Super-Gaussian: Interactive Scene Editing for 3D Gaussian Splatting and NLI-Based Volume Visualization in Virtual Reality

Chaoli Wang; Kaiyuan Tang; Suemin Jeon; Won-Ki Jeong

REVIEW 3 major objections 5 minor 65 references

Super-Gaussian: Interactive Scene Editing for 3D Gaussian Splatting and NLI-Based Volume Visualization in Virtual Reality

T0 review · 3 major / 5 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 Super-Gaussian claims that grouping 3D Gaussian splats into feature-aware clusters makes volumetric region selection in VR more accurate and less effortful than existing 2D and point-based tools.

desk verdict A genuinely useful Gaussian-clustering interaction unit, wrapped in a headline claim that the benchmark evidence doesn't quite support. read the letter →

arxiv 2608.04475 v1 pith:5H4L5KYY submitted 2026-08-05 cs.HC

Suemin Jeon , Kaiyuan Tang , Chaoli Wang , Won-Ki Jeong This is my paper

classification cs.HC

keywords volumevisualization3DGaussiansplattingvirtualrealitynaturallanguageinteractionregion-of-interestselectionrandomwalkpropagationgeodesicclusteringsceneediting

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

Super-Gaussian is a VR volume-visualization framework built on editable 3D Gaussian splatting. It claims that grouping Gaussian primitives into structure-aware clusters, then selecting through a random-walk propagate-and-refine loop, lets users specify complex volumetric regions such as tumors, vascular necks, or cosmic filaments more accurately and with far fewer interactions than existing 2D screen-space or point-by-point VR tools. On top of selection, the system runs speech-driven natural-language commands, so a user-defined region can be labeled, queried, recolored, or made transparent by talking to the scene. If the claim holds, exploring medical and simulation volumes in VR stops being a painstaking selection chore and becomes a fast, semantic, semi-automatic workflow.

What carries the argument

The load-bearing object is the Super-Gaussian cluster: an over-segmentation of the Gaussian splat set computed by extending SLIC superpixels to 3D, with distance $D(i,j)=d_g(i,j)^2/S^2 + m^2 \|\mathbf{f}_i-\mathbf{f}_j\|^2$, where $d_g$ is a geodesic shortest-path distance on a kNN graph over Gaussian centers and $\mathbf{f}_i$ is the 13-dimensional geometric feature vector of mean scale, normal, and principal axes. Selection is then propagated by solving the random-walk graph-Laplacian equation $L_{uu}\boldsymbol{\mu}_u=-L_{ul}\boldsymbol{\mu}_l$ on a cluster-level graph whose edge weights blend geodesic proximity and feature affinity, $w_{uv}=0.9\exp(-(g_{uv}/\sigma_s)^2)+0.1\exp(-\|\tilde{\mathbf{f}}_u-\tilde{\mathbf{f}}_v\|/\sigma_f)^2$. This collapses the selection problem from individual splats to a compact graph, which is what makes region propagation near-instant and the select-and-refine loop practical in VR.

What would settle it

Repeat the vortex task from Table 1 with the geodesic distance in Equation 5 replaced by Euclidean distance while keeping all other settings fixed: if accuracy and completion time do not degrade, the geodesic feature is not what carries the reported gains. Alternatively, on a dataset with weak structural contrast (e.g., noisy medical scans), measure the IoU of Super-Gaussian clusters against manual region annotation; if it falls far below the roughly 0.85 boundary agreement the paper reports for pseudo-mask reconstruction, the structure-alignment premise fails.

Watch

Extended reading notes

Core claim

The paper's central claim is that Super-Gaussian consistently outperforms the baseline methods SuperSplat and GSVR in both selection accuracy and interaction cost. The mechanism is a two-level representation: individual Gaussians are first merged into Super-Gaussians by geodesic SLIC clustering over geometric features (mean scale, normal, and principal axes), and selection is then propagated over a cluster graph by random walks seeded by 3D brush strokes, with cluster-level toggling and point-level refinement available as corrections. Because the graph has at most a few hundred to a thousand nodes per scene, propagation runs in roughly ten milliseconds. Quantitative results on vortex, lobster, and ionization report higher MCC, F1, and IoU than both baselines while reducing completion time and interaction count, and four case studies show the workflow separating and labeling anatomical structures (aneurysm sac, neck, parent vessel), a bonsai's pot and leaves, and the core and filaments of a cosmological simulation.

Load-bearing premise

The most load-bearing premise is that a fixed set of geometric features and clustering hyperparameters (window size, compactness, random-walk sigmas, and the 0.9/0.1 weight balance) will carve every new dataset into structure-aligned groups without per-dataset tuning; if clusters cross anatomical or physical boundaries, the random-walk propagation and refinement steps inherit the error.

Editorial extensions

If this is right

Users can define semantically meaningful regions of interest directly in 3D without pre-segmentation or offline 2D mask lifting, removing what the paper identifies as the main bottleneck of prior Gaussian segmentation tools.
On the reported benchmark, Super-Gaussian reaches higher MCC, F1, and IoU than both SuperSplat and GSVR on all three datasets while lowering completion time and interaction count; the largest gains appear in transparent and sparse regions such as vortex and ionization.
The editable Gaussian renderer keeps average VR frame rates near 120 FPS even for multi-gigabyte volumes, staying above the 90 FPS cybersickness threshold, while the naive ray-marched DVR baseline drops below 5 FPS or runs out of memory on the same scenes.
Natural-language commands can be attached to user-defined selections through on-the-fly CLIP text labels, letting a selected region be referenced, queried, recolored, or made transparent by speech within the same visualization-perception-action loop.

