Pith. sign in

REVIEW 4 major objections 6 minor 56 references

F-Hash: Feature-Based Hash Design for Time-Varying Volume Visualization via Multi-Resolution Tesseract Encoding

T0 review · 4 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read F-Hash is a feature-based multi-resolution tesseract hash encoding that claims to cut INR training time for time-varying volumes by 10–100x while using fewer parameters.

desk verdict A genuinely new 4D Tesseract hash encoding for time-varying volumes, but the headline 10x-100x speedup is measured against baselines trained on full volumes while F-Hash trains on a feature coreset, so the encoding's contribution is not isolated. read the letter →

arxiv 2507.03836 v1 pith:KK66FNWV submitted 2025-07-04 cs.GR cs.CV

Jianxin Sun , David Lenz , Hongfeng Yu , Tom Peterka This is my paper
classification cs.GRcs.CV
keywords time-varyingvolumevisualizationimplicitneuralrepresentationmulti-resolutionhashencodingtesseractembeddinggridminimalperfectfunctioncoresetselectionadaptiveraymarchinginput
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

F-Hash is an input-encoding method for implicit neural representations of time-varying volumetric data. It places trainable embeddings on a multi-resolution 4D "tesseract" grid spanning space and time, maps every grid corner to a hash bucket with a collision-free bijective function, and trains only on a feature-based coreset selected from key frames. The paper argues that this combination cuts INR training time by one to two orders of magnitude while keeping reconstruction quality, and that the same encoding serves interval, isosurface, and segmentation features as well as evolution visualization. The practical stake is that a representation previously requiring hours or days of training can instead be fit in minutes and rendered interactively.

What carries the argument

The load-bearing object is the multi-resolution Tesseract embedding grid: for each resolution level, a 4D grid over (t,x,y,z) whose corners hold trainable embedding vectors, aggregated by quadrilinear interpolation (spatial trilinear first, then linear in time) into one vector per level. Each grid is indexed by F-Hash, a bijective minimal perfect hash that linearizes the 4D coordinate into a single bucket index, so every corner maps to exactly one occupied bucket and each table size scales with the level's grid size. A feature-based coreset-selection pipeline determines where and when to train: key-frame detection, feature extraction, dilation to cover interpolation stencils, temporal fusion into a feature bounding box, and coordinate normalization into that box. The concatenated per-level embedding vectors feed a shallow MLP that outputs the scalar volume value.

What would settle it

Use a synthetic time-varying volume with a feature moving along a fast circular or accelerating path, keep only a few widely spaced key frames, train F-Hash, and render an intermediate time; if the PSNR at that interpolated time falls far below the PSNR of the same model trained on all frames, the coreset-plus-linear-interpolation premise is falsified.

Watch

Extended reading notes

Core claim

The paper's central claim is that the input encoding, not the neural network, is the bottleneck when training implicit neural representations on time-varying volumes, and that a feature-adaptive 4D encoding removes that bottleneck. F-Hash constructs per-resolution-level Tesseract embedding grids whose spatial and temporal resolutions are derived from a feature bounding box around detected features at key frames, and indexes each grid with a minimal perfect hash function built from simple 4D linearization. The hash is bijective and gives 100% bucket utilization, so no collisions waste learning capacity and no empty buckets waste parameters. Because only dilated feature regions at key frames are used as training samples, every epoch is much cheaper, and because the grid is quadrilinearly interpolated over time, the trained network still answers queries between key frames. Across the Combustion, Argon Bubble, and Supernova datasets, F-Hash reaches 20/30/40 dB PSNR thresholds faster than dense-grid and multi-resolution-hash baselines, ends the fixed iteration budget with higher PSNR/SSIM and fewer encoding parameters, and its adaptive ray marching lowers rendering latency; the paper also states that its compression ratio, while better than the input-encoding baselines, is below specialized volume compressors.

Load-bearing premise

The method assumes that the dilated feature regions around a handful of detected key frames, stitched together by linear interpolation in time, contain enough information to reconstruct the whole time-varying volume at every instant, including between key frames.

