Pith. sign in

REVIEW 6 major objections 5 minor 59 references

MSNeRV: Neural Video Representation with Multi-Scale Feature Fusion

T0 review · 6 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 MSNeRV outperforms the VTM random-access codec on dynamic 1080p video under both PSNR and MS-SSIM.

desk verdict Solid incremental INR codec with believable representation gains, but the VTM-beating headline is under-evidenced and needs code or per-sequence data. read the letter →

arxiv 2506.15276 v1 pith:W57PU5OM submitted 2025-06-18 cs.CV cs.MMeess.IV

Jun Zhu , Xinfeng Zhang , Lv Tang , JunHao Jiang This is my paper · ORCID
classification cs.CVcs.MMeess.IV
keywords implicitneuralrepresentationvideocompressionmulti-scalefeaturefusiontemporalwindowGoP-levelgridmulti-resolutionsupervisionhigh-frequencyboostingrate-distortionoptimization
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 argues that implicit neural representations can be made competitive with, and in dynamic scenes better than, a state-of-the-art conventional video codec, if the network is designed around the multi-scale structure of video. It introduces MSNeRV, an INR codec whose encoder fuses frame-level, window-level, and GoP-level grids for temporal consistency, and whose decoder supervises intermediate resolutions and boosts high-frequency residuals. The reported result is a 4.5% BD-rate saving under PSNR and 30.5% under MS-SSIM over VTM-23.7 Random Access on HEVC ClassB, together with 42% and 28% bitrate savings versus HiNeRV on ClassB and UVG. The reason to care: if this holds, a parameter-based codec that needs no training prior can edge past a mature block-based standard on content where motion and detail are hardest.

What carries the argument

The central object is the multi-scale fused grid: a per-frame learnable base grid $\gamma_{\mathrm{base}}(t)$, a sliding temporal-window fusion producing $X_{\mathrm{temporal}}$, and a GoP-level background grid $\gamma_{\mathrm{GoP}}(k)$, summed into $X_{\mathrm{fused}}$. This feeds a decoder built from Multi-Scale Feature blocks, each combining bilinear and pixel-shuffle upsampling with a hierarchical local-grid encoding, fusing depth-wise convolutions of different kernel sizes, and concatenating features across depths. The machinery works by making the network reuse low- and mid-resolution features through multi-resolution supervision, high-frequency boosting, and a scale-adaptive loss, so that representation capacity rises without increasing model size.

What would settle it

Rerun the comparison with per-sequence BD-rate tables and confidence intervals across the five ClassB sequences under at least two distinct VTM-23.7 random-access configurations; if the average PSNR BD-rate advantage over VTM fails to reproduce, or inverts on any single sequence, the central claim is not robust.

Watch

Extended reading notes

Core claim

MSNeRV claims to be an INR-based codec that, on average, surpasses VTM-23.7 with Random Access configuration on HEVC ClassB, a dataset dominated by dynamic and detailed 1080p content. The central discovery is that intermediate features of the upsampling decoder and multi-scale temporal context are under-used resources: capturing them yields better rate-distortion performance at the same model size. The paper's mechanism combines temporal windows with learnable weights, GoP-level background grids, multi-resolution supervision, high-frequency boosting, and multi-scale feature blocks with cross-depth fusion. In video representation experiments, MSNeRV reports higher PSNR than HiNeRV and HNeRV-Boost at comparable model sizes, and in compression experiments it reports the BD-rate savings summarized above.

Load-bearing premise

The headline rests on averaging BD-rate over five HEVC ClassB sequences under one fixed VTM-23.7 random-access configuration and one fixed training and QAT schedule; if that protocol is not representative, the reported 4.5% PSNR advantage could invert.

Editorial extensions

If this is right

  • If the reported results hold, INR-based codecs can surpass a mature block-based codec on dynamic 1080p content under PSNR, not only under perceptual metrics.
  • Because the gains come from fusing existing internal features rather than enlarging the network, the approach points to a way of increasing representation capacity without increasing bitrate.
  • Multi-resolution supervision and high-frequency boosting imply that INR training can use intermediate decoder features as explicit targets, which may improve training efficiency and stability.
  • On UVG, MSNeRV is within 7.8% of VTM under PSNR and better than VTM by 26.2% under MS-SSIM, suggesting the codec is competitive across content types while being perceptually strong.
  • Compared with HiNeRV under the same coding scheme, MSNeRV saves 42% bitrate on ClassB and 28% on UVG, a direct corollary of its representation experiments.

