Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

ScaleFree: Dynamic KDE for Multiscale Point Cloud Exploration in VR

T0 review · 3 major / 4 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read ScaleFree claims that GPU-accelerated adaptive kernel density estimation can recompute density fields on the fly in VR, and that dynamic fields raise multiscale selection accuracy well above precomputed ones.

desk verdict ScaleFree delivers a real user-validated improvement in VR density-based selection by recomputing adaptive KDE on the GPU, but the billion-particle scalability claim outruns the demonstrated algorithm. read the letter →

arxiv 2601.20758 v2 pith:POT2LITB submitted 2026-01-28 cs.HC

Lixiang Zhao , Fuqi Xie , Tobias Isenberg , Hai-Ning Liang , Lingyun Yu This is my paper
classification cs.HC
keywords multiscalevisualizationkerneldensityestimationadaptiveKDEGPUaccelerationvirtualrealityinteractivedataselectionpointcloudnavigationcosmologicalsimulation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper addresses the bottleneck that keeps kernel density estimation—converting discrete point clouds into continuous scalar fields—from being used dynamically in virtual reality. It argues that by parallelizing an adaptive KDE on the GPU, with a k-d-tree range search for the pilot density and thread-group shared-memory reduction for per-particle smoothing lengths, the density field can be recomputed whenever the user's scale of view changes. In a 24-participant VR study comparing dynamic recomputation with precomputed single-resolution and precomputed multi-resolution fields, the dynamic field produced a combined precision-recall score (F1) of 0.85 versus 0.64 and 0.50, with lower completion time and workload. If this holds, multiscale exploration of unstructured point clouds—including cosmological simulations—could abandon precomputed density ladders and instead build selection volumes and navigation targets from the scale actually on screen. The paper's performance tests show order-of-magnitude speedups over CPU baselines, though the demonstrated datasets top out at 442,000 particles.

What carries the argument

Load-bearing object: modified adaptive KDE with a finite-support ellipsoidal kernel—each particle contributes to a grid node only inside an ellipsoid whose semi-axes equal that particle's smoothing length, which is large in sparse regions and small in dense ones, capped at five grid spacings. Speedup mechanism: a staged GPU pipeline—gather-style pilot-density kernel with k-d-tree range queries (the only spatially culled stage), an adaptive-length kernel using thread-group shared-memory reduction for mean pilot density, and a final node-per-thread kernel that accumulates all particles. The paper's complexity analysis covers the first two stages; the final stage is a direct product of node and

What would settle it

Run the final-density kernel alone on a fixed 64^3 grid while increasing particle count from 10^5 to 10^8 on the same GPU and plot time against particle count. If the time grows linearly—reaching well beyond an interactive budget at 10^8—then the billion-particle recomputation claim is refuted; at 10^9 particles the kernel as written requires about 2.6×10^14 ellipsoidal-kernel evaluations per density field.

Watch

Extended reading notes

Core claim

On its own terms, the discovery is that the practical obstacle to dynamic kernel density estimation is not the estimator but the pipeline. A modified adaptive estimator with a finite-support ellipsoidal kernel—small smoothing lengths in dense regions, large ones in sparse regions—can be staged as three GPU kernels: pilot density via k-d-tree range queries, per-particle adaptive smoothing lengths via thread-group parallel reduction, and a final per-node accumulation over all particles. With this pipeline, a fresh density field is ready shortly after a scale change. In the paper's user study, selection built from the dynamic field reached F1 = 0.85 and MCC = 0.83, versus F1 = 0.64/MCC = 0.63 f

Load-bearing premise

Every grid node in the final density field is recomputed by looping over every particle, and the interactive timings cover at most 442,000 particles; the claim that the same pipeline serves billion-particle simulations rests on a scaling step the paper does not demonstrate.

Editorial extensions

If this is right

  • Density fields no longer need to be precomputed as single-resolution or mipmap structures; a field can be generated for the exact scale the user is viewing, which is the property the study connects to higher selection accuracy.
  • Unstructured datasets without predefined hierarchies become explorable by a 'show me that in more detail' interaction: selecting a region triggers recomputation and progressive zooming toward it.
  • Precomputed low-resolution fields systematically omit fine-scale structures, so any visualization tool that relies on them will force users into extra refinement; the study's accuracy gap (0.85 vs 0.64 vs 0.50 F1) is the measured cost of that omission.
  • The GPU-to-CPU transfer of the density field is the current latency bottleneck, so moving selection-volume extraction entirely to the GPU would make updates still smoother.
  • A fixed 64^3 grid still bounds the finest captured detail; even with on-the-fly recomputation, features smaller than the grid spacing require a feature-aware resolution scheme.