Reading between the lines

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

Because the clustering uses only geometric attributes, selection should remain valid across transfer functions and lighting changes; a direct test would be to re-color or re-light a labeled region under a different transfer function and check that the same cluster selection still matches the target.
The same cluster graph could host a learned seed proposer that guesses foreground and background clusters from the current view, turning the random-walk step from a brush stroke into a single click and removing most of the remaining manual effort.
The paper's latency data suggest text labeling cost is dominated by 2D rasterization of selected Gaussians, so caching per-cluster renderings would cut labeling latency further without changing the CLIP embeddings.

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.

Desk Editor's Note

read the letter

Bottom line: the paper has a genuinely useful idea — grouping 3D Gaussian splats into geodesic SLIC clusters using geometric features, then letting users select by cluster, refine by random walk and point edits — and it is packaged with an unusually honest limitations section. But the headline claim that it “consistently outperforms” existing methods is not supported by the evidence as presented. The selection benchmark is four author-operators, one task per dataset, no variance, no significance, and the ground-truth mask definition is never stated. The paper itself says at the end of Section 6.1 to treat it as a controlled technical comparison, not a usability study — which is the right call, but that should be the framing from the start, not a caveat.

What is actually new is the Super-Gaussian unit itself. The ablation (geodesic vs Euclidean distance) looks convincing qualitatively, and the clustering robustness appendix across 3DGS, 2DGS, Scaffold-GS, and iVR-GS is a nice touch that suggests the idea isn't tied to one representation. The rendering table is also useful: the editable Gaussian representation holds roughly 120 FPS where a naive DVR baseline falls off or runs out of memory on larger volumes.

The soft spots are real but contained. The RW ablation in Appendix C is the kind of thing that should be in the main text: on vortex, their method gets IoU 0.52 while plain Gaussian RW gets 1.0 under the same seeds. That is a boundary condition on the central “structure-aware propagation” claim, and the paper does not discuss it. Hyperparameter sensitivity (S, m, the RW sigmas) is acknowledged in the limitations but not analyzed; for a method whose whole pitch is structure-aligned over-segmentation, showing how cluster boundaries change with those parameters would make the contribution much more robust. The NLI components are largely a re-application of NLI4VolVis; the novelty is in the selection, not the language-model integration.

Verdict: this qualifies for peer review and, I think, eventual publication after the accuracy claim is scoped and the evaluation is either expanded or honestly relabeled. If I were the editor, I would send it out; if I were a reviewer, I would push hard for variance data, a defined ground-truth protocol, and a discussion of the vortex failure mode. The gap between the systems contribution and the evaluation is a teachable example, so I would bring it to a reading group.

Referee Report

3 major / 5 minor

Summary. The paper presents Super-Gaussian, a VR volume visualization framework built on editable 3D Gaussian splatting. It introduces a feature-aware clustering of Gaussian primitives (Super-Gaussian) as higher-level interaction units, a hierarchical select-and-refine workflow combining random-walk propagation, cluster selection, and point refinement, and an NLI multi-agent system for on-the-fly labeling and semantic editing of user-selected regions. The system is evaluated through four case studies (bonsai, aneurysm, Nyx, chameleon), a quantitative selection comparison against SuperSplat and GSVR on three datasets, rendering efficiency measurements against a naive DVR baseline, and ablation studies of the geodesic SLIC clustering and RW propagation. The central claim, stated in Section 6.1, is that Super-Gaussian consistently outperforms the two baselines in both accuracy and interaction cost.

Significance. The work addresses a real gap in VR volume visualization: existing editable 3DGS pipelines rely on offline segmentations or tedious point-by-point selection, and this paper proposes a concrete interaction design that combines structure-aware over-segmentation with hierarchical refinement and NLI. If the quantitative claims are substantiated, the contribution would be valuable to the TVCG community. I credit the authors for extensive supplementary evaluations (RW ablation, clustering robustness across four Gaussian representations, text-labeling latency, unstructured data applicability) and for transparently stating limitations in Section 7. The rendering efficiency measurements (Table 3) show a clear advantage over a naive DVR baseline. However, the central quantitative comparison currently rests on four author-operators with no variance or significance analysis, which prevents the paper from establishing its headline claim in its current form.

major comments (3)

[Section 6.1, Table 1, Fig. 8] The claim that 'Super-Gaussian consistently outperforms baseline methods in both accuracy and interaction cost' is not supported by the evidence as reported. Table 1 contains only mean values from four trained author-operators, with no error bars, per-trial variance, or significance tests, and the protocol states that operators saw the target region highlighted in green (Fig. 8) before the recorded trials. The ground-truth target definition itself is not specified, so it is unclear whether the highlighted masks were independently defined or derived from the method being evaluated. These issues make it impossible to distinguish genuine method effects from operator familiarity or experimenter bias; for example, the vortex completion-time difference (21.54 s vs 138.30 s) could reflect operator practice rather than the interaction design. I recommend adding inferential statistics, reporting per-operator/trial data, specifying and ideally externalizing the ground-truth definition, and either recruiting independent operators or explicitly scoping the claim to a controlled technical demonstration.
[Section 4.2.1, Eq. (5), and Section 7] The Super-Gaussian construction and RW propagation depend on several hand-set hyperparameters (voxel cell size S, compactness m in Eq. (5); sigma_s, sigma_f, and the 0.9/0.1 balance in Eq. (7); and the 0.5 threshold in Section 4.2.2). The paper itself states in Section 7 that 'cluster size and boundary quality can vary significantly across datasets' and that S 'directly control[s] the clustering granularity.' However, the quantitative evaluation in Table 1 does not report the hyperparameter values used or include a sensitivity analysis, so the reader cannot assess whether the reported selection accuracy holds across datasets or only for a favorable parameter choice. Since the central claim is that the method is consistently superior across datasets, this is a load-bearing gap. I recommend reporting the chosen values, adding a sensitivity study (e.g., varying S and m), or narrowing the claim to the demonstrated settings.
[Section 6.1, Results] The paragraph following Table 1 contains an internal tension: it states that 'for datasets with clearly defined boundaries (e.g., vortex and ionization), SuperSplat and GSVR can sometimes achieve comparable or slightly higher accuracy,' while the same section's opening asserts consistent superiority in accuracy. With no statistical tests, the reader cannot determine whether the small differences in metrics such as precision (e.g., 0.9751 vs 0.9766 for vortex) are meaningful. The claim should be aligned with the reported numbers and either qualified (e.g., 'higher overall accuracy and lower interaction cost on these datasets') or backed by significance testing.

