Pith. sign in

REVIEW 3 major objections 5 minor 62 references

GaussianSelector: Lightweight Human-Guided Object Selection in 3D Gaussian Splatting with Graph Optimization

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read Training-free graph cuts on 3D Gaussian superpoints select an object from one scribble view at 85.3 mIoU.

desk verdict A competent training-free graph-cut approach to sparse-scribble 3DGS selection, but the undocumented iterative loop in Algorithm 1 and loosely specified multi-round protocol make the headline numbers provisional until clarified. read the letter →

arxiv 2608.01492 v1 pith:ECTRY4NH submitted 2026-08-02 cs.CV

classification cs.CV
keywords 3DGaussianSplattinginteractiveobjectselectiongraphcutsuperpointsscribbleliftingtraining-freesegmentationhuman-in-the-loopvisibility-aware
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

GaussianSelector aims to show that complete 3D object selection in a reconstructed Gaussian Splatting scene can be driven by very sparse user input—one or a few 2D scribble strokes—without retraining the scene representation and without any pretrained image segmentation model. It turns the native Gaussian primitives into a superpoint graph: appearance is reparameterized by a deterministic Canonical Axis Color descriptor, Gaussians are grouped by spatial and appearance proximity, and edges are weighted by continuity of position, color, and opacity. User scribbles are lifted into 3D by alpha-transmittance coverage, and the object is recovered as the global minimum of a binary graph-cut energy. On the NVOS benchmark the paper reports 85.3 mIoU from a single scribble view and 92.2 mIoU after two additional refinement rounds, close to the 92.5 of a strong all-views, SAM-mask method, while running in minutes on CPU without GPU inference. If these numbers hold, interactive 3D selection becomes practical where dense multi-view supervision and large GPU memory are unavailable.

What carries the argument

The load-bearing construction is the continuity-weighted superpoint graph combined with visibility-aware scribble lifting. The edge weight $w_{ij} = \exp(-d_{ij}^2/\sigma_{ij}^2)$ with $d_{ij} = w_x d_x + w_c d_c + w_o d_o$, self-tuned $\sigma_{ij}$ from local k-NN distance medians, encodes where label boundaries should and should not fall; the scribble-lifting coverage $\rho_i = \sum_p \alpha_{ip} T_{ip} M(p) / (\sum_p \alpha_{ip} T_{ip} + \epsilon)$ maps 2D strokes onto Gaussian seeds through $\alpha$-composited visibility. Together they reduce interactive 3D selection to a one-time cached graph construction plus a fast per-round min-cut, which is what makes the claimed CPU-only, multi-round

What would settle it

On a single NVOS view, shift the foreground scribble by one pixel and rerun the single-round pipeline; if mIoU drops by more than a few points, the transmittance-weighted lifting in Eq. (6) is too fragile. In a custom scene with a transparent object in front of a similarly colored object, scribble only the front object from one view: the alpha-composited coverage will mix both surfaces, and if the resulting selection still achieves the reported ~85 mIoU regime, the lifting assumption survives; if not, the central claim is contradicted.

Watch

Extended reading notes

Core claim

GaussianSelector's central discovery is that a 3D object's identity is already encoded in the intrinsic structure of Gaussian primitives, so it can be recovered by a graph-cut labeled over scene-native superpoints, with no learned feature field or lifted SAM masks. Each Gaussian is described by an 18-dimensional Canonical Axis Color (CAC) feature—SH radiance evaluated along the six canonical axes of the primitive's anisotropic support—which avoids the non-unique representation ambiguity of raw SH coefficients. Gaussians are aggregated into superpoints via Leiden community detection on spatial and appearance proximity, and a k-NN superpoint graph carries continuity weights from normalized spa

Load-bearing premise

The load-bearing premise is that a scribble's visible footprint in a rendered view, weighted by how much each Gaussian contributes to those pixels, reliably reveals the user's intended object; at object boundaries, with transparent or reflective surfaces, or when one Gaussian serves both scribbled and unscribbled pixels, wrong seed labels can be created that later stages cannot fully repair.