Reading between the lines

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

  • The billion-particle framing is not yet covered by the measurements: the final density kernel is O(grid nodes × particles) with no culling, and at 10^9 particles on a 64^3 grid a single recomputation would need about 2.6×10^14 kernel evaluations, so the scalability claim would require a spatial acceleration for the final pass that the paper does not implement.
  • A clean ablation would isolate why dynamic fields help: keep the threshold-based selection algorithm fixed and vary only whether the density field is recomputed at the current scale versus precomputed, to separate the benefit of freshness from the benefit of finer effective resolution.
  • The study's threshold-based selection may gain as much from re-deriving the density threshold at the current scale as from KDE speed; comparing against a condition that recomputes thresholds but not adaptive bandwidths would test this.
  • A practical extension, already sketched in the limitations, is to trigger recomputation only when view change exceeds a threshold and to recompute only the region around the user's focus; that would convert the current 'recompute on every scale change' policy into a budget-aware one.
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

3 major / 4 minor

Summary. ScaleFree is a GPU-accelerated adaptive KDE pipeline for dynamic density-field recomputation during multiscale point-cloud exploration in VR. The method has three stages: pilot density estimation using k-d-tree range queries (Alg. 2), adaptive smoothing-length computation via parallel reduction (Alg. 3), and final density estimation over the grid (Alg. 4). The authors report strong speedups over single-core/multi-core CPU baselines on cosmological point clouds (76k–442k particles), and a preregistered 24-participant VR user study comparing the dynamic approach (DR) against precomputed single-resolution (PS) and multi-resolution (PM) density fields. They report substantially higher F1/MCC for DR, no clear time difference, lower workload, and higher preference.

Significance. If the claims are taken at their demonstrated scale, the paper makes a useful contribution: it shows that adaptive KDE can be recomputed on the GPU fast enough for interactive VR selection on datasets up to about 442k particles, and the user study provides credible evidence (F1 0.85 vs. 0.64/0.50; MCC ratios with CIs excluding 1) that dynamic density fields improve multiscale selection accuracy relative to precomputed fields. The preregistration, open data/code links, and estimation-style reporting with CIs are methodological strengths. However, the abstract's 'billions of particles' claim is not supported by the algorithm as written, because the final density estimation (Alg. 4) is O(M·N) with no spatial culling, and the paper's own complexity discussion and future-work section explicitly limit the current implementation to much smaller scales. This gap affects the central significance claim and needs to be resolved before publication.

major comments (3)
  1. [Abstract; Sec. 3.2.3 (Alg. 4); Sec. 3.3; Sec. 6.3] Algorithm 4's FDE kernel loops over every particle for every grid node (line 8: for i in 0 : particles.count-1), with no k-d-tree query or spatial culling; each recomputation is O(M·N). The complexity analysis in Sec. 3.3 covers only the pilot-density stage (O(M√N)) and explicitly states 'hundreds of thousands of particles,' not billions. Sec. 6.3 confirms the current implementation computes KDE over the whole data space. At res=64^3 (M=262,144) and N=1e9, FDE alone would require ~2.6e14 kernel evaluations, extrapolating to roughly 700 s per recomputation on the measured RTX 4090—far outside interactive rates. Please either add spatial culling/query to FDE and analyze it, or revise the billion-particle claim and all derivative statements to the tested range (≤442k).
  2. [Sec. 3.3 / Table 1; Sec. 6.3] The reported ScaleFree timings (0.042–0.309 s) explicitly exclude CPU–GPU data transfer, yet Sec. 6.3 states that the main time cost of recomputation delay lies in transferring the density field from GPU to CPU. Thus Table 1 does not demonstrate the end-to-end latency of the actual VR interaction pipeline. Please report transfer-inclusive timings (or list transfer separately) and re-evaluate the 'real-time' and 'smooth navigation' claims against those numbers.
  3. [Sec. 5.3 / Fig. 6(c); Table 3] Hypothesis H3 is declared supported even though the paper's own pairwise ratio CIs include 1: DR/PM time ratio = 0.89 [0.70,1.08] and DR/PS = 0.86 [0.66,1.08]. Under the estimation framework the authors adopt, a 95% CI covering 1 means no clear evidence of a difference; the text even acknowledges some participants were faster with PM or PS. Please rephrase as 'no conclusive completion-time benefit' or restrict the claim to the observed direction of the means.
minor comments (4)
  1. [Eq. (2) vs. Eq. (4) and Alg. 2 line 15] The kernel argument is written as E(||r^(n;j)||) in Eqs. (2)/(7), while Eq. (4) and Alg. 2 use E(x)=1−x^2 with what is effectively ||r||^2. Please make the notation consistent (e.g., E(||r||^2)).
  2. [Sec. 3.1] Typo: 'the the jth particle's position' should read 'the jth particle's position.'
  3. [Alg. 1] The last dispatch in Algorithm 1 uses 'FSDEtz' where it should be 'FDEtz'.
  4. [Alg. 3] If particles.count is not divisible by ASLtx, the last thread group may contain fewer than ASLtx active threads, and sharedDen entries for inactive lanes are not initialized before the parallel reduction. Please clarify how partial groups are handled.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found; the KDE derivation and user-study comparisons are self-contained, though the billion-particle scalability claim is unsupported (a correctness concern, not circularity).