minor comments (5)

[Appendix C, Table 1] The 'cluster RW (Euclidean)' rows for chameleon and lobster report identical accuracy metrics (IoU 0.5538, accuracy 0.9854, F1 0.5538, precision 1.0000) despite different datasets and runtimes; please verify whether these are copy-paste errors.
[Section 4.2.1, Eq. (5)] The role of compactness m is described as balancing 'spatial proximity and feature similarity,' but in the formula m^2 multiplies the feature term, so increasing m increases the weight of feature distance; please clarify the intended interpretation.
[Section 6.1] The definition of a 'selection action' is ambiguous for polygon selection in SuperSplat, which requires multiple clicks; please specify whether a polygon is one interaction or one per click.
[Section 4.1 vs 4.2.1] The term 'Super-Gaussian' is used for both the overall framework (Section 4.1) and the clustering unit (Section 4.2.1); consider using distinct names to avoid confusion.
[Fig. 4 caption] The caption mentions projecting feature vectors onto the first PCA axis, but the PCA procedure is not described in Section 4.2.1; please add a brief explanation or remove the reference.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the system is built from externally published components and evaluated against external baselines; the core selection workflow is not derived from its own outputs.

full rationale

The paper's derivation chain is empirical and self-contained rather than circular. Super-Gaussian clustering (Section 4.2.1) extends published SLIC superpixel methodology to Gaussian primitives using geometric features and a geodesic kNN graph; the equations define the clustering procedure without presupposing the selection results it later supports. The random-walk propagation (Section 4.2.2) applies a standard graph-based algorithm, and the NLI labeling (Section 4.3) reuses the published NLI4VolVis embedding strategy as a component, not as proof of the paper's central claim. The quantitative claim that Super-Gaussian 'consistently outperforms baseline methods' is supported by Table 1, an empirical measurement against external baselines (SuperSplat and GSVR); no fitted parameter is renamed as a prediction, and no equation computes the evaluation metrics from the method's own outputs. The self-citations to iVR-GS, NLI4VolVis, and TexGS-VolVis provide the base editable-Gaussian representation and NLI machinery, but these are peer-reviewed published building blocks with independent content, and the paper's contribution is an extension rather than a derivation from an unverified self-citation. Concerns about the four author-operators, lack of variance reporting, or unspecified ground-truth definition are evaluation-validity issues, not circularity: the paper itself labels the comparison a 'controlled technical comparison' rather than a usability evaluation. No step in the paper reduces, by construction or by citation chain, to its own inputs.

Assumptions & free parameters 6 free parameters · 7 assumptions · 1 invented entities

The central claim depends on a modest set of hand-set clustering and propagation parameters that are not given values and are not ablated for sensitivity. No new physical entities are posited; the Super-Gaussian is an algorithmic grouping. The main untested premises are the structure-alignment of the geometric feature set and the transferability of CLIP to volume-rendered imagery.

free parameters (6)

Compactness factor m (Eq. 5)
Balances spatial proximity and feature similarity in the SLIC-style clustering. No value is reported and no sensitivity analysis is provided.
Voxel cell size S (Eq. 5)
Sets the seed grid density and therefore the number of super-Gaussians. Section 7 states it directly controls clustering granularity.
RW edge weight spatial sigma (sigma_s, Eq. 7)
Controls sensitivity of geodesic distance in the random-walk graph. Value is not reported.
RW edge weight feature sigma (sigma_f, Eq. 7)
Controls sensitivity of feature affinity in the random-walk graph. Value is not reported.
RW edge weight balance 0.9/0.1 (Eq. 7) = 0.9 and 0.1
Fixed hand-set relative weight of geodesic versus feature terms in the edge weight. No ablation on this ratio is given.
RW probability threshold 0.5 (Section 4.2.2) = 0.5
Converts random-walk probabilities into binary selection. Chosen without discussion.