Reading between the lines

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

  • Editorial inference: the temporal-window and GoP-grid fusion mechanism is not specific to video regression; the same idea could be dropped into any coordinate-based representation of dynamic scenes to improve temporal consistency.
  • Editorial inference: because the paper reports bitrate savings only under one VTM-23.7 random-access configuration, the 4.5% figure should be read as the result of that protocol; whether it transfers to other VTM configurations or content classes is untested.
  • Editorial inference: the 30-epoch QAT schedule and channel-based bitrate control suggest two cheap experiments the paper does not run—longer QAT or a stronger entropy model—that could either widen or shrink the reported gap.
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

6 major / 5 minor

Summary. The paper proposes MSNeRV, an implicit neural representation (INR) framework for video compression. The method introduces a multi-scale temporal encoder that fuses base grids with sliding-window temporal features and GoP-level background grids, a multi-scale spatial decoder with multi-resolution supervision and a high-frequency boosting target, a scale-adaptive loss, and multi-scale feature blocks with cross-depth fusion. Experiments on HEVC ClassB and UVG datasets report PSNR gains over existing INR methods (e.g., 0.9–1.4 dB over HiNeRV at comparable model size) and claim BD-rate savings of 4.5% in PSNR and 30.5% in MS-SSIM against VTM-23.7 (Random Access) on ClassB, plus 42% and 28% bitrate savings versus HiNeRV on ClassB and UVG, respectively. The central claim is that MSNeRV outperforms the state-of-the-art conventional codec VTM on dynamic 1080p content.

Significance. If the reported results are reproducible, MSNeRV would be among the first INR-based video codecs to beat VTM random-access compression on 1080p dynamic content under PSNR, which is a notable advance for the field. The per-sequence representation results in Table 1 and the systematic ablations in Tables 3, 5, and 6 are useful empirical contributions, and the architecture components are clearly motivated. However, the headline VTM comparison rests on aggregate BD-rate numbers over only five ClassB sequences, with no per-sequence breakdown, no confidence intervals, and an incompletely specified codec configuration; the UVG PSNR result is 7.8% worse than VTM, which contradicts the unqualified 'superior on average' claim. The absence of released code and the selection of hyperparameters on the same evaluation set further reduce confidence. These issues make the central empirical claim conditional pending a more rigorous evaluation.