full rationale

Walking the claimed derivation chain: the KDE formulation (Eqs. 1–8) is a standard modified-Breiman adaptive Epanechnikov estimator; no parameter is fitted to the outcome being predicted, and no `prediction` is a renamed fit. The user study compares DR against precomputed single-resolution (PS) and multi-resolution (PM) density fields, which are external references; the DR advantage (F1 0.85 vs 0.64/0.50) is an empirical result, not an algebraic consequence of the definitions. Self-citations to MeTAPoint [75] and CAST [68,69] are used as the common selection interaction across all conditions, so they do not load-bear the comparison; no uniqueness theorem is invoked, and no ansatz is smuggled via self-citation. The main weakness is a scalability gap: Algorithm 4 (FDE, Sec. 3.2.3) is plainly O(M·N) with a loop over all particles for every grid node and no spatial culling, Sec. 3.3 only claims 'hundreds of thousands of particles', and Sec. 6.3 admits the current implementation computes KDE over 'the whole data space' rather than viewport subsets. The abstract's 'billions of particles' claim is therefore unsupported beyond the tested 442k-particle regime (about 2.6e14 kernel evaluations per recomputation at N=1e9, roughly 700 s on the reported hardware). This is a correctness/overclaim issue, not circularity, because the paper's equations and comparisons do not presuppose the target result.

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

The paper introduces no new entities or objects; it combines known algorithmic primitives (adaptive KDE, k-d trees, GPU reduction) into a system. The free parameters are design/performance choices, not fitted to force the reported outcomes.

free parameters (4)
  • Grid resolution res = 64^3 (262,144 nodes)
    Fixed resolution for the density field in dynamic and PS conditions; chosen as a performance/accuracy tradeoff, and effectively gives higher local resolution when the box B follows the viewport, contributing to the user-study accuracy advantage.
  • Smoothing-length cap 5·s_k = 5·node spacing per axis
    Hand-chosen upper bound in Eq. 6 to limit kernel extent and computational cost; not derived from data, affects final density smoothness.
  • Initial smoothing length ℓ_k = 2(P80−P20)/log N = Data-dependent percentiles
    Heuristic base bandwidth adopted from prior work (Zhao et al. 2024 [75]); determines the pilot density and indirectly the adaptive bandwidths.
  • Thread group dimensions = PDE/FDE tx=8; ASL tx=1024
    Implementation tuning for GPU occupancy; affects performance but not the algorithmic result.
assumptions (3)
  • standard math Modified Breiman adaptive KDE with Epanechnikov kernel yields a valid density estimate for multiscale point clouds.
    Invoked in Sec. 3.1 (Eqs. 1–8) as the foundation of the density fields; the paper does not prove properties but relies on established KDE theory.
  • domain assumption k-d tree range-query complexity is O(√N + K) for the pilot density stage.
    Sec. 3.2.1 assumes this complexity after citing Kakde [34]; it depends on low-dimensional balanced data and degrades for adversarial distributions.
  • domain assumption The NASA-TLX workload instrument and participant rankings adequately measure cognitive load and preference.
    Standard HCI practice invoked in Sec. 5.1–5.3 to support H5/H6; accepted as domain convention.

how reviews work

0 comments
Cite this review

Pith. "Pith review of ScaleFree: Dynamic KDE for Multiscale Point Cloud Exploration in VR." pith.science (2026). https://pith.science/paper/POT2LITB

@misc{pith2026260120758,
  author       = {Pith},
  title        = {Pith review of: ScaleFree: Dynamic KDE for Multiscale Point Cloud Exploration in VR},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/POT2LITB}},
  note         = {Machine review of arXiv:2601.20758}
}
read the original abstract

We present ScaleFree, a GPU-accelerated adaptive Kernel Density Estimation (KDE) algorithm for scalable, interactive multiscale point cloud exploration. With this technique, we cater to the massive datasets and complex multiscale structures in advanced scientific computing, such as cosmological simulations with billions of particles. Effective exploration of such data requires a full 3D understanding of spatial structures, a capability for which immersive environments such as VR are particularly well suited. However, simultaneously supporting global multiscale context and fine-grained local detail remains a significant challenge. A key difficulty lies in dynamically generating continuous density fields from point clouds to facilitate the seamless scale transitions: while KDE is widely used, precomputed fields restrict the accuracy of interaction and omit fine-scale structures, while dynamic computation is often too costly for real-time VR interaction. We address this challenge by leveraging GPU acceleration with k-d-tree-based spatial queries and parallel reduction within a thread group for on-the-fly density estimation. With this approach, we can recalculate scalar fields dynamically as users shift their focus across scales. We demonstrate the benefits of adaptive density estimation through two data exploration tasks: adaptive selection and progressive navigation. Through performance experiments, we demonstrate that ScaleFree with GPU-parallel implementation achieves orders-of-magnitude speedups over sequential and multi-core CPU baselines. In a controlled experiment, we further confirm that our adaptive selection technique improves accuracy and efficiency in multiscale selection tasks.