Editorial extensions

If this is right

  • Training an INR for a large time-varying volume drops from hours or days to minutes: the three test datasets fit in 7.4, 8.5, and 15.7 minutes respectively under F-Hash, versus roughly 43–123 minutes for the compared encodings.
  • The same trained representation supports feature tracking, evolution visualization, and temporal super-resolution because any query at time t is defined by interpolation between neighboring key frames.
  • Encoding parameter counts fall by roughly an order of magnitude, shrinking model memory and speeding up every backpropagation step.
  • Rendering can stay interactive: occupancy grids skip empty space and the adaptive ray marching algorithm cuts average frame latency by roughly a third relative to sample streaming alone.
  • Because the encoding is agnostic to how features are detected, interval, isosurface, and segmentation features can share one unified encoding and rendering pipeline.

Reading between the lines

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

  • The speedup is not purely architectural: F-Hash trains only on dilated feature regions at key frames while the baselines train on full frames, so part of the 10–100x gain likely comes from the smaller and easier training set; a controlled comparison with baselines restricted to the same coreset would isolate the encoding's contribution.
  • The temporal interpolation is linear between the two nearest key frames, so features that accelerate, rotate, or change topology between key frames will be smeared; datasets with such nonlinear evolution would need more key frames or higher-order temporal interpolation.
  • The collision-free hash and resolution-dependent tables are general, so the same design could be applied to video, climate, or other spatiotemporal arrays, and a data-adaptive fold parameter could trim parameters further.
  • Given the paper's admitted compression gap versus specialized lossy compressors, F-Hash's natural role is a fast interactive representation for analysis and feature visualization rather than a long-term archival codec.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The manuscript proposes F-Hash, a multi-resolution hash-based input encoding for implicit neural representations (INRs) of time-varying volumetric data. The method selects a feature-based coreset from key frames, constructs 4D multi-resolution "Tesseract" embedding grids, maps grid corners to dense per-level hash tables via the bijective linearization in Eq. (9), and feeds quadrilinearly interpolated embeddings to a small MLP. A rendering pipeline with an adaptive ray marching (ARM) algorithm is also proposed. Experiments on Combustion, Argon Bubble, and Supernova report 10–100× faster convergence, fewer encoding parameters, and lower rendering latency than MHE, MHE Tinycudann, DG Single-Res, and DG Multi-Res.

Significance. If the convergence claims were supported by controlled experiments, the work would be practically significant: it offers a simple, collision-free dense hash construction, a natural temporal extension of hash-based encoding, and a coreset/occupancy-grid strategy that can reduce both training and rendering workload. The idea of deriving the multi-resolution configuration from a feature bounding box is a useful departure from manually tuned fixed-resolution grids. The central derivation in Eq. (9) is sound, and the paper is honest about several limitations, including the residual gap to specialized compressors. However, the main empirical claim is not yet established because the comparison simultaneously changes the training data (coreset), the temporal modeling strategy (one 4D model versus per-key-frame models), and the encoding architecture. The rendering-latency results are likewise plausible but confounded by occupancy-grid culling and smaller model size.