major comments (6)
  1. [§4.2, Table 2] The central claim that MSNeRV outperforms VTM-23.7 (RA) is based solely on aggregate BDBR values over five HEVC ClassB sequences, with no per-sequence BDBR, no confidence intervals, and no description of the VTM configuration (QP ladder, intra period, bit depth, encoder options). The PSNR BDBR gain is only -4.5% on ClassB while the UVG PSNR row is +7.8% (worse), so the observed advantage is small relative to typical cross-sequence dispersion. Please provide per-sequence BDBR tables, specify the exact VTM settings used, report error bars or the per-sequence range, and explicitly scope the claim to the ClassB PSNR comparison.
  2. [§4.2, Figure 6] The rate-distortion comparison does not state how many RD points were used for each codec, which QPs were selected for VTM and DCVC, or how the bitrates of INR-based methods were matched to those points. BDBR is sensitive to the placement and number of RD points, so without this information the -4.5% value is not reproducible. Please tabulate the RD points and describe the bitrate-control protocol for each method.
  3. [§4.1, §4.2, Appendix C] Several design choices—temporal window size (Appendix C.2), layer depths (Appendix C.3), SA-loss coefficients (Table 4), and downsampling type (Table 5)—are tuned on the same HEVC ClassB sequences used for the headline VTM comparison. This selection on the test set can bias the measured BDBR gain. Please either validate on a held-out set or report a sensitivity analysis showing how the ClassB BDBR changes when these hyperparameters are varied within reasonable ranges.
  4. [§3.2.2, Eq. (5)] The high-frequency boosting target Vboost = Vgt + H(|Vgt - XN|) is defined using the absolute value of the prediction residual, which is a non-smooth, phase-rectified signal; filtering this quantity does not extract the true high-frequency content in a phase-correct manner, and the target is non-stationary because XN changes during training. The high-pass filter is not specified (filter type, kernel, normalization). Please justify this target, provide the exact HPF implementation, and show that the 0.10 dB gain in the V3 ablation is robust across bitrates and training seeds; if it is not, consider removing the component or replacing it with a more conventional high-frequency loss.
  5. [§4.3, Table 3] The ablation study reports a single operating point per variant, and several differences are 0.05–0.10 dB PSNR, which is within typical run-to-run variation for neural representation training. Without multiple seeds or confidence intervals, the claim that each component contributes meaningfully is not supported. Please add variance estimates or demonstrate deterministic training with a fixed seed.
  6. [Abstract, §1, §4.2] The claim of 'superior compression performance to VTM on average' is overbroad: the UVG PSNR row in Table 2 is +7.8% worse, and the average over ClassB and UVG is not defined. The abstract's statement that MSNeRV 'surpasses VTM-23.7 (Random Access) in dynamic scenarios' should be scoped explicitly to the HEVC ClassB dataset and to the PSNR metric; otherwise the claim is misleading.
minor comments (5)
  1. [§3.1, Eq. (2), Eq. (4)] The notation in the temporal encoder is inconsistent: the summation in Eq. (2) is written as 'i=t+l−1X i=t', and Eq. (4) introduces 'γ base.(t)' while Eq. (2) uses 'X i base'; please unify the notation and define all variables (e.g., w_t^i, γ_GoP(k)) clearly. Also, 'By apply a sliding window mechanism' should be 'By applying a sliding window mechanism'.
  2. [§3.2.1] The subscript in '3x3 convolution layer Nr' should be N_r, not 'Nr'. Additionally, the sentence 'Since this metric primarily focuses on the perceptual similarity that is crucial for the final reconstructed video.' is a sentence fragment; please revise it.
  3. [§4.2] The text says 'MS-SSIM is used to evaluate the subjective quality of reconstructed videos.' MS-SSIM is an objective metric, not a subjective one; please replace 'subjective quality' with 'perceptual quality'.
  4. [References] Reference [11] appears to be an unrelated paper on measuring ocular torsion, and reference [50] describes VTM 17 while the text reports VTM-23.7; please correct the citations to match the VTM version actually used.
  5. [Figure 6] The figure caption for Figure 6 does not state which dataset or bitrate/quality axes are shown, nor whether the RD curves are averaged over sequences or represent a single sequence; please add a descriptive caption.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: all central results are direct experimental measurements against external baselines, with no fitted constant used to predict them.

full rationale

MSNeRV is an empirical systems paper. The central claims—BD-rate savings versus HiNeRV, DCVC-DC, DCVC-FM, and VTM-23.7 (RA)—are measured outputs of the experiments reported in Table 2 and Figure 6, not derived quantities. The SA-loss coefficients in Table 4 and the temporal window size are tuning choices validated by ablations (Tables 3, 5, 6, 7), but they are not fitted to the external BD-rate comparisons and then re-predicted, so there is no self-definitional reduction. The paper does cite prior work by overlapping authors, notably [14] for high-frequency networks and [52] for scene-based deep video compression, but these citations are contextual and not load-bearing: the architecture's components are described with their own equations (Eqs. 1–11), and the compression gains are attributed to measured rate-distortion performance against independently published methods and codecs. No uniqueness theorem, imported ansatz, or renamed empirical pattern is used to force the claimed result. The strongest caveats are experimental robustness issues—only five ClassB sequences, no per-sequence BDBR, no error bars, and no released code—but these are correctness or reproducibility concerns, not circularity. Under the hard rules, a paper whose claims are direct measurements has essentially zero circularity burden, so the score is 0.

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

The central claim rests mainly on empirical training and evaluation rather than on derivation. The ledger lists the hand-tuned coefficients and architecture hyperparameters selected with ClassB ablations as free parameters. The key axioms are the standard INR premise that a per-video network can act as a compact codec, the assumption that max-pooled low-resolution targets are good supervision, and the unusual assumption that the high-frequency-boosted reference in Eq. (5) is a valid training target. No new physical or conceptual entities are introduced.