Figures

Figures reproduced from arXiv: 2601.20758 by the authors.

Figure 1
Figure 1. ScaleFree dynamically recomputes density fields on the GPU, enabling on-the-fly density estimation of regions of interest [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Experiment datasets: (a) Nbody 1 (76k points), (b) Nbody 2 [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Our progressive navigation technique. it to the GPU. A kernel then executes the Marching Cubes algo￾rithm [41,64] to extract the selection volume as an iso-surface based on this threshold. We then select points within the extracted vol￾ume. Thanks to our optimized computation, the selection volume is generated rapidly. Even with the MeTAPoint technique [75], which detects density thresholds on-the-fly at the user’s … view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: One of the datasets we used in our study. [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 6
Figure 6. Figure 6: Pairwise ratio of (a) F1, (b) MCC, and (c) time, 95% CIs. [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: User workload and performance. Error bars: 95% CIs. [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 10
Figure 10. Figure 10: Workflow of progressive navigation method. [PITH_FULL_IMAGE:figures/full_fig_p012_10.png]
Figure 11
Figure 11. Figure 11: Workflow of the GPU-accelerated KDE algorithm. [PITH_FULL_IMAGE:figures/full_fig_p012_11.png]
Figure 12
Figure 12. Figure 12: Experiment datasets [PITH_FULL_IMAGE:figures/full_fig_p014_12.png]

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. DP-LENS: A Density-Aware Polyfocal Lens with Topology-Driven Auto-Routing for Occlusion Management in Immersive 3D Analytics

    cs.HC 2026-07 conditional novelty 6.0 of 10

    A density-aware polyfocal fisheye lens with voice-initiated topology-driven auto-routing reduced task time, cognitive load, and physical fatigue in two VR user studies.

Reference graph

Works this paper leans on

75 extracted references · cited by 1 Pith paper

  1. [1]

    Argelaguet and M

    F. Argelaguet and M. Maignant. GiAnt: Stereoscopic-compliant multi- scale navigation in VEs. InProc. VRST, pp. 269–277. ACM, New York,

  2. [2]

    Bacim, D

    F. Bacim, D. Bowman, and M. Pinho. Wayfinding techniques for multiscale virtual environments. InProc. 3DUI, pp. 67–74. IEEE Comp. Soc., Los Alamitos, 2009. doi: 10/fg3dbh 2

    2009

  3. [3]

    Backurs, P

    A. Backurs, P. Indyk, and T. Wagner. Space and time efficient kernel density estimation in high dimensions.Adv Neural Inf Process Syst, 32, art. no. 9251, 10 pages, 2019. url: papers.nips.cc/paper files/ paper/2019/hash/a2ce8f1706e52936dfad516c23904e3e -Abstract.html. 3

    2019

  4. [4]

    T. Baguley. Standardized or simple effect size: What should be re- ported?Br J Psychol, 100(3):603–617, 2009. doi: 10/bnw2nb 7

    2009

  5. [5]

    Besanc ¸on, A

    L. Besanc ¸on, A. Ynnerman, D. F. Keefe, L. Yu, and T. Isenberg. The state of the art of spatial interfaces for 3D visualization.Comput Graph Forum, 40(1):293–326, 2021. doi: 10/gjbpxp 6

    2021

  6. [6]

    R. Brown. HTC vive pro2. Web site: vr-compare.com/headset/ htcvivepro2. Last accessed: Sep. 2025. 7

    2025

  7. [7]

    Brunsdon

    C. Brunsdon. Estimating probability surfaces for geographical point data: An adaptive kernel algorithm.Comput GeoSci, 21(7):877–894,

  8. [8]

    Bueckle, K

    A. Bueckle, K. Buehling, P. C. Shih, and K. B¨orner. 3D virtual real- ity vs. 2D desktop registration user interface comparison.PloS one, 16(10):e0258103, 2021. doi: 10/qmkr 1

    2021

Show all 75 references
  1. [9]

    J. E. Burt, G. M. Barber, and D. L. Rigby.Elementary Statistics for Geographers. Guilford Press, New York, 3rd ed., 2009. doi: 10/br6ngd 3

    2009

  2. [10]