assumptions (7)

standard math SLIC superpixel over-segmentation is a valid basis for region grouping (used as the model for Super-Gaussian, Section 4.2.1).
The algorithm is standard and cited as [1]; its extension to 3D Gaussian centers is the paper's contribution.
standard math Shortest-path distances on a kNN graph approximate geodesic distances on the underlying point cloud (Eq. 6).
This is a standard manifold-learning assumption, cited via [21,48].
standard math The random-walk graph Laplacian solution gives meaningful foreground probabilities for unlabeled nodes (Eq. 8).
Standard RW formulation from [34]; the paper applies it at the cluster level.
domain assumption Editable Gaussian attributes from iVR-GS/TexGS faithfully represent volume geometry and appearance (Sections 3 and 4).
The framework inherits the trained Gaussian representation and its editability from prior published systems [2,45,47] without re-validating reconstruction fidelity.
domain assumption Geometric Gaussian attributes (mean scale, normal, principal axes) are stable, structure-aligned descriptors, while color and opacity are not (Section 4.2.1 and Appendix Fig. 2).
The paper provides only qualitative visualization evidence for this claim; it is load-bearing for clustering quality.
domain assumption CLIP embeddings computed on volume-rendered images retain semantic alignment for scientific and medical content (Sections 4.3 and 5.4).
The paper acknowledges in Section 7 that CLIP is trained on natural images and may degrade on volume-rendered images with domain-specific characteristics.
ad hoc to paper Hand-set clustering and RW hyperparameters generalize across datasets (Sections 4.2.1 and 4.2.2).
No sensitivity analysis is provided, and Section 7 concedes that cluster size and boundary quality vary significantly across datasets.

invented entities (1)

Super-Gaussian cluster unit independent evidence
purpose: Higher-level interaction unit grouping Gaussian primitives by geometric structure, enabling brush-based selection, RW propagation, and semantic labeling without pre-segmentation.
The Super-Gaussian is an algorithmic construct, not a physical entity. It is directly evaluated through selection benchmarks, ablations, and case studies, so it carries no explanatory burden of the graviton type.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Super-Gaussian: Interactive Scene Editing for 3D Gaussian Splatting and NLI-Based Volume Visualization in Virtual Reality." pith.science (2026). https://pith.science/paper/5H4L5KYY

@misc{pith2026260804475,
  author       = {Pith},
  title        = {Pith review of: Super-Gaussian: Interactive Scene Editing for 3D Gaussian Splatting and NLI-Based Volume Visualization in Virtual Reality},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5H4L5KYY}},
  note         = {Machine review of arXiv:2608.04475}
}

read the original abstract

Despite the promise of virtual reality (VR) for intuitive spatial interaction, volume visualization (VolVis) in VR remains constrained by high rendering costs and motion discomfort. Recent advances have shown that representing volumetric scenes with 3D Gaussian splatting enables high-performance rendering, making this representation well-suited for VR. However, existing Gaussian-based scene editing workflows remain limited by slow offline segmentation and fatigue-inducing manual selection. To address these challenges, we present Super-Gaussian, a novel VolVis framework that enhances scene editing and interaction in VR through intuitive 3D Gaussian selection and natural language interaction (NLI). Our approach groups Gaussian primitives into higher-level units via feature-aware clustering, enabling efficient selection of complex volumetric regions, such as tumors in medical images or filaments in cosmological data, without point-by-point interaction. Building on this, we introduce a hierarchical select-and-refine workflow that combines random-walk-based region propagation, cluster selection, and point refinement, allowing users to progressively specify regions of interest with reduced effort. We further support on-the-fly text labeling of selected regions using NLI, allowing users to semantically query, interpret, and manipulate content within a visualization-perception-action loop. By integrating multimodal interaction, including speech, visual feedback, and spatial manipulation in VR, our framework supports intuitive exploration, editing, and scientific analysis of volumetric data. We demonstrate the effectiveness of Super-Gaussian through four case studies, quantitative selection benchmarks against existing Gaussian-based techniques, and system-level evaluations. Implementation details and experiments can be found on the project page: https://smin0136.github.io/super-gaussian-project/

Figures

Figures reproduced from arXiv: 2608.04475 by the authors.

**Figure 2.** Comparison of Gaussian selection in semi-transparent volumetric [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 3.** Block diagram of the Super-Gaussian clustering pipeline. (a) (b) (c) [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Figures from the paper (10 more)

**Figure 4.** Figure 4: Super-Gaussian for (a) aneurysm, (b) beetle, and (c) ionization, color-mapped from cluster-level features by projecting each feature vector onto the first PCA axis. structural transitions, whereas appearance-related features (e.g., offset color, opacity) show minimal …

**Figure 5.** Figure 5: The select-and-refine workflow with text labeling example of (a) [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]

**Figure 6.** Figure 6: Case study with the aneurysm and Nyx datasets demonstrating the use of NLI with user-defined text labels. Each schematic VR panel summarizes the text labels and colors of editable units. 5.3 Nyx For the Nyx dataset, we demonstrate how users can explore and interpret co…

**Figure 8.** Figure 8: Datasets used in the quantitative selection comparison: (a) [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]

**Figure 9.** Figure 9: Ablation on distance formulation in SLIC-based Super-Gaussian [PITH_FULL_IMAGE:figures/full_fig_p009_9.png]

**Figure 2.** Figure 2: Gaussian attributes of the chameleon dataset color-mapped onto the 3D scene to illustrate its spatial distribution, including anisotropy, opacity, scale, offset color, normal directions, and principal axes [PITH_FULL_IMAGE:figures/full_fig_p012_2.png]

**Figure 1.** Figure 1: User interface of Super-Gaussian in the VR environment. (a) [PITH_FULL_IMAGE:figures/full_fig_p012_1.png]

**Figure 3.** Figure 3: Datasets used in the RW ablation: (a) chameleon, (b) lobster, and (c) vortex. The target selected is highlighted with a green dotted line. D CLUSTERING ROBUSTNESS To evaluate the robustness of our clustering algorithm across different Gaussian scene representations, w…

**Figure 4.** Figure 4: Datasets used in the clustering robustness experiments: (a) [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]