major comments (4)
  1. [Sec. 5.2.1, Eq. (3), Fig. 16 caption] The convergence comparison does not isolate the contribution of the Tesseract encoding. F-Hash is trained only on the feature coreset inside the Feature Bounding Box at 9–10 key frames, while the baseline methods are trained separately on full key-frame volumes, as stated in Sec. 5.2.1. The FBB covers only 63.95%, 32.30%, and 52.41% of the Combustion, Argon Bubble, and Supernova frame volumes (Fig. 16 caption), and the coreset is a further subset of the FBB (Eq. 3). Consequently, the reported 10–100× speedup couples data-set reduction, one 4D model versus per-frame 3D models, and the encoding design. I request an ablation in which the baselines are trained on the same coreset and the same key-frame schedule, and a version of F-Hash trained on the full volume or on all key frames, so that the speedup attributable to the encoding itself can be separated from data-selection and joint-modeling effects.
  2. [Sec. 3.1, Sec. 6, Introduction] The paper claims a meta-learning component, but none is implemented or evaluated. Section 3.1 says the coreset selection is "for Meta-learning," and the Introduction lists it as a component, yet there is no pretraining, warm-start, or meta-learning procedure anywhere in the method or experiments. Section 5.2.1 itself attributes the speedup to "coreset selection reduces the total number of training samples," which is ordinary data subsampling, not meta-learning. Please remove the meta-learning terminology or add and evaluate an actual meta-learning stage.
  3. [Sec. 3.3, Eq. (12), Sec. 5.2.1] The claimed temporal super-resolution ability is not evaluated. The paper states that the INR trained on key frames can be queried at arbitrary time steps via temporal interpolation (Eq. 12), but the convergence and accuracy metrics in Sec. 5.2.1 and Table 4 are computed on key frames. No PSNR/SSIM or visual comparison is reported for interpolated non-key frames. If features move or change nonlinearly between key frames, the reconstruction error at intermediate times may grow, which would qualify the practical value of the speedup. Please report accuracy at unobserved time steps, ideally as a function of temporal distance from the nearest key frame, and compare with per-frame models that train on all frames.
  4. [Tables 3 and 4, Sec. 5.1.3] The training schedule and comparison protocol are described inconsistently. Sec. 5.1.3 says training stops at the 60th epoch/iteration, while Sec. 5.3.1 compares methods at the 30th iteration, and Table 3 reports a 60-iteration budget with "NR" for methods that never reach the target PSNR. An "NR" result gives no information about how close the method came, and a fixed iteration budget combined with different per-iteration costs can bias wall-clock comparisons. Please specify which schedule each table uses, report time-to-target at common accuracy levels, and include multiple seeds with error bars, since the reported convergence times are single runs.
minor comments (6)
  1. [Introduction] The text promises that "Detailed supporting quantitative results can be found in the Appendix," but no appendix appears in the submitted manuscript. Either include the appendix or remove the reference.
  2. [Sec. 3.2.1, Eqs. (7)–(8)] The resolution recursion is defined for l in {2, …, Ls−1}, but the Tail function mentions the stopping condition l ≤ Ls, leaving the resolution of the top level (l = Ls) ambiguous. Please clarify the intended construction.
  3. [Sec. 3.2.3, Eq. (9)] Calling Eq. (9) a "feature-based hash function" is misleading: it is a deterministic coordinate linearization independent of the detected features, and the collision-free property is guaranteed by construction. The feature-dependent part of the method lies in the coreset selection and the FBB-derived grid configuration, not in the hash itself. Consider renaming it a "dense linearization" or "bijective hash" for clarity.
  4. [Sec. 3.1] The key-frame detection methods [17,37,49] and the feature extraction thresholds/dilation radius are not specified in sufficient detail for reproduction. Since the number and placement of key frames are free parameters that affect the coreset and thus the speedup, please report the concrete choices used for each dataset.
  5. [Tables 3–6] The tables are difficult to read because of the dense multi-column layout, especially Table 3, where it is hard to see which convergence times correspond to which feature type. Please reformat, for example by separating the feature columns or using clearer headers.
  6. [Sec. 5.3.2] The compression comparison is a useful strength, but the comparison with NeurComp, SZ3, and TTHRESH uses a different quality target (≥45 dB frame PSNR). Please state the bitrate/quality operating points explicitly so the reader can see that the comparison is apples-to-apples.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the hash properties are true by construction (Eq. 9), and the speedup claims are empirical comparisons against external baselines, not derivations from fitted constants.

full rationale

The paper's only 'by construction' claim is the bijective/collision-free hash (Eq. 9), which is a direct linearization of 4D grid coordinates; this is a design definition, not a circular derivation of a prediction from its own output. The convergence-speed claims are measured wall-clock/PSNR curves against external baselines (MHE, DG Single-Res, DG Multi-Res, Tinycudann), with no parameter fitted to the convergence outcome. The only self-citation ([33]) appears in a related-work list for rendering and is not load-bearing. The skeptic's coreset critique—F-Hash trains on a feature-bounded subset of key frames while baselines model full frames separately—is a valid fairness/soundness concern about attributing the speedup to the encoding, but it is not circularity: the paper does not define speed in terms of coreset size or rename a fitted quantity as a prediction. The missing meta-learning pretraining is an overclaim/missing-support issue, not a circular reduction. Therefore no circular step meets the quoted-evidence bar.

Assumptions & free parameters 5 free parameters · 5 assumptions · 1 invented entities

The central claim depends on several hand-chosen parameters (fold, embedding size, MLP size, key-frame selection, feature thresholds) and on domain assumptions about the sufficiency of key frames, linear temporal interpolation, and the coreset. No new physical entities are introduced; the tesseract grid is an internal computational structure.