    Z. Cao, J. Han, S. Yang, and X. Jin. Fast best viewpoint selection with geometry-enhanced multiple views and cross-modal distillation.Vis Comput, 41(7):5075–5086, 2025. doi: 10/p5p3 7

    2025

  3. [11]

    S. K. Card, G. G. Robertson, and J. D. Mackinlay. The information visualizer, an information workspace. InProc. CHI, pp. 181–186. ACM, New York, 1991. doi: 10/cvtdps 6

    1991

  4. [12]

    Cautun, R

    M. Cautun, R. van de Weygaert, and B. J. Jones. NEXUS: Tracing the cosmic web connection.Mon Not R Astron Soc, 429(2):1286–1308,

  5. [13]

    T. N. Chan, R. Cheng, and M. L. Yiu. QUAD: Quadratic-bound-based kernel density visualization. InProc. SIGMOD, pp. 35–50. ACM, New York, 2020. doi: 10/gkpt7c 3

    2020

  6. [14]

    T. N. Chan, P. L. Ip, B. Zhu, D. Wu, J. Xu, C. S. Jensen et al. Large- scale spatiotemporal kernel density visualization. InProc. ICDE, pp. 99–113. IEEE Comp. Soc., Los Alamitos, 2025. doi: 10/qmw4 3, 9

    2025

  7. [15]

    Y .-C. Chen, S. Ho, P. E. Freeman, C. R. Genovese, and L. Wasserman. Cosmic web reconstruction through density ridges: Method and algo- rithm.Mon Not R Astron Soc, 454(1):1140–1156, 2015. doi: 10/f7wrzq 1

    2015

  8. [16]

    Chheang, B

    V . Chheang, B. T. Weston, R. W. Cerda, B. Au, B. Giera, P.-T. Bremer et al. A virtual environment for collaborative inspection in additive manufacturing. InProc. CHI, art. no. 26, 7 pages. ACM, New York,

  9. [17]

    I. Cho, J. Li, and Z. Wartell. Multi-scale 7DOF view adjustment.IEEE Trans Vis Comput Graph, 24(3):1331–1344, 2018. doi: 10/gcx789 2

    2018

  10. [18]

    G. Cumming. The new statistics: Why and how.Psychol Sci, 25(1):7– 29, 2014. doi: 10/5k3 7

    2014

  11. [19]

    P. Diggle. A kernel method for smoothing point process data.J R Stat Soc C (Appl Stat), 34(2):138–147, 1985. doi: 10/bsh8ht 1, 3

    1985

  12. [20]

    Dragicevic

    P. Dragicevic. Fair statistical communication in HCI. In J. Robertson and M. Kaptein, eds.,Modern Statistical Methods for HCI, chap. 13, pp. 291–330. Springer, Cham, 2016. doi: 10/ggd8gc 7

    2016

  13. [21]

    Dragicevic, F

    P. Dragicevic, F. Chevalier, and S. Huot. Running an HCI experiment in multiple parallel universes. InCHI Extended Abstracts, pp. 607–618. ACM, New York, 2014. doi: 10/gpcs8n 7

    2014

  14. [22]

    V . A. Epanechnikov. Non-parametric estimation of a multivariate probability density.Theory Probab Appl, 14(1):153–158, 1969. doi: 10/bjdbsc 3

    1969

  15. [23]

    Ferdosi, H

    B. Ferdosi, H. Buddelmeijer, S. Trager, M. Wilkinson, and J. Roerdink. Comparison of density estimation methods for astronomical datasets. Astron Astrophys, 531, art. no. A114, 16 pages, 2011. doi: 10/cj3s5m 3

    2011

  16. [24]

    C. H. Fleming and J. M. Calabrese. A new kernel density estimator for accurate home-range and species-range area estimation.Methods Ecol Evol, 8(5):571–579, 2017. doi: 10/f98k6z 3

    2017

  17. [25]

    Fu, W.-B

    C.-W. Fu, W.-B. Goh, and J. A. Ng. Multi-touch techniques for ex- ploring large-scale 3D astrophysical simulations. InProc. CHI, pp. 2213–2222. ACM, New York, 2010. doi: 10/drxzf7 2

    2010

  18. [26]

    Gansner, Y

    E. Gansner, Y . Koren, and S. North. Topological fisheye views for visualizing large graphs.IEEE Trans Vis Comput Graph, 11(4):457– 468, 2005. doi: 10/cgmnfg 3

    2005

  19. [27]

    R. J. Garc ´ıa-Hern´andez, C. Anthes, M. Wiedemann, and D. Kran- zlm¨uller. Perspectives for using virtual reality to extend visual data mining in information visualization. InProc. Aerospace Conf., pp. 1151–1161. IEEE Comp. Soc., Los Alamitos, 2016. doi: 10/gtmkhw 1

    2016

  20. [28]

