Pith. sign in

REVIEW 5 major objections 5 minor 58 references

Efficient Semantic Splatting for Remote Sensing Multi-view Segmentation

T0 review · 5 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read An explicit Gaussian-splatting renderer with per-point semantic features achieves higher average mIoU and roughly 90% lower rendering latency than NeRF-based methods in sparse-label remote sensing multi-view segmentation.

desk verdict A plausible first application of 3D Gaussian splatting to remote sensing multi-view segmentation under sparse labels, with genuine speed advantages, but the evaluation leaks tuning choices into the test set and the SAM2 pseudo-label branch is under-specified. read the letter →

arxiv 2412.05969 v2 pith:RVBMGWED submitted 2024-12-08 cs.CV

Zipeng Qi , Hao Chen , Haotian Zhang , Zhengxia Zou , Zhenwei Shi This is my paper · ORCID
classification cs.CV
keywords remotesensingsemanticsegmentationmulti-viewGaussianSplattingSAM2sparselabelspseudo-labelsaggregationloss
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

Remote sensing scenes photographed from many viewpoints usually need either dense annotations for per-image networks or slow per-scene optimization in an implicit neural field to stay view-consistent. This paper claims that an explicit Gaussian-splatting representation can replace the implicit NeRF stage: each 3D point carries color and a continuous semantic feature vector, and the same alpha-blending splat renders both an RGB image and a semantic feature map that a small decoder converts into class labels. With only 2–6% of views labeled, the method reports higher average mIoU (mean Intersection-over-Union) than the compared training- and optimization-based methods on six synthetic and three real sub-datasets, while cutting per-view rendering time by about 90% and optimization time to about ten minutes. To make up for the explicit point cloud's weaker spatial continuity and for missing boundary supervision, the paper adds SAM2-generated boundary pseudo-labels and two aggregation losses that pull neighboring semantic features together in the rendered feature map and in 3D space. If the claims hold, sparse-label multi-view segmentation of remote sensing scenes becomes fast enough for practical downstream use, with the optimized point cloud itself serving as a semantic 3D representation.

What carries the argument

The central object is an explicit 3D Gaussian semantic field: a COLMAP-initialized point cloud in which each Gaussian stores position, covariance, opacity, spherical-harmonic color coefficients, and a 16-dimensional continuous semantic feature vector. Splatting projects these Gaussians to the image plane and alpha-blends color and semantic features through the same opacity weights, producing a semantic feature map that a lightweight MLP decoder maps to class probabilities. The same explicit structure that makes rendering fast also motivates two aggregation losses, one on the rendered 2D feature map and one on the 3D point cloud, plus SAM2-derived boundary pseudo-labels, to restore the spatial continuity that explicit point clouds lack compared to implicit neural fields. The load-bearing identity is the shared alpha-blending pass: because color and semantics are rendered together, RGB reconstruction supervises geometry while semantic losses supervise the feature vector, and all views inherit one consistent 3D field.

What would settle it

Run the same nine-scene evaluation while replacing SAM2 pseudo-labels with ground-truth labels in the boundary mask B; if the mIoU gap between the two conditions is near zero, the pseudo-label stream is not the source of the claimed gains. A cheaper check is to compute per-pixel agreement between SAM2 pseudo-labels and ground truth inside B across all unlabeled views and confirm that it exceeds chance in boundary bands.

Watch

Extended reading notes

Core claim

The paper's central claim is that multi-view semantic segmentation of remote sensing scenes under sparse labels can be solved by treating it as a rendering problem in an explicit 3D Gaussian field, and that this approach is both more accurate and dramatically faster than the implicit NeRF-based alternatives. The design renders color and a continuous per-point semantic feature vector through the same splatting and alpha-blending operation, then maps the rendered feature map to class probabilities with a small MLP decoder, so that view consistency is a property of the shared explicit scene rather than of per-image networks. On the nine sub-datasets used in the evaluation, the paper reports the highest average mIoU, an average mIoU of 61.21 against 58.67 for the closest NeRF baseline IRT (summarized in the abstract as up to 3.5% improvement), a reduction of per-view rendering time by about 90% (under one second), and an optimization time of about ten minutes. The paper further claims that the gains over a plain semantic-splatting baseline come from SAM2 boundary pseudo-labels and from two aggregation losses that enforce feature similarity among neighboring points in the rendered feature map and in 3D space.