free parameters (5)
  • fold parameter f = 2
    Controls the resolution reduction factor between multi-resolution levels for all axes (Section 3.2.1). Chosen by hand, fixed for all experiments, and acknowledged as predefined in Section 6.
  • embedding size = 2
    Number of trainable embedding weights per grid corner; selected via ablation on the Argon Bubble dataset in Section 4.1.
  • MLP depth and width = 2 layers, 64 neurons
    Chosen via ablation in Section 4.2 and fixed across all datasets; these define the decoder capacity.
  • key frame count and detection method = 10, 9, and 10 key frames for Combustion, Argon Bubble, Supernova
    The paper cites generic key-frame methods [17,37,49] without naming the algorithm or parameters used, so the exact coreset depends on an unspecified choice.
  • feature thresholds and dilation radius = Not reported
    Interval, isosurface, and segmentation features in Section 5.2 require value ranges and neighborhood dilation parameters that are not specified, and these determine the training coreset.
assumptions (5)
  • domain assumption Key frames capture all critical evolutionary events, and linear temporal interpolation between them is sufficient to reconstruct intermediate time steps.
    Invoked in Section 3.1 (key frame detection) and Section 3.2.2 / Eq. 12 (temporal interpolation); no validation on held-out time steps is reported.
  • domain assumption The dilated feature regions (coreset) contain all samples needed to train a representation that visualizes the features well.
    Stated in Section 3.1; the evaluation uses the same features for coreset construction and for measuring convergence, so this is not independently tested.
  • domain assumption A shallow MLP (2 hidden layers, 64 neurons) is a sufficient decoder for the Tesseract encoding across all tested datasets.
    Fixed after an ablation on one dataset (Section 4.2) and carried to the other datasets.
  • domain assumption The multi-resolution configuration derived from the FBB with fold=2 gives approximation quality comparable to or better than uniform-resolution grids with the same parameter budget.
    Argued in Section 3.2.1, but the supporting experiment is the confounded comparison in Section 5.2.
  • standard math The MLP with parametric encoding can represent the continuous spatiotemporal field; standard neural-network approximation properties are assumed.
    The paper relies on the universal approximation behavior of MLPs, as is standard in INR literature.
invented entities (1)
  • Tesseract embedding grid (4D multi-resolution parametric grid)
    purpose: Stores trainable embedding vectors indexed by (t,x,y,z) so a small MLP can decode the time-varying volume; enables joint spatial-temporal encoding.
    Introduced in Section 3.2 as the core data structure. It has no observable signature outside the method; its benefit is evidenced only by the paper's dataset-specific experiments, so independent_evidence is False.

how reviews work

0 comments
Cite this review

Pith. "Pith review of F-Hash: Feature-Based Hash Design for Time-Varying Volume Visualization via Multi-Resolution Tesseract Encoding." pith.science (2026). https://pith.science/paper/KK66FNWV

@misc{pith2026250703836,
  author       = {Pith},
  title        = {Pith review of: F-Hash: Feature-Based Hash Design for Time-Varying Volume Visualization via Multi-Resolution Tesseract Encoding},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KK66FNWV}},
  note         = {Machine review of arXiv:2507.03836}
}
read the original abstract

Interactive time-varying volume visualization is challenging due to its complex spatiotemporal features and sheer size of the dataset. Recent works transform the original discrete time-varying volumetric data into continuous Implicit Neural Representations (INR) to address the issues of compression, rendering, and super-resolution in both spatial and temporal domains. However, training the INR takes a long time to converge, especially when handling large-scale time-varying volumetric datasets. In this work, we proposed F-Hash, a novel feature-based multi-resolution Tesseract encoding architecture to greatly enhance the convergence speed compared with existing input encoding methods for modeling time-varying volumetric data. The proposed design incorporates multi-level collision-free hash functions that map dynamic 4D multi-resolution embedding grids without bucket waste, achieving high encoding capacity with compact encoding parameters. Our encoding method is agnostic to time-varying feature detection methods, making it a unified encoding solution for feature tracking and evolution visualization. Experiments show the F-Hash achieves state-of-the-art convergence speed in training various time-varying volumetric datasets for diverse features. We also proposed an adaptive ray marching algorithm to optimize the sample streaming for faster rendering of the time-varying neural representation.

Figures

Figures reproduced from arXiv: 2507.03836 by the authors.