**Figure 5.** Figure 5: presents the resulting editable Gaussian representations together with the corresponding Super-Gaussian clustering results. Despite being trained on unstructured tetrahedral meshes rather than structured volumes, our method successfully produces spatially coherent s…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

65 extracted references · 28 canonical work pages

[1]

Achanta, A

R. Achanta, A. Shaji, K. Smith, A. Lucchi, P. Fua, and S. Süsstrunk. SLIC superpixels compared to state-of-the-art superpixel methods.IEEE Transactions on Pattern Analysis and Machine Intelligence, 34(11):2274– 2282, 2012. doi: 10.1109/TPAMI.2012.120 4

doi:10.1109/tpami.2012.120 2012
[2]

K. Ai, K. Tang, and C. Wang. NLI4V olVis: Natural language interaction for volume visualization via LLM multi-agents and editable 3D Gaussian splatting.IEEE Transactions on Visualization and Computer Graphics, 32(1):46–56, 2026. doi: 10.1109/TVCG.2025.3633888 2, 3, 5, 6, 9

arXiv 2026
[3]

Bauer, Q

D. Bauer, Q. Wu, H. Gadirov, and K.-L. Ma. GSCache: Real-time radiance caching for volume path tracing using 3D Gaussian splatting.IEEE Transactions on Visualization and Computer Graphics, 32(1):901–911,

work page
[4]

J. F. Blinn. Models of light reflection for computer synthesized pictures. InProceedings of ACM SIGGRAPH Conference, pp. 192–198, 1977. doi: 10.1145/965141.563893 3

work page arXiv 1977
[5]

Boorboor, M

S. Boorboor, M. S. Castellana, Y . Kim, C. Zhu-Tian, J. Beyer, H. Pfister et al. V oxAR: Adaptive visualization of volume rendered objects in optical see-through augmented reality.IEEE Transactions on Visualization and Computer Graphics, 30(10):6801–6812, 2023. doi: 10.1109/TVCG.2023. 3340770 1, 2

doi:10.1109/tvcg.2023 2023
[6]

J. Cen, J. Fang, C. Yang, L. Xie, X. Zhang, W. Shen et al. Segment any 3D Gaussians. InProceedings of AAAI Conference on Artificial Intelligence, pp. 1971–1979, 2025. doi: 10.1609/AAAI.V39I2.32193 2, 3

work page doi:10.1609/aaai.v39i2.32193 1971
[8]

Y . J. Cho and H. Y . Yoon. Numerical analysis of the rocom boron dilution benchmark experiment using the cupid code.Nuclear Engineering and Design, 341:167–175, 2019. doi: 10.1016/j.nucengdes.2018.11.005 14

work page doi:10.1016/j.nucengdes.2018.11.005 2019
[9]

J. Choi, H. Kim, H. An, and Y . Jung. Real-time transfer function editor for direct volume rendering in mixed reality. InProceedings of ACM SIGGRAPH Asia Posters, pp. 38:1–38:2, 2024. doi: 10.1145/3681756. 3697953 2

work page doi:10.1145/3681756 2024

Show all 65 references

[10]

Dyken, A

L. Dyken, A. Sewell, W. Usher, N. Debardeleben, S. Petruzza, and S. Ku- mar. V olume Encoding Gaussians: Transfer function-agnostic 3D Gaus- sians for volume rendering.arXiv preprint arXiv:2504.13339, 2025. doi: 10.48550/arXiv.2504.13339 2

2025 doi
[11]

Goswami and O

A. Goswami and O. Sharma. HoloView: An immersive mixed- reality visualization system for anatomical education.arXiv preprint arXiv:2501.08736, 2025. doi: 10.48550/arXiv.2501.08736 1, 2

2025 doi arXiv
[12]

M. Han, A. Sewell, J. Insley, J. Knowles, V . A. Mateevitsi, M. E. Papka et al. Toward distributed 3D Gaussian splatting for high-resolution isosur- face visualization. InProceedings of IEEE International Conference on eScience, pp. 331–332, 2025. doi: 10.1109/eScience65000.20...

2025
[13]

Hrycak and J

C. Hrycak and J. Krüger. An investigation of the Apple Vision Pro for out-of-core ray-guided volume rendering with BorgVR. InProceedings of Vision, Modeling, and Visualization, 2025. doi: 10.2312/vmv.20251236 2

2025 doi
[14]

Hrycak, D

C. Hrycak, D. Lewakis, and J. Krüger. Investigating the Apple Vision Pro spatial computing platform for GPU-based volume visualization. In Proceedings of IEEE VIS Conference (Short Papers), pp. 181–185, 2024. doi: 10.1109/VIS55277.2024.00044 2

2024
[15]

Huang, Z

B. Huang, Z. Yu, A. Chen, A. Geiger, and S. Gao. 2D Gaussian splat- ting for geometrically accurate radiance fields. InProceedings of ACM SIGGRAPH Conference, pp. 32:1–32:11, 2024. doi: 10.1145/3641519. 3657428 12

2024 doi
[16]

Jadhav and A

S. Jadhav and A. E. Kaufman. MD-Cave: An immersive visualization workbench for radiologists.IEEE Transactions on Visualization and Computer Graphics, 29(12):4832–4844, 2022. doi: 10.1109/TVCG.2022. 3193672 1, 2

2022 doi
[17]

U. Jain, A. Mirzaei, and I. Gilitschenski. GaussianCut: Interactive seg- mentation via graph cut for 3D Gaussian splatting. InProceedings of Advances in Neural Information Processing Systems, pp. 89184–89212,
[18]