    O. Hahn, C. Porciani, C. M. Carollo, and A. Dekel. Properties of dark matter haloes in clusters, filaments, sheets and voids.Mon Not R Astron Soc, 375(2):489–499, 2007. doi: 10/dxbdjj 1, 2

    2007

  21. [29]

    Halladjian, D

    S. Halladjian, D. Kouˇril, H. Miao, M. E. Gr¨oller, I. Viola, and T. Isen- berg. Multiscale unfolding: Illustratively visualizing the whole genome at a glance.IEEE Trans Vis Comput Graph, 28(10):3456–3470, 2022. doi: 10/kt3m 2

    2022

  22. [30]

    Halladjian, H

    S. Halladjian, H. Miao, D. Kouˇril, M. E. Gr¨oller, I. Viola, and T. Isen- berg. Scale trotter: Illustrative visual travels across negative scales. IEEE Trans Vis Comput Graph, 26(1):654–664, 2020. doi: 10/kt3k 2

    2020

  23. [31]

    S. Hart. Nasa-task load index (NASA-TLX); 20 years later.Proc Hum Factors Ergon Soc Annu Meet, 50(9):904–908, 2006. doi: 10/fzvtd4 7

    2006

  24. [32]

    B. He, N. K. Govindaraju, Q. Luo, and B. Smith. Efficient gather and scatter operations on graphics processors. InProc. SC, art. no. 46, 12 pages. ACM, New York, 2007. doi: 10/fgtjzk 4

    2007

  25. [33]

    W.-H. Hsu, Y . Zhang, and K.-L. Ma. A multi-criteria approach to camera motion design for volume data animation.IEEE Trans Vis Comput Graph, 19(12):2792–2801, 2013. doi: 10/f5h3m3 6

    2013

  26. [34]

    H. M. Kakde. Range searching using KD tree. Technical report, Dept. of Computer Science, Florida State Univ., 2005. Available: https: //users.cs.utah.edu/%7Elifeifei/cis5930/kdtree.pdf. 4

    2005

  27. [35]

    Klacansky, H

    P. Klacansky, H. Miao, A. Gyulassy, A. Townsend, K. Champley, J. Tringe et al. Virtual inspection of additively manufactured parts. In Proc. PacificVis, pp. 81–90. IEEE Comp. Soc., Los Alamitos, 2022. doi: 10/g92v9x 2

    2022

  28. [36]

    Kopper, T

    R. Kopper, T. Ni, D. A. Bowman, and M. Pinho. Design and evaluation of navigation techniques for multiscale virtual environments. InProc. VR, pp. 175–182. IEEE Comp. Soc., Los Alamitos, 2006. doi: 10/ cjs6g6 2

    2006

  29. [37]

    Kouˇril, T

    D. Kouˇril, T. Isenberg, B. Kozl´ıkov´a, M. Meyer, M. E. Gr ¨oller, and I. Viola. HyperLabels: Browsing of dense and hierarchical molecular 3D models.IEEE Trans Vis Comput Graph, 27(8):3493–3504, 2021. doi: 10/kt4m 2

    2021

  30. [38]

    B. Laha, K. Sensharma, J. D. Schiffbauer, and D. A. Bowman. Effects of immersion on visual analysis of volume data.IEEE Trans Vis Comput Graph, 18(4):597–606, 2012. doi: 10/fx9b5g 1

    2012

  31. [39]

    O. D. Lampe and H. Hauser. Interactive visualization of streaming data with kernel density estimation. InProc. Pacific Vis, pp. 171–178. IEEE Comp. Soc., Los Alamitos, 2011. doi: 10/c866v9 3

    2011

  32. [40]

    J. J. LaViola, D. A. Feliz, D. F. Keefe, and R. C. Zeleznik. Hands-free multi-scale navigation in virtual environments. InProc. I3D, pp. 9–15. ACM, New York, 2001. doi: 10/bdmkfc 2

    2001

  33. [41]

    W. E. Lorensen and H. E. Cline. Marching cubes: A high resolution 3D surface construction algorithm.ACM SIGGRAPH Comput Graph, 21(4):163–169, 1987. doi: 10/ft9gsh 6

    1987

  34. [42]

    Olsen, R

    K. Olsen, R. M. H. Lindrup, and M. Mørup. Think global, adapt local: Learning locally adaptive k-nearest neighbor kernel density estimators. PMLR, 238:4114–4122, 2024. url: proceedings.mlr.press/v238/ olsen24a.html. 9

    2024

  35. [43]

    Pavanatto, A

    L. Pavanatto, A. Giovannelli, B. Giera, T. Bremer, H. Miao, and D. A. Bowman. Exploring multiscale navigation of homogeneous and dense objects with progressive refinement in virtual reality. InProc. VR, pp. 228–237. IEEE Comp. Soc., Los Alamitos, 2025. doi: 10/p5fw 2