Editorial extensions

If this is right

  • Single-view scribble interaction becomes a usable mode: the paper reports 85.3 mIoU from one NVOS scribble view, improving on the 70.1 of the original NVOS method.
  • A user can refine a selection round by round from new viewpoints, with each round reusing the cached graph and costing roughly 0.2 minutes on CPU; the paper reports 89.6 mIoU after one refinement round and 92.2 after two.
  • No pretrained segmentation network or GPU feature-field training is needed for the interaction itself, so the approach drops VRAM requirements and removes dependency on SAM-quality masks.
  • On 3D-OVS the method reaches 93.2 mIoU (93.6 with the ROI variant), within reach of the 94.4 reported by the all-views GaussianCut baseline.
  • The graph-cut formulation makes the result globally optimal for the given evidence and prior; the ablation shows scribbles alone yield 33.2 mIoU, graph alone 61.0, and the full model 85.3.

Reading between the lines

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

  • If the Neyman-Pearson framing is taken literally, the method's ceiling is set by how well two GMMs capture true foreground/background appearance; replacing them with nonparametric density estimates would be a direct test of whether the 85.3 single-view number is appearance-model-limited.
  • Because CAC samples only the DC and first-order directional SH response, specular or view-dependent materials may be conflated; a testable extension is to evaluate on scenes with highlights or transparent objects, where Section 3.3's transmittance lifting is most fragile.
  • The same superpoint-graph-plus-visibility-lifting scaffold could be transferred to other primitive-based radiance representations or point clouds once an equivalent appearance descriptor and coverage map are defined; the paper does not claim this.
  • A likely practical consequence the paper leaves implicit: because the scene graph is cached once, the per-round update cost is small enough to support real-time interactive editing on a laptop CPU, which the reported 0.2-minute rounds suggest.
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 / 5 minor

Summary. GaussianSelector proposes a training-free interactive 3D object selection method for 3D Gaussian Splatting. From sparse user scribbles on one or a few views, it builds a scene-native superpoint graph: Gaussians are reparameterized with a Canonical Axis Color (CAC) descriptor, coarsened via Leiden clustering, and connected with continuity-weighted k-NN edges. Scribbles are lifted to Gaussian-level seeds through transmittance-weighted coverage, aggregated to superpoint seeds, and used to fit GMM appearance models. Selection is posed as a binary submodular graph-cut energy (Eq. 1) solved exactly by s-t min-cut. The paper reports 85.3 mIoU on NVOS with a single scribble view and 92.2 mIoU with three interaction rounds, competitive with SAM-lifting baselines that use all views, while requiring no retraining and substantially less computation.

Significance. If the central claim holds, the contribution is significant: a training-free, GPU-light, sparse-scribble method that matches dense multi-view SAM-based selection would lower the barrier for interactive 3DGS editing and asset extraction. Strengths include the clean separation of scribble-independent scene encoding from interactive inference, the use of an exact graph-cut solver, and concrete runtime measurements showing large gains over feature-field and SAM-lifting baselines. The method is falsifiable on standard benchmarks and the component ablation in Table 4, once clarified, can support the design. However, the current manuscript leaves a load-bearing algorithmic loop unspecified and the multi-round protocol under-controlled, so the headline numbers are not yet reproducible as described.