Jiang, C

Y . Jiang, C. Yu, T. Xie, X. Li, Y . Feng, H. Wang et al. VR-GS: A physical dynamics-aware interactive Gaussian splatting system in virtual reality. In Proceedings of ACM SIGGRAPH Conference, pp. 78:1–78:12, 2024. doi: 10.1145/3641519.3657448 2, 3

2024
[19]

Kerbl, G

B. Kerbl, G. Kopanas, T. Leimkühler, and G. Drettakis. 3D Gaussian splatting for real-time radiance field rendering.ACM Transactions on Graphics, 42(4):139:1–139:14, 2023. doi: 10.1145/3592433 2, 3, 4, 12

2023 doi
[20]

Kirillov, E

A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson et al. Segment anything. InProceedings of IEEE/CVF International Conference on Computer Vision, pp. 3992–4003, 2023. doi: 10.1109/ICCV51070. 2023.00371 3

2023
[21]

Klein and G

J. Klein and G. Zachmann. Proximity graphs for defining surfaces over point clouds. InProceedings of Eurographics Symposium on Point-Based Graphics, pp. 131–138, 2004. doi: 10.2312/SPBG/SPBG04/131-138 4

2004 doi
[22]

Kleinbeck, H

C. Kleinbeck, H. Schieber, K. Engel, R. Gutjahr, and D. Roth. Multi- layer Gaussian splatting for immersive anatomy visualization.IEEE Transactions on Visualization and Computer Graphics, 31(1):2353–2363,
[23]

M. Lavik. Unity V olume Rendering.https://github.com/mlavik1/ UnityVolumeRendering, 2024. Accessed: 2026-04-06. 8

2024
[24]

J. J. LaViola Jr. A discussion of cybersickness in virtual environments. ACM SIGCHI Bulletin, 32(1):47–56, 2000. doi: 10.1145/333329.333344 2, 8

2000
[25]

Q. Li, Y . Wang, Y . Cai, H. Du, and L. Liu. Variable basis mapping for real-time volumetric visualization.arXiv preprint arXiv:2601.09417, 2026. doi: 10.48550/arXiv.2601.09417 2

2026 doi
[26]

K. Liu, F. Zhan, M. Xu, C. Theobalt, L. Shao, and S. Lu. StyleGaussian: Instant 3D style transfer with Gaussian splatting. InProceedings of ACM SIGGRAPH Asia Technical Communications, pp. 21:1–21:4, 2024. doi: 10.1145/3681758.3698002 3

2024
[27]

Q. Liu, S. Qiu, Y . Wang, X. Wu, K. S. H. Chok, C.-W. Fu et al. Coordinated 2D-3D visualization of volumetric medical data in XR with multimodal interactions. InProceedings of IEEE VIS Conference (Short Papers), pp. 331–335, 2025. doi: 10.1109/VIS60296.2025.00072 2

2025
[28]

S. Liu, H. Miao, Z. Li, M. L. Olson, V . Pascucci, and P.-T. Bremer. A V A: Towards autonomous visualization agents through visual perception-driven decision-making.Computer Graphics Forum, 43(3):e15093, 2024. doi: 10.1111/cgf.15093 3

2024 doi
[29]

Z. Liu, H. Ouyang, Q. Wang, K. L. Cheng, J. Xiao, K. Zhu et al. InFusion: Inpainting 3D Gaussians via learning depth completion from diffusion prior.arXiv preprint arXiv:2404.11613, 2024. doi: 10.48550/arXiv.2404. 11613 2, 3

2024 doi arXiv
[31]

X. Meng, R. Du, and A. Varshney. Eye-dominance-guided foveated rendering.IEEE Transactions on Visualization and Computer Graphics, 26(5):1972–1980, 2020. doi: 10.1109/TVCG.2020.2973442 2

1972
[32]

Mirzaei, R

A. Mirzaei, R. De Lutio, S. W. Kim, D. Acuna, J. Kelly, S. Fidler et al. RefFusion: Reference adapted diffusion models for 3D scene inpainting. arXiv preprint arXiv:2404.10765, 2024. doi: 10.48550/arXiv.2404.10765 2, 3

2024 doi arXiv
[33]

Niedermayr, C

S. Niedermayr, C. Neuhauser, K. Petkov, K. Engel, and R. Westermann. Application of 3D Gaussian splatting for cinematic anatomy on consumer class devices. InProceedings of Vision, Modeling, and Visualization, pp. 1–13, 2024. doi: 10.2312/vmv.20241195 2

2024 doi
[34]

K. Pearson. The problem of the random walk.Nature, 72(1867):342–342,
[35]

Pidhorskyi, M

S. Pidhorskyi, M. Morehead, Q. Jones, G. Spirou, and G. Doretto. syGlass: Interactive exploration of multidimensional images using virtual reality head-mounted displays.arXiv preprint arXiv:1804.08197, 2018. doi: 10. 48550/arXiv.1804.08197 1, 2

2018 doi arXiv
[36]

SuperSplat v2.23.2

PlayCanvas. SuperSplat v2.23.2. https://playcanvas.com/ supersplat/editor, 2024. Accessed: 2026-03-09. 2, 3, 7 10 © 2026 IEEE. This is the author’s version of the article that has been published in IEEE Transactions on Visualization and Computer Graphics. The final version of ...