Figure 1
Figure 1. Time-varying volume visualization via proposed F-Hash multi-resolution tesseract encoding on Combustion(a), Argon Bubble [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Coreset selection process to retrieve the coreset for training, occupancy grid, and the FBB. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Occupancy grid (red empty squares) extraction for three types of [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (10 more)
Figure 4
Figure 4. Figure 4: Multi-resolution configuration of Grid-based methods and F-Hash [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Comparison between MHE and F-Hash on collision and bucket [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 7
Figure 7. Figure 7: Quadrilinear interpolation. 3.2.4 Input Encoding The encoding steps for the proposed multi-resolution Tesseract encod￾ing consist of the following steps: 1. Locate neighboring time steps: The input, (t, x, y,z), needs to find its neighboring time steps across the tempo…
Figure 8
Figure 8. Figure 8: Hash function design for mapping the multi-resolution Tesseract [PITH_FULL_IMAGE:figures/full_fig_p006_8.png]
Figure 9
Figure 9. Figure 9: Prediction accuracy (the 10th iteration, results of more iterations [PITH_FULL_IMAGE:figures/full_fig_p006_9.png]
Figure 10
Figure 10. Figure 10: Convergence speed for training the INR using different input [PITH_FULL_IMAGE:figures/full_fig_p007_10.png]
Figure 11
Figure 11. Figure 11: Convergence visualization for the Argon Bubble dataset. [PITH_FULL_IMAGE:figures/full_fig_p007_11.png]
Figure 12
Figure 12. Figure 12: Convergence speed for training the INR using different input [PITH_FULL_IMAGE:figures/full_fig_p008_12.png]
Figure 15
Figure 15. Figure 15: Convergence visualization for the Supernova dataset. [PITH_FULL_IMAGE:figures/full_fig_p008_15.png]
Figure 16
Figure 16. Figure 16: Evolution visualization using different input encoding methods. Reconstruction accuracy is labeled in the format of PSNR/SSIM for each [PITH_FULL_IMAGE:figures/full_fig_p009_16.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

56 extracted references · 27 canonical work pages

  1. [1]

    Z. Bai, Y . Tao, and H. Lin. Time-varying volume visualization: A survey. Journal of Visualization, 23:745–761, 2020. 2, 3

    work page 2020

  2. [2]

    Z. Bai, Y . Tao, and H. Lin. Time-varying volume visualization: a survey. J. Vis., 23(5):745–761, 17 pages, Oct. 2020. doi: 10.1007/s12650-020 -00654-x 7

    work page doi:10.1007/s12650-020 2020

  3. [3]

    Ballester-Ripoll, P

    R. Ballester-Ripoll, P. Lindstrom, and R. Pajarola. Tthresh: Tensor com- pression for multidimensional visual data. IEEE Transactions on Vi- sualization and Computer Graphics , 26(9):2891–2903, 2020. doi: 10. 1109/TVCG.2019.2904063 8

    arXiv 2020

  4. [4]

    J. T. Barron, B. Mildenhall, M. Tancik, P. Hedman, R. Martin-Brualla, and P. P. Srinivasan. Mip-nerf: A multiscale representation for anti-aliasing neural radiance fields. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pp. 5855–5864, October 2021. 2, 3

    work page 2021

  5. [5]

    Bauer, Q

    D. Bauer, Q. Wu, and K.-L. Ma. Fovolnet: Fast volume rendering using foveated deep neural networks. IEEE Transactions on Visualization and Computer Graphics, 29(1):515–525, 2023. doi: 10.1109/TVCG.2022. 3209498 2

    doi:10.1109/tvcg.2022 2023

  6. [6]

    E. R. Chan, C. Z. Lin, M. A. Chan, K. Nagano, B. Pan, S. De Mello, O. Gallo, L. J. Guibas, J. Tremblay, S. Khamis, T. Karras, and G. Wet- zstein. Efficient geometry-aware 3d generative adversarial networks. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 16123–16133, June 2022. 7

    work page 2022

  7. [7]

    Devkota and S

    S. Devkota and S. Pattanaik. Efficient neural representation of volumet- ric data using coordinate-based networks. Computer Graphics Forum, 42(7):e14955, 2023. doi: 10.1111/cgf.14955 2

    work page doi:10.1111/cgf.14955 2023

  8. [8]

    Dutta and H.-W

    S. Dutta and H.-W. Shen. Distribution driven extraction and tracking of features for time-varying data analysis. IEEE Transactions on Visualiza- tion and Computer Graphics, 22(1):837–846, 2016. doi: 10.1109/TVCG. 2015.2467436 2

    arXiv 2016

Show all 56 references
  1. [9]

    Fridovich-Keil, A

    S. Fridovich-Keil, A. Yu, M. Tancik, Q. Chen, B. Recht, and A. Kanazawa. Plenoxels: Radiance fields without neural networks. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 5501–5510, June 2022. 2, 3