major comments (3)
  1. [Algorithm 1 / §3.5] Algorithm 1 lines 8–13 contain a repeat loop that is never described in the method text. Section 3.5 states: 'Once the optimal labeling L* is obtained, it is broadcasted back...', presenting Eq. (1) as a single optimization. In contrast, Algorithm 1, after each min-cut, restricts the superpoint graph to the foreground subgraph {k: L*_k = F} and re-estimates the likelihood models on the restricted graph. The convergence criterion is unspecified, the number of iterations is not reported, and this loop is not ablated in Table 4. Moreover, restricting to foreground nodes removes all background seeds, making re-estimation of p_B undefined unless additional context nodes are retained, which is not described. As written, the reported 85.3/92.2 mIoU may be produced by an undocumented iterative post-processing step rather than by the energy in Eq. (1). Please either remove the loop, specify it pr
  2. [Table 1 / §4.2] The multi-round refinement protocol is not reproducible from the manuscript. Section 4.2 states that 'users are allowed to select informative views and inject scribbles', but no procedure is given for selecting these views, how many strokes are used, how ambiguous cases are resolved, or how the '2 Rounds' and '3 Rounds' rows were obtained. Since mIoU is reported without standard deviations or multiple annotators, the 89.6 and 92.2 numbers could depend strongly on favorable view selection. The paper should specify a fixed view schedule, report per-task breakdowns, and ideally include multiple-user variability. Without this, the central 'fewer views' claim is not yet a controlled comparison.
  3. [Table 1 / Fig. 3] The comparison to GaussianCut under sparse views is qualitative only. Table 1 reports GaussianCut only with ALL views, while the text and Fig. 3 claim GaussianCut degrades when given fewer views. Since GaussianCut is the most directly comparable graph-cut baseline on 3DGS, the paper should report quantitative IoU for GaussianCut under the same 1-view and 3-view settings (or explain why such a comparison is not possible). Without this, the claim that GaussianSelector matches dense multi-view methods using fewer views is not directly established for the closest baseline.
minor comments (5)
  1. [Table 4] The check marks in Table 4 appear inconsistent with the text in §4.4. The text says adding graph propagation improves from 33.2 to 61.0, and adding unary modeling reaches 79.6, which implies the 'w/o unary' row should have Unary=×, Graph=✓, and the 'w/o graph' row should have Unary=✓, Graph=×. The table as printed shows the opposite. Please fix the row labels or the marks.
  2. [§3.4, Eq. (9)] The affine normalization parameters m and s are described only in words ('midpoint between seed medians' and 'adaptive scale determined by separation') and the target seed confidence of 0.95 is listed in Implementation Details but not connected to Eq. (9). Please give the exact formula for s and how the 0.95 confidence is used.
  3. [§3.2] The distances d_x, d_c, d_o in Eq. (4) are called 'normalized', but no normalization procedure is defined. Also, the gating threshold is a quantile of which distribution? Specify these details for reproducibility.
  4. [§3.3] Equation (6) uses α_ip and T_ip without defining the pixel index convention and how the per-view threshold is set. The phrase 'per-view threshold' is not specified numerically or algorithmically; please clarify how thresholds are chosen and how the majority vote tie rule works.
  5. [§4.1] All experiments are reported without error bars or significance tests. At minimum, report standard deviations over multiple runs or multiple view selections for the key Table 1 rows.

Circularity Check

1 steps flagged · score 2.0 of 10

No significant circularity: reported mIoU is measured against held-out ground truth; only a minor overlapping-author self-citation motivates the CAC reparameterization.

  1. self citation load bearing [Section 3.1, Eq. (3)]
    "raw SH coefficients suffer from non-unique representation ambiguity, the same rendered appearance can be expressed by multiple distinct coefficient vectors (Xin et al. 2025), making coefficient-space distances an unreliable proxy for perceptual similarity."

    The sole support for the premise that raw SH coefficient distances are unreliable is (Xin et al. 2025), whose author list overlaps with the present paper (Yuheng Liu, Xiaohui Xie, Xinke Li). This premise motivates the CAC reparameterization that underlies the superpoint graph and unary evidence. However, the paper's headline mIoU numbers are measured against held-out ground truth, and the ablation 'w/o CAC' (80.3 vs 85.3) shows CAC is empirically beneficial, so the self-citation is motivational rather than the load-bearing source of the reported performance.

full rationale

The central derivation chain is self-contained: superpoints are built from Gaussian geometry/appearance, user scribbles are lifted by transmittance-weighted coverage (Eq. 6), GMMs are fit to those lifted seeds, and the graph-cut energy (Eq. 1) is minimized and then evaluated against ground-truth masks. The GMM fitting to user-provided seeds is standard supervised evidence modeling, not a fitted parameter renamed as a prediction. The only overlapping-author citation is the SH-ambiguity motivation for CAC; it does not force the experimental outcome, and the ablation shows CAC contributes beyond the citation. Algorithm 1's iterative foreground-restriction loop (lines 8-13) is undocumented and unablated, which is a reproducibility/correctness concern, but it does not make the derivation circular because the final evaluation is still against ground truth. Overall, no prediction reduces by construction to its inputs.