2024
[37]

Pranckeviˇcius

A. Pranckeviˇcius. Unity Gaussian Splatting. https://github.com/ aras-p/UnityGaussianSplatting, 2025. Accessed: 2026-02-01. 3

2025
[38]

M. Qin, W. Li, J. Zhou, H. Wang, and H. Pfister. LangSplat: 3D language Gaussian splatting. InProceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 20051–20060, 2024. doi: 10.1109/ CVPR52733.2024.01895 3

2024
[39]

Radford, J

A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal et al. Learning transferable visual models from natural language supervision. InProceedings of International Conference on Machine Learning, pp. 8748–8763, 2021. doi: 10.48550/arXiv.2103.00020 3

2021 doi arXiv
[40]

N. Ravi, V . Gabeur, Y .-T. Hu, R. Hu, C. Ryali, T. Ma et al. SAM 2: Segment anything in images and videos. InProceedings of International Conference on Learning Representations, pp. 28085–28128, 2025. doi: 10 .48550/arXiv.2408.00714 13

2025 arXiv
[41]

Scholl, S

I. Scholl, S. Suder, and S. Schiffer. Direct volume rendering in virtual real- ity. InProceedings of Bildverarbeitung für die Medizin 2018: Algorithmen- Systeme-Anwendungen. Proceedings des Workshops vom 11. bis 13. März 2018 in Erlangen, pp. 297–302, 2018. doi: 10.1007/978-3-...

2018 doi
[42]

Y . Shen, B. Li, J. Huang, D. Yip, and Z. Wang. GaussianShopVR: Fa- cilitating immersive 3D authoring using Gaussian splatting in VR. In Proceedings of ACM Symposium on User Interface Software and Tech- nology, pp. 188:1–188:14, 2025. doi: 10.1145/3746059.3747803 2, 3, 7

2025
[43]

Shetty, A

N. Shetty, A. Chaudhary, D. Coming, W. R. Sherman, P. O’Leary, E. T. Whiting et al. Immersive ParaView: A community-based, immersive, universal scientific visualization application. InProceedings of IEEE Virtual Reality Conference, pp. 239–240, 2011. doi: 10.1109/VR.2011. 5759487 1, 2

2011 doi
[44]

Taibo and J

J. Taibo and J. A. Iglesias-Guitian. Immersive 3D medical visualization in virtual reality using stereoscopic volumetric path tracing. InProceedings of IEEE Virtual Reality Conference, pp. 1044–1053, 2024. doi: 10.1109/ VR58804.2024.00123 1, 2

2024
[45]

K. Tang, K. Ai, J. Han, and C. Wang. TexGS-V olVis: Expressive scene editing for volume visualization via textured Gaussian splatting.IEEE Transactions on Visualization and Computer Graphics, 32(1):933–943,
[46]

Tang and C

K. Tang and C. Wang. ECoNGS: Efficient compressive neural Gaussian splats for volume visualization.IEEE Transactions on Visualization and Computer Graphics, 33(1), 2027. Accepted. 2

2027
[47]

K. Tang, S. Yao, and C. Wang. iVR-GS: Inverse volume rendering for explorable visualization via editable 3D Gaussian splatting.IEEE Trans- actions on Visualization and Computer Graphics, 31(6):3783–3795, 2025. doi: 10.1109/TVCG.2025.3567121 2, 3, 4, 12

2025
[48]

J. B. Tenenbaum, V . d. Silva, and J. C. Langford. A global geometric frame- work for nonlinear dimensionality reduction.Science, 290(5500):2319– 2323, 2000. doi: 10.1126/science.290.5500.2319 4
[49]

C. F. Tsang, A. Hu, O. Perel, C. Kolve, and M. Shugrina. ArtisanGS: Interactive tools for Gaussian splat selection with AI and human in the loop.arXiv preprint arXiv:2602.10173, 2026. doi: 10.48550/arXiv.2602. 10173 3

2026 doi
[50]

doi: 10.1109/TVCG.2025.3634643 2, 3

2025
[51]

S. Wang, D. Zhu, H. Yu, and Y . Wu. Immersive WYSIWYG (what you see is what you get) volume visualization. InProceedings of IEEE Pacific Vi- sualization Symposium, pp. 166–170, 2020. doi: 10.1109/PacificVis48177 .2020.1001 2

2020
[52]

Y . Wang, Q. Wu, G. Zhang, and D. Xu. Learning 3D geometry and feature consistent Gaussian splatting for object removal. InProceedings of European Conference on Computer Vision, pp. 1–17, 2024. doi: 10. 1007/978-3-031-72646-0_1 3

2024
[53]

Waschk and J

A. Waschk and J. Krüger. FA VR-accelerating direct volume rendering for virtual reality systems. InProceedings of IEEE VIS Conference (Short Papers), pp. 106–110, 2020. doi: 10.1109/VIS47514.2020.00028 2

2020
[54]

C. Xu, H. Cheng, Z. Chen, J. Wang, Y . Chen, and L. Zhao. Real-time realistic volume rendering of consistently high quality with dynamic illu- mination.IEEE Transactions on Visualization and Computer Graphics, 31(9):5288–5303, 2024. doi: 10.1109/TVCG.2024.3445339 1, 2

2024
[55]

Usher, P

W. Usher, P. Klacansky, F. Federer, P.-T. Bremer, A. Knoll, J. Yarch et al. A virtual reality visualization tool for neuron tracing.IEEE Transactions on Visualization and Computer Graphics, 24(1):994–1003, 2018. doi: 10. 1109/TVCG.2017.2744079 2

2018
[56]

M. Ye, M. Danelljan, F. Yu, and L. Ke. Gaussian Grouping: Segment and edit anything in 3D scenes. InProceedings of European Conference on Computer Vision, pp. 162–179, 2024. doi: 10.1007/978-3-031-73397 -0_10 3