free parameters (4)
  • SA loss coefficients alpha_r, beta_r = alpha: 0.9, 0.8, 0.7, 0.6; beta: 0.1, 0.2, 0.3, 0.3 across resolutions (Table 4)
    Chosen by hand for each resolution and used in the total loss; no sensitivity analysis is reported.
  • Temporal window size l and overlap = l=5, overlap=4
    Selected via ablation on HEVC ClassB (Table 6); affects all temporal fusion.
  • Fusion layer count and layer depth configuration = 3 fusion layers, depths (3, 3, 3, 3)
    Selected by ablation on HEVC ClassB (Table 7); affects model capacity and bitrate.
  • Downsampling function for multi-resolution supervision = Max pooling
    Selected by comparison on ClassB (Table 5); controls all low-resolution supervision targets.
assumptions (4)
  • domain assumption A per-video neural network overfit by gradient descent can serve as a compact video representation whose bitrate is the cost of its compressed weights.
    This is the standard INR premise inherited from NeRV/HiNeRV (Sections 1 and 2.2).
  • domain assumption Max pooling preserves structure better than smoothing for low-resolution supervision targets.
    Justified only by Table 5 on ClassB, then assumed for the method.
  • ad hoc to paper Vboost = Vgt + H(|Vgt - XN|) is a valid training reference for improving high-frequency fidelity.
    Eq. 5 introduces a target that depends on the current reconstruction error and can exceed valid pixel ranges; no theoretical or empirical justification beyond the ablation.
  • domain assumption PSNR and MS-SSIM BDBR over 12 sequences measure compression superiority.
    Standard metrics, but the headline VTM comparison rests on 5 ClassB sequences for PSNR.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MSNeRV: Neural Video Representation with Multi-Scale Feature Fusion." pith.science (2026). https://pith.science/paper/W57PU5OM

@misc{pith2026250615276,
  author       = {Pith},
  title        = {Pith review of: MSNeRV: Neural Video Representation with Multi-Scale Feature Fusion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/W57PU5OM}},
  note         = {Machine review of arXiv:2506.15276}
}
read the original abstract

Implicit Neural representations (INRs) have emerged as a promising approach for video compression, and have achieved comparable performance to the state-of-the-art codecs such as H.266/VVC. However, existing INR-based methods struggle to effectively represent detail-intensive and fast-changing video content. This limitation mainly stems from the underutilization of internal network features and the absence of video-specific considerations in network design. To address these challenges, we propose a multi-scale feature fusion framework, MSNeRV, for neural video representation. In the encoding stage, we enhance temporal consistency by employing temporal windows, and divide the video into multiple Groups of Pictures (GoPs), where a GoP-level grid is used for background representation. Additionally, we design a multi-scale spatial decoder with a scale-adaptive loss function to integrate multi-resolution and multi-frequency information. To further improve feature extraction, we introduce a multi-scale feature block that fully leverages hidden features. We evaluate MSNeRV on HEVC ClassB and UVG datasets for video representation and compression. Experimental results demonstrate that our model exhibits superior representation capability among INR-based approaches and surpasses VTM-23.7 (Random Access) in dynamic scenarios in terms of compression efficiency.

Figures

Figures reproduced from arXiv: 2506.15276 by the authors.

Figure 1
Figure 1. Reconstruction quality over iterations for implicit neural [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Left top: Main architecture of MSNeRV. Right: High-frequency boosting. The residual between XN and Vgt is processed by the high pass filter (HPF) and selectively added to Vgt. Left bottom: Multi-resolution supervision. Low-resolution videos are generated by cascaded max-pooling. Scale-adaptive loss (SA Loss) functions are used to incorporate multi-scale spatial information into the network. ...... ...... t = 0 1 2 T… view at source ↗
Figure 3
Figure 3. The multi-scale temporal encoder. We use a temporal [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: (a) Video Ground Truth. (b) Optical flow map where dif￾ferent colors indicate varying motion directions and the white areas represent regions with no motion. (c) High-Frequency components extracted by a high-pass filter. (d)(e)(f) Feature maps extracted by the 3 × 3 co…
Figure 6
Figure 6. Figure 6: Video compression results on the HEVC ClassB [ [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: Compression results on different videos. [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: Parameter distributions of MSNeRV (Standard Deviation [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 9
Figure 9. Figure 9: Inpainting results. Ground Truth MSNeRV HiNeRV t = 4 4 t = 4 6 [PITH_FULL_IMAGE:figures/full_fig_p013_9.png]
Figure 10
Figure 10. Figure 10: Interpolation results. We interpolate the grids corresponding to the training set to obtain the required grids for the test set. These grids are then upsampled to generate the test frames using the trained decoder. This approach ensures that no information from the te…
Figure 11
Figure 11. Figure 11: Interpolation results. 14 [PITH_FULL_IMAGE:figures/full_fig_p014_11.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 53 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,”Communications of the ACM, vol. 65, no. 1, pp. 99–106, 2021. 3

    work page 2021

  2. [2]