Assumptions & free parameters 11 free parameters · 7 assumptions · 1 invented entities

The method rests on a set of hand-set hyperparameters and modeling assumptions. The most consequential are the unreported continuity weights in Eq. (4), the unquantified coverage threshold in Eq. (6), and the user-adjustable superpoint resolution; none receive sensitivity analysis. The CAC and superpoint assumptions are introduced specifically for this pipeline.

free parameters (11)
  • k-NN neighborhood size k = 8
    Fixed for all experiments; controls graph connectivity and superpoint structure.
  • Continuity weight hyperparameters w_x, w_c, w_o = not reported; only constraint w_x+w_c+w_o=1
    These weights balance spatial, CAC, and opacity distances in Eq. (4); without exact values the graph prior is not fully specified.
  • Edge gating quantile threshold = 0.95
    Removes outlier edges beyond the 95th percentile per distance type in Sec. 3.2.
  • GMM components for appearance contrast = 3
    Number of mixture components for p_F and p_B in Eq. (8).
  • Target seed confidence / affine scale s = 0.95
    Used in Eq. (9) for contrastive affine normalization; exact mapping to m and s not specified.
  • Seed-evidence weight beta = 4.0
    Weight for reinforcing seed labels in the unary term Eq. (10).
  • Over-segmentation resolution (Leiden resolution) = user-adjustable, default not given
    Controls granularity of the superpoint abstraction; user-tuned per task.
  • Connected component filtering threshold = user-adjustable, default not given
    Removes isolated regions after graph-cut optimization.
  • Scale-outlier criterion = user-adjustable, default not given
    Rejects excessively large Gaussian primitives caused by highlight or shadow artifacts.
  • Seed coverage threshold and per-view majority vote tie rule = not specified
    Eq. (6) requires a per-view threshold to label Gaussians as F/B/U; this threshold is not quantified.
  • Maximum iterations in Algorithm 1 = not specified
    Algorithm 1 repeats foreground-subgraph restriction until convergence or max iterations; max is not specified.
assumptions (7)
  • standard math The binary Potts energy E(L) in Eq. (1) is submodular, so s-t min-cut returns the exact global optimum.
    Standard result for binary submodular energy minimization (Boykov and Funka-Lea 2006), invoked in Sec. 3.5.
  • domain assumption The pairwise continuity prior with weights w_ij from Eq. (4) correctly aligns label boundaries with object seams.
    The graph-cut approach assumes that appearance and spatial distances on superpoints are predictive of object boundaries; if this fails, the propagated labels are wrong. Introduced in Sec. 3.2.
  • ad hoc to paper The CAC feature in Eq. (3), SH radiance sampled along six local axes, captures the perceptual appearance of a Gaussian sufficiently for contrast modeling.
    CAC is introduced in this paper as an appearance descriptor. It is not derived from a uniqueness theorem and is justified partly by a self-cited reference (Xin et al. 2025).
  • domain assumption Leiden community detection on a k-NN graph of Gaussians produces superpoints whose boundaries align with object boundaries.
    If the superpoints cut through an object or merge object and background, the graph-cut solution cannot recover the true selection. Used in Sec. 3.2.
  • domain assumption Transmittance-weighted scribble coverage rho_i (Eq. 6) is a reliable proxy for user intent when lifting 2D strokes to Gaussian seeds.
    The alpha-composited footprint of a Gaussian is used to decide whether it is foreground or background; this ignores multi-view consistency and can be unreliable for transparent or boundary Gaussians (Sec. 3.3).
  • domain assumption The GMM likelihood ratio delta_k is the most powerful discriminator between foreground and background given the node features.
    Neyman-Pearson optimality applies to known densities; here p_F and p_B are estimated from possibly sparse seeds, so the guarantee is an assumption rather than a theorem (Sec. 3.4).
  • domain assumption Hyperparameters (k=8, quantile 0.95, beta=4.0, 3 GMM components, seed confidence 0.95) transfer across scenes without per-scene tuning.
    The paper sets these defaults once and evaluates on multiple benchmarks; no sensitivity analysis is provided (Sec. 4.1).
invented entities (1)
  • Canonical Axis Color (CAC) feature
    purpose: Appearance descriptor for each 3D Gaussian, evaluating SH radiance along six local canonical axes to support superpoint clustering and foreground/background contrast.
    A hand-defined descriptor introduced in this paper; its sufficiency is only validated through end-to-end IoU on benchmarks, not through an independent prediction.