2024 doi
[57]

L. Yu, K. Efstathiou, P. Isenberg, and T. Isenberg. Efficient structure-aware selection techniques for 3D point cloud visualizations with 2DOF input. IEEE Transactions on Visualization and Computer Graphics, 18(12):2245– 2254, 2012. doi: 10.1109/TVCG.2012.217 8

2012 doi
[58]

L. Yu, K. Efstathiou, P. Isenberg, and T. Isenberg. CAST: Effective and efficient user interaction for context-aware selection in 3D particle clouds. IEEE Transactions on Visualization and Computer Graphics, 22(1):886– 895, 2015. doi: 10.1109/TVCG.2015.2467202 8

2015
[59]

Z. Yuan, S. He, Y . Liu, and L. Yu. MEinVR: Multimodal interaction techniques in immersive exploration.Visual Informatics, 7(3):37–48,
[60]

Yao and C

S. Yao and C. Wang. V olSegGS: Segmentation and tracking in dynamic volumetric scenes via deformable 3D Gaussians.IEEE Transactions on Visualization and Computer Graphics, 32(1):407–417, 2026. doi: 10. 1109/TVCG.2025.3642516 2

2026
[61]

Zhuang, D

J. Zhuang, D. Kang, Y .-P. Cao, G. Li, L. Lin, and Y . Shan. TiP-Editor: An accurate 3D editor following both text-prompts and image-prompts. ACM Transactions on Graphics, 43(4):121:1–121:12, 2024. doi: 10.1145/ 3658205 2, 3

2024
[62]

Zwicker, H

M. Zwicker, H. Pfister, J. van Baar, and M. Gross. EW A volume splatting. InProceedings of IEEE Visualization Conference, pp. 29–538, 2001. doi: 10.1109/VISUAL.2001.964490 3 11 A USERINTERFACE OFSUPER-GAUSSIAN Super-Gaussian provides three main user interfaces. Two of them are...

2001
[66]

L. Zhao, T. Isenberg, F. Xie, H.-N. Liang, and L. Yu. MeTACAST: Target-and context-aware spatial selection in VR.IEEE Transactions on Visualization and Computer Graphics, 30(1):480–494, 2023. doi: 10. 1109/TVCG.2023.3326517 8

2023
[1905]

doi: 10.1038/072294b0 5

doi
[2023]

doi: 10.1016/J.VISINF.2023.06.001 1, 2

2023 doi
[2024]

doi: 10.48550/arXiv.2411.07555 3

doi arXiv
[2025]

doi: 10.1109/TVCG.2025.3549882 2

2025

Pith tools

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

[1] [1]

Achanta, A

R. Achanta, A. Shaji, K. Smith, A. Lucchi, P. Fua, and S. Süsstrunk. SLIC superpixels compared to state-of-the-art superpixel methods.IEEE Transactions on Pattern Analysis and Machine Intelligence, 34(11):2274– 2282, 2012. doi: 10.1109/TPAMI.2012.120 4

doi:10.1109/tpami.2012.120 2012

[2] [2]

K. Ai, K. Tang, and C. Wang. NLI4V olVis: Natural language interaction for volume visualization via LLM multi-agents and editable 3D Gaussian splatting.IEEE Transactions on Visualization and Computer Graphics, 32(1):46–56, 2026. doi: 10.1109/TVCG.2025.3633888 2, 3, 5, 6, 9

arXiv 2026

[3] [3]

Bauer, Q

D. Bauer, Q. Wu, H. Gadirov, and K.-L. Ma. GSCache: Real-time radiance caching for volume path tracing using 3D Gaussian splatting.IEEE Transactions on Visualization and Computer Graphics, 32(1):901–911,

work page

[4] [4]

J. F. Blinn. Models of light reflection for computer synthesized pictures. InProceedings of ACM SIGGRAPH Conference, pp. 192–198, 1977. doi: 10.1145/965141.563893 3

work page arXiv 1977

[5] [5]

Boorboor, M

S. Boorboor, M. S. Castellana, Y . Kim, C. Zhu-Tian, J. Beyer, H. Pfister et al. V oxAR: Adaptive visualization of volume rendered objects in optical see-through augmented reality.IEEE Transactions on Visualization and Computer Graphics, 30(10):6801–6812, 2023. doi: 10.1109/TVCG.2023. 3340770 1, 2

doi:10.1109/tvcg.2023 2023

[6] [6]

J. Cen, J. Fang, C. Yang, L. Xie, X. Zhang, W. Shen et al. Segment any 3D Gaussians. InProceedings of AAAI Conference on Artificial Intelligence, pp. 1971–1979, 2025. doi: 10.1609/AAAI.V39I2.32193 2, 3

work page doi:10.1609/aaai.v39i2.32193 1971

[7] [8]

Y . J. Cho and H. Y . Yoon. Numerical analysis of the rocom boron dilution benchmark experiment using the cupid code.Nuclear Engineering and Design, 341:167–175, 2019. doi: 10.1016/j.nucengdes.2018.11.005 14

work page doi:10.1016/j.nucengdes.2018.11.005 2019

[8] [9]

J. Choi, H. Kim, H. An, and Y . Jung. Real-time transfer function editor for direct volume rendering in mixed reality. InProceedings of ACM SIGGRAPH Asia Posters, pp. 38:1–38:2, 2024. doi: 10.1145/3681756. 3697953 2

work page doi:10.1145/3681756 2024