    Nerv: Neural representations for videos,

    H. Chen, B. He, H. Wang, Y . Ren, S. N. Lim, and A. Shrivas- tava, “Nerv: Neural representations for videos,”Advances in Neural Information Processing Systems, vol. 34, pp. 21 557– 21 568, 2021. 1, 2, 3

    work page 2021

  3. [3]

    Hnerv: A hybrid neural representation for videos,

    H. Chen, M. Gwilliam, S.-N. Lim, and A. Shrivastava, “Hnerv: A hybrid neural representation for videos,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 10 270–10 279. 1, 3

    work page 2023

  4. [4]

    Hin- erv: Video compression with hierarchical encoding-based neural representation,

    H. M. Kwan, G. Gao, F. Zhang, A. Gower, and D. Bull, “Hin- erv: Video compression with hierarchical encoding-based neural representation,”Advances in Neural Information Pro- cessing Systems, vol. 36, 2024. 1, 2, 3, 5, 6

    work page 2024

  5. [5]

    Dvc: An end-to-end deep video compression framework,

    G. Lu, W. Ouyang, D. Xu, X. Zhang, C. Cai, and Z. Gao, “Dvc: An end-to-end deep video compression framework,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 11 006–11 015. 1, 2

    work page 2019

  6. [6]

    Deep contextual video compression,

    J. Li, B. Li, and Y . Lu, “Deep contextual video compression,” Advances in Neural Information Processing Systems, vol. 34, pp. 18 114–18 125, 2021. 1, 3

    work page 2021

  7. [7]

    Neural video compression with diverse contexts,

    ——, “Neural video compression with diverse contexts,” in Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, 2023, pp. 22 616–22 626. 1, 3, 6

    work page 2023

  8. [8]

    Neural video compression with feature modulation,

    ——, “Neural video compression with feature modulation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 26 099–26 108. 1, 3, 6

    work page 2024

Show all 59 references
  1. [9]

    Nvrc: Neural video representation compression,

    H. M. Kwan, G. Gao, F. Zhang, A. Gower, and D. Bull, “Nvrc: Neural video representation compression,”arXiv preprint arXiv:2409.07414, 2024. 2, 3

    2024

  2. [10]

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

    B. Bross, Y .-K. Wang, Y . Ye, S. Liu, J. Chen, G. J. Sullivan, and J.-R. Ohm, “Overview of the versatile video coding (vvc) standard and its applications,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 31, no. 10, pp. 3736– 3764, 2021. 1, 2, 3

    2021

  3. [11]

    Vtm—an image- processing system for measuring ocular torsion,

    S. Moore, I. Curthoys, and S. McCoy, “Vtm—an image- processing system for measuring ocular torsion,”Computer methods and programs in biomedicine, vol. 35, no. 3, pp. 219–230, 1991. 1

    1991

  4. [12]

    High efficiency video coding,

    M. Wien, “High efficiency video coding,”Coding Tools and specification, vol. 24, p. 1, 2015. 1

    2015

  5. [13]

    Algorithm description for versa- tile video coding and test model 5,

    J. Chen, Y . Ye, and S. Kim, “Algorithm description for versa- tile video coding and test model 5,”JVET-L1002, Joint Video Exploration Team (JVET), 2018

    2018

  6. [14]

    Canerv: Content adaptive neural representation for video compression,

    L. Tang, J. Zhu, X. Zhang, L. Zhang, S. Ma, and Q. Huang, “Canerv: Content adaptive neural representation for video compression,”arXiv preprint arXiv:2502.06181, 2025. 1, 4

    2025 arXiv

  7. [15]

    Ffnerv: Flow-guided frame-wise neural representations for videos,

    J. C. Lee, D. Rho, J. H. Ko, and E. Park, “Ffnerv: Flow-guided frame-wise neural representations for videos,” inProceedings of the 31st ACM International Conference on Multimedia, 2023, pp. 7859–7870. 2, 3

    2023

  8. [16]

    Overview of the high efficiency video coding (hevc) stan- dard,

    G. J. Sullivan, J.-R. Ohm, W.-J. Han, and T. Wiegand, “Overview of the high efficiency video coding (hevc) stan- dard,”IEEE Transactions on circuits and systems for video technology, vol. 22, no. 12, pp. 1649–1668, 2012. 2, 3, 6, 7

    2012

  9. [17]