    2022

  2. [10]

    Hadadan, S

    S. Hadadan, S. Chen, and M. Zwicker. Neural radiosity. ACM Trans. Graph., 40(6), article no. 236, 11 pages, Dec. 2021. doi: 10.1145/3478513 .3480569 7

    2021 doi

  3. [11]

    Han and C

    J. Han and C. Wang. Ssr-tvd: Spatial super-resolution for time-varying data analysis and visualization. IEEE Transactions on Visualization and Computer Graphics, 28(6):2445–2456, 2022. doi: 10.1109/TVCG.2020. 3032123 1

    2022 doi

  4. [12]

    Han and C

    J. Han and C. Wang. Tsr-vfd: Generating temporal super-resolution for unsteady vector field data. Computers & Graphics, 103:168–179, 2022. doi: 10.1016/j.cag.2022.02.001 1

    2022 doi

  5. [13]

    Han and C

    J. Han and C. Wang. Coordnet: Data generation and visualization gen- eration for time-varying volumes via a coordinate-based neural network. IEEE Transactions on Visualization and Computer Graphics, 29(12):4951– 4963, 2023. doi: 10.1109/TVCG.2022.3197203 1, 2

    2023

  6. [14]

    J. Han, H. Zheng, and C. Bi. Kd-inr: Time-varying volumetric data com- pression via knowledge distillation-based implicit neural representation. IEEE Transactions on Visualization and Computer Graphics, 30(10):6826– 6838, 2024. doi: 10.1109/TVCG.2023.3345373 1, 2

    2024

  7. [15]

    J. Han, H. Zheng, D. Z. Chen, and C. Wang. Stnet: An end-to-end generative framework for synthesizing spatiotemporal super-resolution volumes. IEEE Transactions on Visualization and Computer Graphics, 28(1):270–280, 2022. doi: 10.1109/TVCG.2021.3114815 1

    2022

  8. [16]

    Hospedales, A

    T. Hospedales, A. Antoniou, P. Micaelli, and A. Storkey. Meta-learning in neural networks: A survey. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(9):5149–5169, 2022. doi: 10.1109/TPAMI.2021. 3079209 2

    2022 doi

  9. [17]

    Huang and H

    C. Huang and H. Wang. A novel key-frames selection framework for comprehensive video summarization. IEEE Transactions on Circuits and Systems for Video Technology, 30(2):577–589, 2020. doi: 10.1109/TCSVT .2019.2890899 3

    2020

  10. [18]

    Kehrer and H

    J. Kehrer and H. Hauser. Visualization and visual analysis of multifaceted scientific data: A survey. IEEE Transactions on Visualization and Com- puter Graphics, 19(3):495–513, 2013. doi: 10.1109/TVCG.2012.110 7

    2013 doi

  11. [19]

    Kumpf, M

    A. Kumpf, M. Rautenhaus, M. Riemer, and R. Westermann. Visual analysis of the temporal evolution of ensemble forecast sensitivities. IEEE Transactions on Visualization and Computer Graphics , 25(1):98–108,

  12. [20]

    Liang, K

    X. Liang, K. Zhao, S. Di, S. Li, R. Underwood, A. M. Gok, J. Tian, J. Deng, J. C. Calhoun, D. Tao, Z. Chen, and F. Cappello. Sz3: A modular framework for composing prediction-based error-bounded lossy compres- sors. IEEE Transactions on Big Data , 9(2):485–498, 2023. doi: 10. ...

    2023

  13. [21]

    Ljung, J

    P. Ljung, J. Krüger, E. Groller, M. Hadwiger, C. D. Hansen, and A. Yn- nerman. State of the art in transfer functions for direct volume rendering. Computer Graphics Forum, 35(3):669–691, 2016. doi: 10.1111/cgf.12934 2

    2016 doi

  14. [22]

    Y . Lu, K. Jiang, J. A. Levine, and M. Berger. Compressive neural represen- tations of volumetric scalar fields. Computer Graphics Forum, 40(3):135– 146, 2021. doi: 10.1111/cgf.14295 2, 8

    2021 doi

  15. [23]

    Lukasczyk, G

    J. Lukasczyk, G. Aldrich, M. Steptoe, G. Favelier, C. Gueunet, J. Tierny, R. Maciejewski, B. Hamann, and H. Leitte. Viscous fingering: A topologi- cal visual analytic approach. In Physical Modeling for Virtual Manufactur- ing Systems and Processes, vol. 869 of Applied Mechanic...