Load-bearing premise

The method's biggest load-bearing premise is that SAM2's boundary pseudo-labels, made from a single annotated view and reused for most of the optimization views, are correct enough in boundary regions to improve the learned semantic field rather than injecting errors; the paper does not test their accuracy directly.

Editorial extensions

If this is right

  • If the central claim is right, sparse-label multi-view segmentation no longer requires hours of NeRF-style optimization; a scene can be optimized in about ten minutes and a novel view rendered in under one second, making interactive or operational use viable.
  • Because every view is rendered from the same optimized Gaussian point cloud, semantic predictions are view-consistent by construction, eliminating the per-image inconsistencies shown for CNN- and Transformer-based baselines.
  • The same optimized representation carries both color and semantic features, so downstream products such as semantically labeled meshes or point clouds can be extracted directly from the splatting field.
  • The reported ablations imply that SAM2 boundary pseudo-labels and the two aggregation losses are all load-bearing: removing any one lowers average mIoU, so the method's accuracy is not solely due to the splatting backbone.

Reading between the lines

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

  • A testable extension would be to use the optimized semantic point cloud directly for downstream 3D tasks such as semantic mesh labeling or object extraction, since the same Gaussians already carry class-related features; the paper mentions such downstream outputs but does not evaluate them.
  • The fixed 1:8 ratio between ground-truth and pseudo-labeled views was tuned on two scenes; an adaptive schedule that lowers pseudo-label weight as optimization converges could generalize better across scenes with different label ratios.
  • Because the two aggregation losses pull all neighboring semantic features together, an edge-aware variant that reduces aggregation across detected boundaries could protect thin structures such as road markings and power lines; this is not explored in the paper.
  • Testing the method at label ratios below 2% would reveal whether the SAM2 pseudo-label stream can carry the method when almost no ground-truth supervision is available.
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

5 major / 5 minor

Summary. The paper proposes an optimization-based semantic segmentation method for remote sensing multi-view imagery under sparse label supervision, built on 3D Gaussian Splatting. Each point carries continuous semantic features that are splatted and rendered into per-view semantic feature maps, followed by an MLP transfer layer to produce segmentation. To handle boundary regions that lack supervision, the method uses SAM2 to generate pseudo-labels from a selected supervisory view, and it adds two aggregation losses, one on rendered 2D feature maps and one on 3D neighboring points. Experiments on the IRT dataset (six synthetic scenes and three real scenes) report mIoU and per-view rendering time, claiming higher accuracy than existing optimization-based methods and roughly 90% faster test-time rendering.

Significance. If the empirical claims are reliable, the paper is a meaningful step toward making multi-view remote sensing segmentation practical under sparse labels: the explicit Gaussian representation avoids the slow per-scene optimization and heavy rendering of NeRF-style baselines while retaining view consistency. The paper includes a broad comparison, an ablation study, and parameter sensitivity experiments, and the central idea of splatting semantic features is well motivated. However, the evaluation contains numerical inconsistencies in the headline claims, and the hyperparameters are tuned on scenes that later appear in the final average, so the significance is currently lower than the paper claims.