    Uvg dataset: 50/120fps 4k sequences for video codec analysis and develop- ment,

    A. Mercat, M. Viitanen, and J. Vanne, “Uvg dataset: 50/120fps 4k sequences for video codec analysis and develop- ment,” inProceedings of the 11th ACM multimedia systems conference, 2020, pp. 297–302. 2, 6, 7

    2020

  10. [18]

    Overview of the h. 264/avc video coding standard,

    T. Wiegand, G. J. Sullivan, G. Bjontegaard, and A. Luthra, “Overview of the h. 264/avc video coding standard,”IEEE Transactions on circuits and systems for video technology, vol. 13, no. 7, pp. 560–576, 2003. 1, 2, 3

    2003

  11. [19]

    Deep learning-based intra pre- diction mode decision for hevc,

    T. Laude and J. Ostermann, “Deep learning-based intra pre- diction mode decision for hevc,” in2016 Picture Coding Symposium (PCS). IEEE, 2016, pp. 1–5. 2

    2016

  12. [20]

    Deep- coder: A deep neural network based video compression,

    T. Chen, H. Liu, Q. Shen, T. Yue, X. Cao, and Z. Ma, “Deep- coder: A deep neural network based video compression,” in 2017 IEEE Visual Communications and Image Processing (VCIP). IEEE, 2017, pp. 1–4

    2017

  13. [21]

    Deep learning in latent space for video prediction and compression,

    B. Liu, Y . Chen, S. Liu, and H.-S. Kim, “Deep learning in latent space for video prediction and compression,” inPro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021, pp. 701–710. 2

    2021

  14. [22]

    Lossy image compression with compressive autoencoders,

    L. Theis, W. Shi, A. Cunningham, and F. Husz ´ar, “Lossy image compression with compressive autoencoders,”arXiv preprint arXiv:1703.00395, 2017

    2017 arXiv

  15. [23]

    Boosting neural representations for videos with a conditional decoder,

    X. Zhang, R. Yang, D. He, X. Ge, T. Xu, Y . Wang, H. Qin, and J. Zhang, “Boosting neural representations for videos with a conditional decoder,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 2556–2566. 3, 6

    2024

  16. [24]

    Implicit neural representations with periodic activa- tion functions,

    V . Sitzmann, J. Martel, A. Bergman, D. Lindell, and G. Wet- zstein, “Implicit neural representations with periodic activa- tion functions,”Advances in neural information processing systems, vol. 33, pp. 7462–7473, 2020. 3

    2020

  17. [25]

    Inras: Implicit neural representation for audio scenes,

    K. Su, M. Chen, and E. Shlizerman, “Inras: Implicit neural representation for audio scenes,”Advances in Neural Infor- mation Processing Systems, vol. 35, pp. 8144–8158, 2022. 3

    2022

  18. [26]

    Coin: Compression with implicit neural represen- tations,

    E. Dupont, A. Goli ´nski, M. Alizadeh, Y . W. Teh, and A. Doucet, “Coin: Compression with implicit neural represen- tations,”arXiv preprint arXiv:2103.03123, 2021. 3

    2021 arXiv

  19. [27]

    Implicit neural representations for image compression,

    Y . Str¨umpler, J. Postels, R. Yang, L. V . Gool, and F. Tombari, “Implicit neural representations for image compression,” in European Conference on Computer Vision. Springer, 2022, pp. 74–91

    2022

  20. [28]

    C3: High-performance and low-complexity neural compres- sion from a single image or video,

    H. Kim, M. Bauer, L. Theis, J. R. Schwarz, and E. Dupont, “C3: High-performance and low-complexity neural compres- sion from a single image or video,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 9347–9358. 3

    2024

  21. [29]

    D-nerf: Neural radiance fields for dynamic scenes,

    A. Pumarola, E. Corona, G. Pons-Moll, and F. Moreno- Noguer, “D-nerf: Neural radiance fields for dynamic scenes,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021, pp. 10 318–10 327. 3

    2021

  22. [30]

    Nerfplayer: A streamable dynamic scene represen- 10 tation with decomposed neural radiance fields,

    L. Song, A. Chen, Z. Li, Z. Chen, L. Chen, J. Yuan, Y . Xu, and A. Geiger, “Nerfplayer: A streamable dynamic scene represen- 10 tation with decomposed neural radiance fields,”IEEE Trans- actions on Visualization and Computer Graphics, vol. 29, no. 5, pp. 2732–2742, 2023. 3

    2023

  23. [31]