    2025

  36. [44]

    M.-T. Pelz, M. Schartau, C. J. Somes, V . Lampe, and T. Slawig. A diffusion-based kernel density estimator (diffKDE, version 1) with optimal bandwidth approximation for the analysis of data in geoscience and ecological research.GeoSci Model Dev, 16(22):6609–6634, 2023. doi: 10/p5f3 3

    2023

  37. [45]

    Pfeifer, N

    S. Pfeifer, N. I. Libeskind, Y . Hoffman, W. A. Hellwing, M. Bilicki, and K. Naidoo. COWS: A filament finder for hessian cosmic web identifiers.Mon Not R Astron Soc, 514(1):470–479, 2022. doi: 10/p5fx 1

    2022

  38. [46]

    Pivovar, J

    J. Pivovar, J. DeGuzman, and E. S. Rosenberg. Virtual reality on a SWIM: Scalable world in miniature. InProc. VRW, pp. 912–913. IEEE Comp. Soc., Los Alamitos, 2022. doi: 10/gt2485 2

    2022

  39. [47]

    Prouzeau, M

    A. Prouzeau, M. Cordeil, C. Robin, B. Ens, B. H. Thomas, and T. Dwyer. Scaptics and highlight-planes: Immersive interaction tech- niques for finding occluded features in 3D scatterplots. InProc. CHI, art. no. 325, 12 pages. ACM, New York, 2019. doi: 10/gnkk76 3

    2019

  40. [48]

    Saito, M

    K. Saito, M. Yano, H. Hino, T. Shoji, A. Asahara, H. Morita et al. Accelerating small-angle scattering experiments on anisotropic samples using kernel density estimation.Sci Rep, 9(1), art. no. 1526, 10 pages,

  41. [49]

    D. W. Scott.Multivariate Density Estimation: Theory, Practice, and Visualization. John Wiley & Sons, Hoboken, 2 nd ed., 2015. doi: 10/ p5f5 3

    2015

  42. [50]

    Shivashankar, P

    N. Shivashankar, P. Pranav, V . Natarajan, R. van de Weygaert, E. P. Bos, and S. Rieder. Felix: A topology based framework for visual explo- ration of cosmic filaments.IEEE Trans Vis Comput Graph, 22(6):1745– 1759, 2016. doi: 10/f8q5rz 1

    2016

  43. [51]

    B. W. Silverman.Density Estimation for Statistics and Data Analysis. Chapman and Hall/CRC, New York, 1986. doi: 10/gqbn6x 3

    1986

  44. [52]

    Song and M

    D. Song and M. Norman. Looking in, looking out: Exploring multiscale data with virtual reality.IEEE Comput Sci Eng, 1(3):53–64, 1994. doi: 10/csw826 2

    1994

  45. [53]

    Sousa Santos, P

    B. Sousa Santos, P. Dias, A. Pimentel, J.-W. Baggerman, C. Ferreira, S. Silva et al. Head-mounted display versus desktop for 3D navigation in virtual reality: A user study.Multimedia Tools Appl, 41(1):161–181,

  46. [54]

    Springel, J

    V . Springel, J. Wang, M. V ogelsberger, A. Ludlow, A. Jenkins, A. Helmi et al. The aquarius project: The subhaloes of galactic haloes. Mon Not R Astron Soc, 391(4):1685–1711, 2008. doi: 10/fsjgzw 1, 5, 7, 12

    2008

  47. [55]

    Springel, S

    V . Springel, S. D. M. White, A. Jenkins, C. S. Frenk, N. Yoshida, L. Gao et al. Simulations of the formation, evolution and clustering of galaxies and quasars.Nature, 435:629–636, 2005. doi: 10/c3cmxr 1, 2, 5

    2005

  48. [56]

    Stoakley, M

    R. Stoakley, M. J. Conway, and R. Pausch. Virtual reality on a WIM: Interactive worlds in miniature. InProc. CHI, pp. 265–272. ACM, New York, 1995. doi: 10/bx762s 2

    1995

  49. [57]

    G. R. VandenBos, ed.Publication Manual of the American Psycholog- ical Association. APA, Washington, DC, 6th ed., 2009. url: apastyle .org/manual. 7

    2009

  50. [58]

    Ware and D

    C. Ware and D. Fleet. Context sensitive flying interface. InProc. I3D, pp. 127–130. ACM, New York, 1997. doi: 10/drbfvq 2

    1997

  51. [59]

    Weissker, M

    T. Weissker, M. Franzgrote, and T. Kuhlen. Try this for size: Multi- scale teleportation in immersive virtual reality.IEEE Trans Vis Comput Graph, 30(5):2298–2308, 2024. doi: 10/p5ft 2

    2024

  52. [60]