major comments (5)
  1. [Abstract and §IV-C, Table II] The reported gains are internally inconsistent. The abstract says 'up to 3.5%' mIoU improvement, and the text says the method surpasses Color-NeRF by 3.54% on average, but Table II gives AVG values of 56.84 for Color-NeRF and 61.21 for the proposed method, a difference of 4.37 mIoU. The same paragraph states the method outperforms IRT 'by over 15.5% in average mIoU,' whereas Table II shows IRT at 58.67 and the proposed method at 61.21, a difference of 2.54. These discrepancies are load-bearing because the abstract and introduction advertise the accuracy gain; the authors should correct the numbers and state exactly which baseline each gain refers to.
  2. [§IV-F and §IV-C, Table II] The hyperparameters—the ground-truth-to-pseudo-label view ratio and the weights a and b of the aggregation losses—are selected on sys #1 and real #1 (Section IV-F), and the same two scenes are then included in the average mIoU reported in Table II and in the ablation study of Table III. This is selection on the test set, which inflates the reported gains. The authors should either tune on held-out scenes (e.g., sys #2 and real #2) and report those results separately, or run multiple random splits and report means with error bars. Given the small per-scene gains in the ablation (e.g., +0.37 mIoU for pseudo-labels), this issue directly affects the strength of the central superiority claim.
  3. [§III-C and Algorithm 1] The SAM2 pseudo-label branch as specified cannot provide the boundary supervision that motivates it. The method randomly selects one supervisory view, segments that view with SAM2, builds a boundary mask B from that view, assigns pseudo-classes by majority vote against that view's label, and propagates the instance masks to all other views. Boundary regions that are not visible in the selected view are never pseudo-labeled, yet Figure 3 and the text identify exactly those 'boundary regions lacking supervision' as the target. Moreover, Eq. (10) applies the same boundary mask B to the rendered target view, which does not align with the target view's actual boundary pixels. The paper reports no accuracy, coverage, or boundary IoU for these pseudo-labels despite their dominating the optimization schedule (8 of every 9 sampled views per Section III-E). This is a load-bearing gap in the method's stated mechanism and should be addressed by specifying per-view mask generation and by evaluating pseudo-label quality or reducing the method's reliance on them.
  4. [§III-D, Eq. (11)] The 2D aggregation loss as written is not a well-defined similarity or divergence. Eq. (11) defines L2D_a = (1/(mk)) sum_i sum_j s_i log(s_i / s_j), where s_i and s_j are 16-dimensional continuous semantic features rather than normalized probability distributions. The expression is asymmetric, scale-dependent, and lacks a summation over feature channels, so it does not reduce to a standard KL divergence or a meaningful continuity loss. The authors should define a proper distance over the feature distribution (for example, softmax-normalized features with KL divergence, or cosine distance) and justify why that objective encourages same-class feature similarity. This matters because the aggregation losses are credited with a large part of the final gain (Section IV-E).
  5. [§IV-C, Table II] All quantitative results are reported as single runs with no error bars, repeated trials, or statistical significance tests. Because the improvements over the strongest baseline (IRT) are modest on several scenes (e.g., sys #2: 38.98 vs. 43.23, where the proposed method is actually lower than Color-NeRF's 38.46? This needs checking) and because hyperparameters are tuned on the test scenes, the current evidence does not support the sweeping claim of superiority across all sub-datasets. The authors should report per-run variance or at least state the number of seeds used.
minor comments (5)
  1. [Throughout] There are many typos and grammatical errors, including 'fellowing' (Section III-C), 'piexe' (Section II-B), 'spital' and 'smantic' (Section III-A), 'Uent' (Figure 1), and 'Villain Gaussian Splatting' (Section III-E, should be 'vanilla'). A thorough proofread is needed.
  2. [§IV-F, Tables V and VI] Table V is captioned 'coefficients for L2D_a' and Table VI is captioned 'coefficients for L2D_a' as well; the second should refer to L3D_a. The sentence after Table VI, 'The best coefficient of L2D_a is 0.1,' should read L3D_a.
  3. [§II-B, Eq. (1)] Equation (1) has mismatched parentheses: 'Φd(Φs(x, y, z), θ, β))' contains an extra closing parenthesis. This makes the equation hard to parse.
  4. [§III-C, Eq. (10)] The notation B is used for the boundary mask, but the paper does not state how B is obtained for a non-supervisory view. If B is always the mask from the selected supervisory view, this should be stated explicitly; if each view has its own B, the generation procedure must be described.
  5. [§IV-D and Figure 7] The view-consistency comparison is qualitative only. A quantitative measure of consistency across views, such as the standard deviation of predictions for corresponding pixels or the agreement between overlapping views, would strengthen the claim and is more informative than the single figure shown.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the reported accuracy and efficiency gains are empirical benchmark measurements, and the authors' prior-work citations supply the dataset and baselines without forcing the outcome.