    Deep learning based hevc in-loop filtering for decoder quality enhancement,

    S. Kuanar, C. Conly, and K. Rao, “Deep learning based hevc in-loop filtering for decoder quality enhancement,” in2018 Picture Coding Symposium (PCS). IEEE, 2018, pp. 164–168. 2

    2018

  24. [32]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 770–

    2016

  25. [33]

    Gradient- based learning applied to document recognition,

    Y . LeCun, L. Bottou, Y . Bengio, and P. Haffner, “Gradient- based learning applied to document recognition,”Proceedings of the IEEE, vol. 86, no. 11, pp. 2278–2324, 1998. 4

    1998

  26. [34]

    J ¨ahne,Digital image processing

    B. J ¨ahne,Digital image processing. Springer Science & Business Media, 2005. 4

    2005

  27. [35]

    Ds- nerv: Implicit neural video representation with decomposed static and dynamic codes,

    H. Yan, Z. Ke, X. Zhou, T. Qiu, X. Shi, and D. Jiang, “Ds- nerv: Implicit neural video representation with decomposed static and dynamic codes,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 23 019–23 029. 4

    2024

  28. [36]

    Metanerv: Meta neural representations for videos with spatial-temporal guidance,

    J. Guo, J. Yao, Z. Wang, J. Bu, H. Wanget al., “Metanerv: Meta neural representations for videos with spatial-temporal guidance,”arXiv preprint arXiv:2501.02427, 2025

    2025 arXiv

  29. [37]

    Loss functions for image restoration with neural networks,

    H. Zhao, O. Gallo, I. Frosio, and J. Kautz, “Loss functions for image restoration with neural networks,”IEEE Transactions on computational imaging, vol. 3, no. 1, pp. 47–57, 2016. 5

    2016

  30. [38]

    On loss functions for deep neural networks in classification,

    K. Janocha and W. M. Czarnecki, “On loss functions for deep neural networks in classification,”arXiv preprint arXiv:1702.05659, 2017. 5

    2017 arXiv

  31. [39]

    Real-time single im- age and video super-resolution using an efficient sub-pixel convolutional neural network,

    W. Shi, J. Caballero, F. Husz ´ar, J. Totz, A. P. Aitken, R. Bishop, D. Rueckert, and Z. Wang, “Real-time single im- age and video super-resolution using an efficient sub-pixel convolutional neural network,” inProceedings of the IEEE conference on computer vision and pattern r...

    2016

  32. [40]

    Channel-wise autoregressive en- tropy models for learned image compression,

    D. Minnen and S. Singh, “Channel-wise autoregressive en- tropy models for learned image compression,” in2020 IEEE International Conference on Image Processing (ICIP). IEEE, 2020, pp. 3339–3343. 2

    2020

  33. [41]

    Recursive residual convolutional neural network-based in-loop filtering for intra frames,

    S. Zhang, Z. Fan, N. Ling, and M. Jiang, “Recursive residual convolutional neural network-based in-loop filtering for intra frames,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 30, no. 7, pp. 1888–1900, 2019. 2

    1900

  34. [42]

    Hypersound: Generating implicit neural rep- resentations of audio signals with hypernetworks,

    F. Szatkowski, K. J. Piczak, P. Spurek, J. Tabor, and T. Trzci´nski, “Hypersound: Generating implicit neural rep- resentations of audio signals with hypernetworks,”arXiv preprint arXiv:2211.01839, 2022. 3

    2022 arXiv

  35. [43]

    The impact of bitrate and gop pattern on the video quality of h. 265/hevc compression standard,

    J. Xu, B. Zhou, C. Zhang, N. Ke, W. Jin, and S. Hao, “The impact of bitrate and gop pattern on the video quality of h. 265/hevc compression standard,” in2018 IEEE International Conference on Signal Processing, Communications and Com- puting (ICSPCC). IEEE, 2018, pp. 1–5

    2018

  36. [44]

    Adam: A method for stochastic optimization,

    D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,”arXiv preprint arXiv:1412.6980, 2014. 6

    2014 arXiv

  37. [45]

    Quantization and training of neural networks for efficient integer-arithmetic-only infer- ence,

    B. Jacob, S. Kligys, B. Chen, M. Zhu, M. Tang, A. Howard, H. Adam, and D. Kalenichenko, “Quantization and training of neural networks for efficient integer-arithmetic-only infer- ence,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. ...