    2017 doi

  16. [24]

    K.-L. Ma. Visualizing time-varying volume data. Computing in Science & Engineering, 5(2):34–42, 2003. doi: 10.1109/MCISE.2003.1182960 2

    2003 arXiv

  17. [25]

    Mehta, M

    I. Mehta, M. Gharbi, C. Barnes, E. Shechtman, R. Ramamoorthi, and M. Chandraker. Modulated periodic activations for generalizable local functional representations. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pp. 14214–14223, October 2021. 2

    2021

  18. [26]

    Mildenhall, P

    B. Mildenhall, P. P. Srinivasan, M. Tancik, J. T. Barron, R. Ramamoorthi, and R. Ng. Nerf: representing scenes as neural radiance fields for view synthesis. Commun. ACM, 65(1):99–106, 8 pages, Dec. 2021. doi: 10. 1145/3503250 2, 3

    2021

  19. [27]

    Müller, A

    T. Müller, A. Evans, C. Schied, and A. Keller. Instant neural graphics primitives with a multiresolution hash encoding. ACM Trans. Graph., 41(4), article no. 102, 15 pages, July 2022. doi: 10.1145/3528223.3530127 2, 3, 7

    2022

  20. [28]

    Müller, F

    T. Müller, F. Rousselle, J. Novák, and A. Keller. Real-time neural radiance caching for path tracing. ACM Trans. Graph., 40(4), article no. 36, 16 pages, July 2021. doi: 10.1145/3450626.3459812 7

    2021

  21. [29]

    Saikia and T

    H. Saikia and T. Weinkauf. Global feature tracking and similarity es- timation in time-dependent scalar fields. Computer Graphics Forum , 36(3):1–11, 2017. doi: 10.1111/cgf.13163 2

    2017 doi

  22. [30]

    Sitzmann, J

    V . Sitzmann, J. Martel, A. Bergman, D. Lindell, and G. Wetzstein. Implicit neural representations with periodic activation functions. In H. Larochelle, M. Ranzato, R. Hadsell, M. Balcan, and H. Lin, eds., Advances in Neural Information Processing Systems, vol. 33, pp. 7462–74...

    2020

  23. [31]

    Sitzmann, M

    V . Sitzmann, M. Zollhoefer, and G. Wetzstein. Scene representation networks: Continuous 3d-structure-aware neural scene representations. In H. Wallach, H. Larochelle, A. Beygelzimer, F. d 'Alché-Buc, E. Fox, and R. Garnett, eds., Advances in Neural Information Processing Syst...

    2019

  24. [32]

    C. Sun, M. Sun, and H.-T. Chen. Direct voxel grid optimization: Super- fast convergence for radiance fields reconstruction. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 5459–5469, June 2022. 2, 3

    2022

  25. [33]

    J. Sun, D. Lenz, H. Yu, and T. Peterka. Make the fastest faster: Importance mask for interactive volume visualization using reconstruction neural networks. arXiv preprint arXiv:2502.06053, 2025. 2

    2025

  26. [34]

    J. Sun, X. Xie, and H. Yu. Rmdncache: Dual-space prefetching neural network for large-scale volume visualization. IEEE Transactions on Visualization and Computer Graphics , pp. 1–13, 2024. doi: 10.1109/ TVCG.2024.3410091 2

    2024

  27. [35]

    Tancik, P

    M. Tancik, P. Srinivasan, B. Mildenhall, S. Fridovich-Keil, N. Raghavan, U. Singhal, R. Ramamoorthi, J. Barron, and R. Ng. Fourier features let networks learn high frequency functions in low dimensional domains. In H. Larochelle, M. Ranzato, R. Hadsell, M. Balcan, and H. Lin, ...

  28. [36]

    D. Tang, S. Singh, P. A. Chou, C. Hane, M. Dou, S. Fanello, J. Taylor, P. Davidson, O. G. Guleryuz, Y . Zhang, et al. Deep implicit volume compression. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 1293–1303, 2020. 2

    2020

  29. [37]

    H. Tang, L. Ding, S. Wu, B. Ren, N. Sebe, and P. Rota. Deep unsuper- vised key frame extraction for efficient video classification. ACM Trans. Multimedia Comput. Commun. Appl., 19(3), article no. 119, 17 pages, Feb

  30. [38]