full rationale

The paper's derivation chain is an empirical method proposal rather than a closed-form prediction: semantic features are rendered via alpha blending (Eq. 8), a transfer layer outputs classes (Eq. 9), and optimization uses CE with ground-truth or SAM2 pseudo-labels plus aggregation losses (Eqs. 10-12). The final mIoU numbers in Table II are measured against annotated ground truth after optimization; they are not algebraically implied by the loss terms or by any fitted constant. The main self-citations (IRT [20] for the dataset and the strongest baseline) are normal reuse of prior empirical resources: the real sub-datasets are Google Maps imagery with manual labels and the synthetic labels come from CARLA, so the comparison is not reduced to a self-citation chain. The SAM2 pseudo-label branch is a legitimate correctness risk because its accuracy is unmeasured and it supervises most optimization steps, but it provides external supervision rather than the model's own output, so it does not make the evaluation circular. One methodological concern is that Section IV-F selects the view ratio and loss weights on sys #1 and real #1, and those same sub-datasets appear in the final Table II; this weakens the independence of those two rows, but the reported mIoU values are measurements under selected hyperparameters, not quantities forced by construction from the fitted parameters. No uniqueness theorem, ansatz, or renamed known result is used as a load-bearing circular step.

Assumptions & free parameters 6 free parameters · 5 assumptions · 1 invented entities

The central claim rests on standard 3DGS rendering, COLMAP initialization, SAM2 pseudo-label accuracy, and the validity of alpha-blending semantic features. Hyperparameters are tuned on two of the nine scenes and reused elsewhere; no held-out validation set is specified.

free parameters (6)
  • a (weight of L2D aggregation loss) = 0.5
    Selected by parameter study on sys #1 and real #1 in Section IV-F; directly affects mIoU.
  • b (weight of L3D aggregation loss) = 0.1
    Selected by parameter study; Table VI header is mislabeled as L2D_a while the text discusses L3D_a.
  • Ground-truth to pseudo-label view ratio = 1:8
    Tuned in Section IV-F; controls the balance between reliable and noisy supervision.
  • Neighbor count k for aggregation losses = 5
    Set by hand in Section III-E; affects the smoothing range of the aggregation losses.
  • Semantic feature dimension = 16
    Set by hand in Section III-E; governs the capacity of the semantic representation.
  • Optimization steps = 30000
    Set by hand in Section III-E; controls convergence and total optimization time.
assumptions (5)
  • domain assumption Gaussian splatting rendering model (Kerbl et al.) accurately represents scene color and geometry for remote sensing multi-view inputs.
    The method reuses 3DGS alpha-blending as its core rendering operator (Eq. 6-8) without re-deriving or validating it for aerial and satellite imagery.
  • domain assumption COLMAP provides accurate camera poses and an initial point cloud for every scene.
    Section II-D states COLMAP estimates intrinsics and extrinsics; if poses or sparse points are inaccurate, splatting will misalign semantic features.
  • domain assumption SAM2 pseudo-labels propagated from one supervisory view are accurate enough in boundary regions to improve supervision.
    Section III-C uses SAM2 to label all unlabeled views and assigns pseudo-classes by majority overlap; no accuracy evaluation of these pseudo-labels is given.
  • domain assumption Continuous semantic features can be alpha-blended like color and decoded to semantic labels by a small MLP.
    Section III-B treats s_i as radiance-like and blends with color opacities; this assumes the feature field is view-independent and linearly decodable.
  • ad hoc to paper KL divergence between raw feature vectors is a valid continuity loss.
    Eq. 11 uses s_i log(s_i/s_j) without normalization or non-negativity constraints; this is a design choice, not a standard result.
invented entities (1)
  • Per-point continuous semantic feature vector s_i
    purpose: Enables differentiable alpha-blended rendering of semantic classes (Eq. 8).
    Latent quantity internal to the model; no external falsifiable prediction is made from the feature vector itself.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Efficient Semantic Splatting for Remote Sensing Multi-view Segmentation." pith.science (2026). https://pith.science/paper/RVBMGWED

