel view synthesis. InProceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 10349–10358, 2024. 3, 1

    2024

  24. [32]

    D-nerf: Neural radiance fields for dynamic scenes

    Albert Pumarola, Enric Corona, Gerard Pons-Moll, and Francesc Moreno-Noguer. D-nerf: Neural radiance fields for dynamic scenes. InProceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 10318–10327, 2021. 2

    2021

  25. [33]

    Masked wavelet representation for compact neural radiance fields

    Daniel Rho, Byeonghyeon Lee, Seungtae Nam, Joo Chan Lee, Jong Hwan Ko, and Eunbyung Park. Masked wavelet representation for compact neural radiance fields. pages 20680–20690, 2023. 3

    2023

  26. [34]

    Binary radiance fields

    Seungjoo Shin and Jaesik Park. Binary radiance fields. 2023. 3

    2023

  27. [35]

    Nerf- player: A streamable dynamic scene representation with de- composed neural radiance fields.IEEE Transactions on Visu- alization and Computer Graphics, 29(5):2732–2742, 2023

    Liangchen Song, Anpei Chen, Zhong Li, Zhang Chen, Lele Chen, Junsong Yuan, Yi Xu, and Andreas Geiger. Nerf- player: A streamable dynamic scene representation with de- composed neural radiance fields.IEEE Transactions on Visu- alization and Computer Graphics, 29(5):2732–2742, 2...

    2023

  28. [36]

    Overview of the high efficiency video coding (hevc) standard.IEEE Transactions on circuits and systems for video technology, 22(12):1649–1668, 2012

    Gary J Sullivan, Jens-Rainer Ohm, Woo-Jin Han, and Thomas Wiegand. Overview of the high efficiency video coding (hevc) standard.IEEE Transactions on circuits and systems for video technology, 22(12):1649–1668, 2012. 5

    2012

  29. [37]

    Variable bitrate neural fields

    Towaki Takikawa, Alex Evans, Jonathan Tremblay, Thomas M¨uller, Morgan McGuire, Alec Jacobson, and Sanja Fidler. Variable bitrate neural fields. pages 1–9, 2022. 3

    2022

  30. [38]

    Neural residual radiance fields for streamably free-viewpoint videos

    Liao Wang, Qiang Hu, Qihan He, Ziyu Wang, Jingyi Yu, Tinne Tuytelaars, Lan Xu, and Minye Wu. Neural residual radiance fields for streamably free-viewpoint videos. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 76–87, 2023. 3

    2023

  31. [39]

    Videorf: Ren- dering dynamic radiance fields as 2d feature video streams

    Liao Wang, Kaixin Yao, Chengcheng Guo, Zhirui Zhang, Qiang Hu, Jingyi Yu, Lan Xu, and Minye Wu. Videorf: Ren- dering dynamic radiance fields as 2d feature video streams. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 470–481, 2024. 3

    2024

  32. [40]

    Vˆ 3: View- ing volumetric videos on mobiles via streamable 2d dynamic gaussians.arXiv preprint arXiv:2409.13648, 2024

    Penghao Wang, Zhirui Zhang, Liao Wang, Kaixin Yao, Siyuan Xie, Jingyi Yu, Minye Wu, and Lan Xu. Vˆ 3: View- ing volumetric videos on mobiles via streamable 2d dynamic gaussians.arXiv preprint arXiv:2409.13648, 2024. 3

    2024 arXiv

  33. [41]

    4d gaussian splatting for real-time dynamic scene rendering

    Guanjun Wu, Taoran Yi, Jiemin Fang, Lingxi Xie, Xiaopeng Zhang, Wei Wei, Wenyu Liu, Qi Tian, and Xinggang Wang. 4d gaussian splatting for real-time dynamic scene rendering. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 20310–20320, ...

    2024

  34. [42]

    Real-time photorealistic dynamic scene representation and rendering with 4d gaussian splatting.arXiv preprint arXiv:2310.10642, 2023

    Zeyu Yang, Hongye Yang, Zijie Pan, and Li Zhang. Real-time photorealistic dynamic scene representation and rendering with 4d gaussian splatting.arXiv preprint arXiv:2310.10642, 2023. 1, 2, 6, 7

    2023 arXiv

  35. [43]

    Deformable 3d gaussians for high- fidelity monocular dynamic scene reconstruction

    Ziyi Yang, Xinyu Gao, Wen Zhou, Shaohui Jiao, Yuqing Zhang, and Xiaogang Jin. Deformable 3d gaussians for high- fidelity monocular dynamic scene reconstruction. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 20331–20341, 2024. 1, 2, 4

    2024

  36. [44]

    Efros, Eli Shecht- man, and Oliver Wang

    Richard Zhang, Phillip Isola, Alexei A. Efros, Eli Shecht- man, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. 2018. 6

    2018

  37. [45]

    Mega: Memory-efficient 4d gaussian splatting for dynamic scenes.arXiv preprint arXiv:2410.13613, 2024

    Xinjie Zhang, Zhening Liu, Yifan Zhang, Xingtong Ge, Dailan He, Tongda Xu, Yan Wang, Zehong Lin, Shuicheng Yan, and Jun Zhang. Mega: Memory-efficient 4d gaussian splatting for dynamic scenes.arXiv preprint arXiv:2410.13613, 2024. 3 GIFStream: 4D Gaussian-based Immersive Video ...

    arXiv

  38. [46]

    Overview In this supplementary material, we present three sections: (I) additional details on the methods, (II) supplementary in- formation on the experiments, and (III) a discussion of lim- itations and future work

  39. [47]

    Supplementary for Methods Details on 3D-to-2D Sorting:We adopt the sorting strat- egy introduced in PLAS [29] to reorganize the Gaussian primitives from 3D space to 2D space based on the simi- larity of their attributes. While [29] sorts static 3D Gaus- sian primitives using t...

  40. [48]

    For the 4DGS, we only keep zero degrees of SH coefficients, therefore the storage will be much smaller than the original version

    Supplementary for Experiments Additional Implementation Details:We implement 4DGaussian, 4DGS, STG and CSTG [18, 23, 41, 42] us- ing their official code bases. For the 4DGS, we only keep zero degrees of SH coefficients, therefore the storage will be much smaller than the origi...

  41. [49]

    We present complete frames from the MPEG dataset along with a comparison of local details in this figure

    Limitation Our representation may exhibit inconsistencies in the back- ground area between different GOPs, particularly in the dis- tant background, where there are insufficient points for ini- GT 4DGaussian CSTG Ours Figure 9.More Qualitative Evaluation. We present complete f...

Pith tools

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