how reviews work

0 comments
Cite this review

Pith. "Pith review of GaussianSelector: Lightweight Human-Guided Object Selection in 3D Gaussian Splatting with Graph Optimization." pith.science (2026). https://pith.science/paper/ECTRY4NH

@misc{pith2026260801492,
  author       = {Pith},
  title        = {Pith review of: GaussianSelector: Lightweight Human-Guided Object Selection in 3D Gaussian Splatting with Graph Optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ECTRY4NH}},
  note         = {Machine review of arXiv:2608.01492}
}
read the original abstract

Selecting a complete 3D object from a reconstructed scene with minimal user effort is essential for practical scene editing and embodied interaction. Existing 3DGS-based methods either retrain the Gaussian representation to embed per-object labels, or build dense multi-view SAM observations, both requiring heavy computation and dense viewpoint coverage that is rarely available in practice. We present GaussianSelector, a training-free framework for interactive 3D object selection from sparse views and sparse scribble guidance. Operating directly on native Gaussian primitives, we coarsen dense Gaussians into geometrically coherent superpoints and construct a continuity-weighted graph using appearance and spatial cues. Sparse user scribbles are lifted into 3D via visibility-aware transmittance coverage, and selection is solved as a global graph-cut energy minimization that propagates sparse evidence to a complete 3D object. This design naturally supports multi-round refinement, where users iteratively correct the selection from additional viewpoints to progressively improve the result. Experiments demonstrate that GaussianSelector achieves competitive selection quality against state-of-the-art multi-view SAM-based methods, while requiring significantly fewer interaction views and substantially lower computational overhead. These properties make it well suited for human-in-the-loop 3D scene editing and 3D asset extraction in real-world deployment scenarios.

Figures

Figures reproduced from arXiv: 2608.01492 by the authors.

Figure 1
Figure 1. Overview of GaussianSelector. We propose a plug-and-play interactive framework for object selection in 3DGS. Given user scribbles on a single view or a few views, GaussianSelector extracts 3D objects efficiently using graph-based optimization in the 3DGS space. Abstract Selecting a complete 3D object from a reconstructed scene with minimal user effort is essential for practical scene edit￾ing and embodied interactio… view at source ↗
Figure 2
Figure 2. Method overview. GaussianSelector first converts a reconstructed 3DGS scene into a superpoint graph using CAC [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Qualitative comparison on NVOS under various [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Qualitative comparison on 3D-OVS. Our method [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Visualization of the intermediate steps. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: User interaction demonstration for human-in-the [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

62 extracted references · 32 canonical work pages

  1. [1]

    , author=

    3d gaussian splatting for real-time radiance field rendering. , author=. ACM Trans. Graph. , volume=

  2. [2]

    Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

    Pointnet: Deep learning on point sets for 3d classification and segmentation , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

  3. [3]

    Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

    Scannet: Richly-annotated 3d reconstructions of indoor scenes , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

  4. [4]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Point transformer , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  5. [5]

    International Journal of Computer Vision , volume=

    Segment anything in 3d with radiance fields , author=. International Journal of Computer Vision , volume=. 2025 , publisher=

  6. [6]

    Proceedings of the European conference on computer vision (ECCV) , pages=

    Depth-aware cnn for rgb-d segmentation , author=. Proceedings of the European conference on computer vision (ECCV) , pages=

  7. [7]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Langsplat: 3d language gaussian splatting , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  8. [8]

    ACM Transactions on Graphics (tog) , volume=

    Dynamic graph cnn for learning on point clouds , author=. ACM Transactions on Graphics (tog) , volume=. 2019 , publisher=

Show all 62 references
  1. [9]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Mip-splatting: Alias-free 3d gaussian splatting , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  2. [10]

    ACM SIGGRAPH 2024 conference papers , pages=

    2d gaussian splatting for geometrically accurate radiance fields , author=. ACM SIGGRAPH 2024 conference papers , pages=

  3. [11]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Structured 3d latents for scalable and versatile 3d generation , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  4. [12]

    arXiv preprint arXiv:2501.16764 , year=

    Diffsplat: Repurposing image diffusion models for scalable gaussian splat generation , author=. arXiv preprint arXiv:2501.16764 , year=

  5. [13]

    arXiv preprint arXiv:2509.22917 , year=

    Learning Unified Representation of 3D Gaussian Splatting , author=. arXiv preprint arXiv:2509.22917 , year=