    Whitlock, S

    M. Whitlock, S. Smart, and D. A. Szafir. Graphical perception for immersive analytics. InProc. VR, pp. 616–625. IEEE Comp. Soc., Los Alamitos, 2020. doi: 10/gktm78 1

    2020

  53. [61]

    M. H. F. Wilkinson and B. C. Meijer. DATAPLOT: A graphical display package for bacterial morphometry and fluorimetry data.Comput Methods Programs Biomed, 47(1):35–49, 1995. doi: 10/dnr53p 3

    1995

  54. [62]

    this is interesting

    G. J. Wills. Selection: 524,288 ways to say “this is interesting”. In Proc. InfoVis, pp. 54–60. IEEE Comp. Soc., Los Alamitos, 1996. doi: 10/bjq6dh 6

    1996

  55. [63]

    C. A. Wingrave, Y . Haciahmetoglu, and D. A. Bowman. Overcoming world in miniature limitations by a scaled and scrolling WIM. In Proc. 3DUI, pp. 11–16. IEEE Comp. Soc., Los Alamitos, 2006. doi: 10/bbp9k4 2

    2006

  56. [64]

    Wyvill, C

    G. Wyvill, C. McPheeters, and B. Wyvill. Data structure for soft objects.Vis Comput, 2(4):227–234, 1986. doi: 10/dndmwc 6

    1986

  57. [65]

    Xie and J

    Z. Xie and J. Yan. Kernel density estimation of traffic accidents in a network space.Comput Env Urban Syst, 32(5):396–406, 2008. doi: 10/b7gxpq 3

    2008

  58. [66]

    C. Yang, Y . Li, C. Liu, and X. Yuan. Deep learning-based viewpoint recommendation in volume visualization.J Vis, 22(5):991–1003, 2019. doi: 10/p5p4 7

    2019

  59. [67]

    Y . Yang, M. Cordeil, J. Beyer, T. Dwyer, K. Marriott, and H. Pfis- ter. Embodied navigation in immersive abstract data visualization: Is overview+detail or zooming better for 3D scatterplots?IEEE Trans Vis Comput Graph, 27(2):1214–1224, 2021. doi: 10/ghgt52 6

    2021

  60. [68]

    L. Yu, K. Efstathiou, P. Isenberg, and T. Isenberg. Efficient structure- aware selection techniques for 3D point cloud visualizations with 2DOF input.IEEE Trans Vis Comput Graph, 18(12):2245–2254, 2012. doi: 10/f4fv9z 1, 3, 7

    2012

  61. [69]

    L. Yu, K. Efstathiou, P. Isenberg, and T. Isenberg. CAST: Effective and efficient user interaction for context-aware selection in 3D particle clouds.IEEE Trans Vis Comput Graph, 22(1):886–895, 2016. doi: 10/kt5n 1, 3, 7

    2016

  62. [70]

    Z. Yuan, M. J. Jarvis, and J. Wang. A flexible method for estimating luminosity functions via kernel density estimation.Astrophys J Suppl Ser, 248(1), art. no. 1, 18 pages, 2020. doi: 10/p5qc 3

    2020

  63. [71]

    Z. Yuan, X. Zhang, J. Wang, X. Cheng, and W. Wang. A flexible method for estimating luminosity functions via kernel density estimation. II. Generalization and python implementation.Astrophys J Suppl Ser, 260(1), art. no. 10, 15 pages, 2022. doi: 10/p5f2 3

    2022

  64. [72]

    Zhang, Q

    G. Zhang, Q. Huang, A.-X. Zhu, and J. H. Keel. Enabling point pattern analysis on spatial big data using cloud computing: Optimizing and accelerating ripley’s k function.Int J Geogr Inf Sci, 30(11):2230–2252,

  65. [73]

    Zhang, A.-X

    G. Zhang, A.-X. Zhu, and Q. Huang. A GPU-accelerated adaptive kernel density estimation approach for efficient point pattern analysis on spatial big data.Int J Geogr Inf Sci, 31(10):2068–2097, 2017. doi: 10/ggfmp5 3, 5

  66. [74]

    L. Zhao, N. Cao, S. He, H.-N. Liang, and L. Yu. L-wim: Collaborative exploration in immersive environments. InISMAR Adjunct Proc., pp. 118–123. IEEE Comp. Soc., Los Alamitos, 2022. doi: 10/gtmh3d 1, 2

    2022

  67. [75]

    L. Zhao, T. Isenberg, F. Xie, H.-N. Liang, and L. Yu. MeTACAST: Target-and context-aware spatial selection in VR.IEEE Trans Vis Comput Graph, 30(1):480–494, 2024. doi: 10/gtnn25 1, 3, 6, 7 ScaleFree: Dynamic KDE for Multiscale Point Cloud Exploration in VR Supplementary Materi...

    2024

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.