    2018

  38. [46]

    Improvements of the bd-psnr model,

    G. Bjontegaard, “Improvements of the bd-psnr model,” VCEG-AI11, 2008. 6

    2008

  39. [47]

    Snerv: Spectra-preserving neural representation for video,

    J. Kim, J. Lee, and J.-W. Kang, “Snerv: Spectra-preserving neural representation for video,” inEuropean Conference on Computer Vision. Springer, 2024, pp. 332–348. 4

    2024

  40. [48]

    Practical full resolution learned lossless image compression,

    F. Mentzer, E. Agustsson, M. Tschannen, R. Timofte, and L. V . Gool, “Practical full resolution learned lossless image compression,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 10 629– 10 638. 3, 6

    2019

  41. [49]

    Ps-nerv: Patch- wise stylized neural representations for videos,

    Y . Bai, C. Dong, C. Wang, and C. Yuan, “Ps-nerv: Patch- wise stylized neural representations for videos,” in2023 IEEE International Conference on Image Processing (ICIP). IEEE, 2023, pp. 41–45

    2023

  42. [50]

    Algorithm description for versatile video coding and test model 17 (vtm 17),

    A. Browne, Y . Ye, and S. Kim, “Algorithm description for versatile video coding and test model 17 (vtm 17),” inJVET- Z2002, 2022. 6

    2022

  43. [51]

    E-nerv: Ex- pedite neural video representation with disentangled spatial- temporal context,

    Z. Li, M. Wang, H. Pi, K. Xu, J. Mei, and Y . Liu, “E-nerv: Ex- pedite neural video representation with disentangled spatial- temporal context,” inEuropean Conference on Computer Vision. Springer, 2022, pp. 267–284. 2, 3

    2022

  44. [52]

    Scene matters: Model-based deep video compression,

    L. Tang, X. Zhang, G. Zhang, and X. Ma, “Scene matters: Model-based deep video compression,” inProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), October 2023, pp. 12 481–12 491. 3

    2023

  45. [53]

    Nerp: implicit neural repre- sentation learning with prior embedding for sparsely sampled image reconstruction,

    L. Shen, J. Pauly, and L. Xing, “Nerp: implicit neural repre- sentation learning with prior embedding for sparsely sampled image reconstruction,”IEEE Transactions on Neural Net- works and Learning Systems, vol. 35, no. 1, pp. 770–782,

  46. [54]

    Fast multi-type tree partitioning for versatile video coding using a lightweight neural network,

    S.-h. Park and J.-W. Kang, “Fast multi-type tree partitioning for versatile video coding using a lightweight neural network,” IEEE Transactions on Multimedia, vol. 23, pp. 4388–4399,

  47. [55]

    Block partitioning structure in the vvc standard,

    Y .-W. Huang, J. An, H. Huang, X. Li, S.-T. Hsiang, K. Zhang, H. Gao, J. Ma, and O. Chubach, “Block partitioning structure in the vvc standard,”IEEE Transactions on Circuits and Sys- tems for Video Technology, vol. 31, no. 10, pp. 3818–3833,

  48. [56]

    Low-complexity ctu partition structure decision and fast intra mode decision for versatile video coding,

    H. Yang, L. Shen, X. Dong, Q. Ding, P. An, and G. Jiang, “Low-complexity ctu partition structure decision and fast intra mode decision for versatile video coding,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 30, no. 6, pp. 1668–1682, 2019

    2019

  49. [57]

    Fast intra mode deci- sion algorithm for versatile video coding,

    X. Dong, L. Shen, M. Yu, and H. Yang, “Fast intra mode deci- sion algorithm for versatile video coding,”IEEE Transactions on Multimedia, vol. 24, pp. 400–414, 2021. 2

    2021

  50. [58]

    Wide angular intra prediction for versatile video coding,

    L. Zhao, X. Zhao, S. Liu, X. Li, J. Lainema, G. Rath, F. Urban, and F. Racap´e, “Wide angular intra prediction for versatile video coding,” in2019 Data Compression Conference (DCC). IEEE, 2019, pp. 53–62. 2

    2019

  51. [59]

    Deep learning-based chroma prediction for intra versatile video coding,

    L. Zhu, Y . Zhang, S. Wang, S. Kwong, X. Jin, and Y . Qiao, “Deep learning-based chroma prediction for intra versatile video coding,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 31, no. 8, pp. 3168–3181, 2020. 2 11 Resolution α β 240×135(r=1) 0.9 0.1 48...

    2020

Pith tools

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