  6. [14]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    4d gaussian splatting for real-time dynamic scene rendering , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  7. [15]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Baking gaussian splatting into diffusion denoiser for fast and scalable single-stage image-to-3d generation and reconstruction , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  8. [16]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Segment anything , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  9. [17]

    arXiv preprint arXiv:2408.00714 , year=

    Sam 2: Segment anything in images and videos , author=. arXiv preprint arXiv:2408.00714 , year=

  10. [18]

    arXiv preprint arXiv:2411.07184 , year=

    Sampart3d: Segment any part in 3d objects , author=. arXiv preprint arXiv:2411.07184 , year=

  11. [19]

    International conference on machine learning , pages=

    Learning transferable visual models from natural language supervision , author=. International conference on machine learning , pages=. 2021 , organization=

  12. [20]

    arXiv preprint arXiv:2309.16671 , year=

    Demystifying clip data , author=. arXiv preprint arXiv:2309.16671 , year=

  13. [21]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Emerging properties in self-supervised vision transformers , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  14. [22]

    arXiv preprint arXiv:2304.07193 , year=

    Dinov2: Learning robust visual features without supervision , author=. arXiv preprint arXiv:2304.07193 , year=

  15. [23]

    Proceedings of the AAAI conference on artificial intelligence , volume=

    Segment any 3d gaussians , author=. Proceedings of the AAAI conference on artificial intelligence , volume=

  16. [24]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Omniseg3d: Omniversal 3d segmentation via hierarchical contrastive learning , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  17. [25]

    European Conference on Computer Vision , pages=

    Click-gaussian: Interactive segmentation to any 3d gaussians , author=. European Conference on Computer Vision , pages=. 2024 , organization=

  18. [26]

    arXiv preprint arXiv:2404.07977 , year=

    Gaga: Group any gaussians via 3d-aware memory bank , author=. arXiv preprint arXiv:2404.07977 , year=

  19. [27]

    GrabCut

    " GrabCut" interactive foreground extraction using iterated graph cuts , author=. ACM transactions on graphics (TOG) , volume=. 2004 , publisher=

  20. [28]

    Advances in Neural Information Processing Systems , volume=

    Gaussiancut: Interactive segmentation via graph cut for 3d gaussian splatting , author=. Advances in Neural Information Processing Systems , volume=

  21. [29]

    arXiv preprint arXiv:2510.21307 , year=

    Towards Physically Executable 3D Gaussian for Embodied Navigation , author=. arXiv preprint arXiv:2510.21307 , year=

  22. [30]

    Advances in neural information processing systems , volume=

    Decomposing nerf for editing via feature field distillation , author=. Advances in neural information processing systems , volume=

  23. [31]

    arXiv preprint arXiv:2412.01583 , year=

    3dsceneeditor: Controllable 3d scene editing with gaussian splatting , author=. arXiv preprint arXiv:2412.01583 , year=

  24. [32]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Neural volumetric object selection , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  25. [33]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Spin-nerf: Multiview segmentation and perceptual inpainting with neural radiance fields , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  26. [34]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Feature 3dgs: Supercharging 3d gaussian splatting to enable distilled feature fields , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  27. [35]

    arXiv preprint arXiv:2410.07577 , year=

    3d vision-language gaussian splatting , author=. arXiv preprint arXiv:2410.07577 , year=

  28. [36]

    European Conference on Computer Vision , pages=

    N2f2: Hierarchical scene understanding with nested neural feature fields , author=. European Conference on Computer Vision , pages=. 2024 , organization=

  29. [37]

    arXiv preprint arXiv:2201.03546 , year=

    Language-driven semantic segmentation , author=. arXiv preprint arXiv:2201.03546 , year=

  30. [38]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Language embedded 3d gaussians for open-vocabulary scene understanding , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  31. [39]

    Proceedings of the 32nd ACM international conference on multimedia , pages=

    Goi: Find 3d gaussians of interest with an optimizable open-vocabulary semantic-space hyperplane , author=. Proceedings of the 32nd ACM international conference on multimedia , pages=

  32. [40]