    Tang and C

    K. Tang and C. Wang. Ecnr: Efficient compressive neural representation of time-varying volumetric datasets. arXiv preprint arXiv:2311.12831,

    arXiv

  31. [39]

    Tang and C

    K. Tang and C. Wang. Stsr-inr: Spatiotemporal super-resolution for mul- tivariate time-varying volumetric data via implicit neural representation. Computers & Graphics, 119:103874, 2024. doi: 10.1016/j.cag.2024.01. 001 1, 2

    2024 doi

  32. [40]

    Teschner, B

    M. Teschner, B. Heidelberger, M. Müller, D. Pomerantes, and M. H. Gross. Optimized spatial hashing for collision detection of deformable objects. In Vmv, vol. 3, pp. 47–54, 2003. 5

    2003

  33. [41]

    Wang, T.-H

    K.-C. Wang, T.-H. Wei, N. Shareef, and H.-W. Shen. Ray-based explo- ration of large time-varying volume data using per-ray proxy distributions. IEEE Transactions on Visualization and Computer Graphics, 26(11):3299– 3313, 2020. doi: 10.1109/TVCG.2019.2920130 2

    2020

  34. [42]

    Weiss, P

    S. Weiss, P. Hermüller, and R. Westermann. Fast neural representations for direct volume rendering. Computer Graphics Forum, 41(6):196–211,

  35. [43]

    Widanagamaachchi, J

    W. Widanagamaachchi, J. Chen, P. Klacansky, V . Pascucci, H. Kolla, A. Bhagatwala, and P.-T. Bremer. Tracking features in embedded surfaces: Understanding extinction in turbulent combustion. In 2015 IEEE 5th Symposium on Large Data Analysis and Visualization (LDAV), pp. 9–16,

    2015

  36. [44]

    Widanagamaachchi, A

    W. Widanagamaachchi, A. Jacques, B. Wang, E. Crosman, P.-T. Bremer, V . Pascucci, and J. Horel. Exploring the evolution of pressure-perturbations to understand atmospheric phenomena. In 2017 IEEE Pacific Visualization Symposium (PacificVis), pp. 101–110, 2017. doi: 10.1109/PAC...

    2017

  37. [45]

    Q. Wu, D. Bauer, M. J. Doyle, and K.-L. Ma. Interactive volume visualiza- tion via multi-resolution hash encoding based neural representation. IEEE Transactions on Visualization and Computer Graphics, 30(8):5404–5418,

  38. [46]

    M. Yang, K. Tang, and C. Wang. Meta-inr: Efficient encoding of volumet- ric data via meta-learning implicit neural representation. arXiv preprint arXiv:2502.09669, 2025. 2

    2025 arXiv

  39. [47]

    S. Yao, Y . Lu, and C. Wang. Visnerf: Efficient multidimensional neu- ral radiance field representation for visualization synthesis of dynamic volumetric scenes. arXiv preprint arXiv:2502.16731, 2025. 2

    2025 arXiv

  40. [48]

    Yariv, J

    L. Yariv, J. Gu, Y . Kasten, and Y . Lipman. V olume rendering of neural implicit surfaces. In M. Ranzato, A. Beygelzimer, Y . Dauphin, P. Liang, and J. W. Vaughan, eds., Advances in Neural Information Processing Systems, vol. 34. Curran Associates, Inc., 2021. 2

    2021

  41. [49]

    Zhong, Y

    Q. Zhong, Y . Zhang, J. Zhang, K. Shi, Y . Yu, and C. Liu. Key frame extrac- tion algorithm of motion video based on priori. IEEE Access, 8:174424– 174436, 2020. doi: 10.1109/ACCESS.2020.3025774 3

    2020

  42. [50]

    Zhou and Y .-J

    B. Zhou and Y .-J. Chiang. Key time steps selection for large-scale time- varying volume datasets using an information-theoretic storyboard. Com- puter Graphics Forum, 37(3):37–49, 2018. doi: 10.1111/cgf.13399 2

    2018 doi

  43. [2015]

    doi: 10.1109/LDA V.2015.7348066 2

    2015

  44. [2019]

    doi: 10.1109/TVCG.2018.2864901 2

    2018

  45. [2022]

    doi: 10.1111/cgf.14578 2

    doi

  46. [2023]

    doi: 10.1145/3571735 3

    doi

  47. [2024]

    doi: 10.1109/TVCG.2023.3293121 2, 3, 6

    2023

  48. [7547]

    Curran Associates, Inc., 2020. 2, 3

    2020

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.