@misc{pith2026241205969,
  author       = {Pith},
  title        = {Pith review of: Efficient Semantic Splatting for Remote Sensing Multi-view Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RVBMGWED}},
  note         = {Machine review of arXiv:2412.05969}
}
read the original abstract

In this paper, we propose a novel semantic splatting approach based on Gaussian Splatting to achieve efficient and low-latency. Our method projects the RGB attributes and semantic features of point clouds onto the image plane, simultaneously rendering RGB images and semantic segmentation results. Leveraging the explicit structure of point clouds and a one-time rendering strategy, our approach significantly enhances efficiency during optimization and rendering. Additionally, we employ SAM2 to generate pseudo-labels for boundary regions, which often lack sufficient supervision, and introduce two-level aggregation losses at the 2D feature map and 3D spatial levels to improve the view-consistent and spatial continuity.

Figures

Figures reproduced from arXiv: 2412.05969 by the authors.

Figure 1
Figure 1. Our method achieves superior efficiency and accuracy for multi-view [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The overview of the proposed model. Our method is an optimization-based semantic splatting approach for multi-view segmentation in remote sensing. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. For a target scene with only a few views, e.g., 3 views with supervision [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: We present input samples of four sub-datasets. [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: The visual results from sys #1, sys #3 and sys #5. The results show that our method achieves more accuracy and outperform other training-based and [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: The visual results from sys #6, real #2 and real #3. The results show that our method achieves more accuracy and outperform other training-based [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: The view-consistent results from sys #4. The results show that our methods can generate results with more accuracy and view consistency. The labels [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: The pseudo labels for boundary regions assist the network in improving accuracy in these challenging areas (shown in white boxes). [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: We calculate the PCA results of splatted semantic feature map. The results show the aggregation losses help the network enhance the stability of [PITH_FULL_IMAGE:figures/full_fig_p012_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 30 canonical work pages

  1. [1]

    Remote sensing image change detection with transformers,

    H. Chen, Z. Qi, and Z. Shi, “Remote sensing image change detection with transformers,” IEEE Transactions on Geoscience and Remote Sensing , vol. 60, pp. 1–14, 2021

    2021

  2. [2]

    Bifa: Remote sensing image change detec- tion with bitemporal feature alignment,

    H. Zhang, H. Chen, C. Zhou, K. Chen, C. Liu, Z. Zou, and Z. Shi, “Bifa: Remote sensing image change detec- tion with bitemporal feature alignment,” IEEE Transac- tions on Geoscience and Remote Sensing , 2024

    work page 2024

  3. [3]

    Pixel-level change detection pseudo-label learn- ing for remote sensing change captioning,

    C. Liu, K. Chen, Z. Qi, Z. Liu, H. Zhang, Z. Zou, and Z. Shi, “Pixel-level change detection pseudo-label learn- ing for remote sensing change captioning,” in IGARSS 2024-2024 IEEE International Geoscience and Remote Sensing Symposium . IEEE, 2024, pp. 8405–8408

    work page 2024

  4. [4]

    Cdmamba: Remote sensing image change detection with mamba,

    H. Zhang, K. Chen, C. Liu, H. Chen, Z. Zou, and Z. Shi, “Cdmamba: Remote sensing image change detection with mamba,” arXiv preprint arXiv:2406.04207 , 2024

    arXiv 2024

  5. [5]