    International Journal of Computer Vision , volume=

    Fmgs: Foundation model embedded 3d gaussian splatting for holistic 3d scene understanding , author=. International Journal of Computer Vision , volume=. 2025 , publisher=

  33. [41]

    Proceedings of the AAAI conference on artificial intelligence , volume=

    Fastlgs: Speeding up language embedded gaussians with feature grid mapping , author=. Proceedings of the AAAI conference on artificial intelligence , volume=

  34. [42]

    arXiv preprint arXiv:2401.17857 , year=

    SAGD: Boundary-enhanced segment anything in 3D Gaussian via Gaussian decomposition , author=. arXiv preprint arXiv:2401.17857 , year=

  35. [43]

    European conference on computer vision , pages=

    Gaussian grouping: Segment and edit anything in 3d scenes , author=. European conference on computer vision , pages=. 2024 , organization=

  36. [44]

    Advances in Neural Information Processing Systems , volume=

    Opengaussian: Towards point-level 3d gaussian-based open vocabulary understanding , author=. Advances in Neural Information Processing Systems , volume=

  37. [45]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Votesplat: Hough voting gaussian splatting for 3d scene understanding , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  38. [46]

    Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

    isegman: Interactive segment-and-manipulate 3d gaussians , author=. Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

  39. [47]

    Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

    Instancegaussian: Appearance-semantic joint gaussian representation for 3d instance-level perception , author=. Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

  40. [48]

    European Conference on Computer Vision , pages=

    Flashsplat: 2d to 3d gaussian splatting segmentation solved optimally , author=. European Conference on Computer Vision , pages=. 2024 , organization=

  41. [49]

    Advances in Neural Information Processing Systems , volume=

    Gaussian graph network: Learning efficient and generalizable gaussian representations from multi-view images , author=. Advances in Neural Information Processing Systems , volume=

  42. [50]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    AG ^2 aussian: Anchor-Graph Structured Gaussian Splatting for Instance-Level 3D Scene Understanding and Editing , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  43. [51]

    2023 IEEE International Conference on Robotics and Automation (ICRA) , pages=

    Interactive object segmentation in 3d point clouds , author=. 2023 IEEE International Conference on Robotics and Automation (ICRA) , pages=. 2023 , organization=

  44. [52]

    arXiv preprint arXiv:2306.00977 , year=

    Agile3d: Attention guided interactive multi-object 3d segmentation , author=. arXiv preprint arXiv:2306.00977 , year=

  45. [53]

    Advances in Neural Information Processing Systems , volume=

    A unified framework for 3d scene understanding , author=. Advances in Neural Information Processing Systems , volume=

  46. [54]

    SIGGRAPH Asia 2024 Conference Papers , pages=

    iseg: Interactive 3d segmentation via interactive attention , author=. SIGGRAPH Asia 2024 Conference Papers , pages=

  47. [55]

    International journal of computer vision , volume=

    Graph cuts and efficient ND image segmentation , author=. International journal of computer vision , volume=. 2006 , publisher=

  48. [56]

    Advances in Neural Information Processing Systems , volume=

    Learning superpoint graph cut for 3d instance segmentation , author=. Advances in Neural Information Processing Systems , volume=

  49. [57]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Open3dis: Open-vocabulary 3d instance segmentation with 2d mask guidance , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  50. [58]

    ACM Transactions on Graphics (ToG) , volume=

    Local light field fusion: Practical view synthesis with prescriptive sampling guidelines , author=. ACM Transactions on Graphics (ToG) , volume=. 2019 , publisher=

  51. [59]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Mip-nerf 360: Unbounded anti-aliased neural radiance fields , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  52. [60]

    Advances in Neural Information Processing Systems , volume=

    Weakly supervised 3d open-vocabulary segmentation , author=. Advances in Neural Information Processing Systems , volume=

  53. [61]

    The Thirteenth International Conference on Learning Representations , year=

    Hqgs: High-quality novel view synthesis with gaussian splatting in degraded scenes , author=. The Thirteenth International Conference on Learning Representations , year=

  54. [62]

    arXiv preprint arXiv:2510.08566 , year=

    D ^2 GS: Depth-and-Density Guided Gaussian Splatting for Stable and Accurate Sparse-View Reconstruction , author=. arXiv preprint arXiv:2510.08566 , year=

Pith tools

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