    Road extraction in remote sensing data: A sur- vey,

    Z. Chen, L. Deng, Y . Luo, D. Li, J. M. Junior, W. N. Gonc ¸alves, A. A. M. Nurunnabi, J. Li, C. Wang, and D. Li, “Road extraction in remote sensing data: A sur- vey,” International journal of applied earth observation and geoinformation , vol. 112, p. 102833, 2022

    work page 2022

  6. [6]

    Topology-guided road graph extraction from remote sensing images,

    Y . Zao, Z. Zou, and Z. Shi, “Topology-guided road graph extraction from remote sensing images,” IEEE Transactions on Geoscience and Remote Sensing , 2023

    work page 2023

  7. [7]

    High- resolution remote sensing image scene understanding: A review,

    Q. Zhu, X. Sun, Y . Zhong, and L. Zhang, “High- resolution remote sensing image scene understanding: A review,” in IGARSS 2019-2019 IEEE International Geoscience and Remote Sensing Symposium . IEEE, 2019, pp. 3061–3064

    work page 2019

  8. [8]

    Mlrsnet: A multi-label high spatial resolution remote sensing dataset for semantic scene understanding,

    X. Qi, P. Zhu, Y . Wang, L. Zhang, J. Peng, M. Wu, J. Chen, X. Zhao, N. Zang, and P. T. Mathiopoulos, “Mlrsnet: A multi-label high spatial resolution remote sensing dataset for semantic scene understanding,” ISPRS Journal of Photogrammetry and Remote Sensing , vol. 169, pp. 337–350, 2020

    work page 2020

Show all 58 references
  1. [9]

    Multi-objects change detection based on res-unet,

    L. Yuan, Y . Li, Y . Si, J. Ren, Y . Yang, Y . Gong, Y . Xia, Z. Tong, and L. Tong, “Multi-objects change detection based on res-unet,” in 2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS . IEEE, 2021, pp. 4364–4367

    2021

  2. [10]

    Encoder-decoder with atrous separable convo- lution for semantic image segmentation,

    L.-C. Chen, Y . Zhu, G. Papandreou, F. Schroff, and H. Adam, “Encoder-decoder with atrous separable convo- lution for semantic image segmentation,” in Proceedings of the European conference on computer vision (ECCV) , 2018, pp. 801–818

    2018

  3. [11]

    Cnn- based segmentation of medical imaging data,

    B. Kayalibay, G. Jensen, and P. van der Smagt, “Cnn- based segmentation of medical imaging data,” arXiv preprint arXiv:1701.03056, 2017

    2017 arXiv

  4. [12]

    Hyperdense-net: a hyper-densely connected cnn for multi-modal image segmentation,

    J. Dolz, K. Gopinath, J. Yuan, H. Lombaert, C. Desrosiers, and I. B. Ayed, “Hyperdense-net: a hyper-densely connected cnn for multi-modal image segmentation,” IEEE transactions on medical imaging , vol. 38, no. 5, pp. 1116–1126, 2018

    2018

  5. [13]

    Rethinking semantic segmentation from a sequence-to-sequence perspective with transformers,

    S. Zheng, J. Lu, H. Zhao, X. Zhu, Z. Luo, Y . Wang, Y . Fu, J. Feng, T. Xiang, P. H. Torr et al., “Rethinking semantic segmentation from a sequence-to-sequence perspective with transformers,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition ...

    2021

  6. [14]

    Seg- menter: Transformer for semantic segmentation,

    R. Strudel, R. Garcia, I. Laptev, and C. Schmid, “Seg- menter: Transformer for semantic segmentation,” in Pro- ceedings of the IEEE/CVF international conference on computer vision , 2021, pp. 7262–7272

    2021

  7. [15]

    Missformer: An effective medical image segmentation transformer,

    X. Huang, Z. Deng, D. Li, and X. Yuan, “Missformer: An effective medical image segmentation transformer,” arXiv preprint arXiv:2109.07162 , 2021

    2021 arXiv

  8. [16]

    Transformer-based visual segmentation: A survey,

    X. Li, H. Ding, H. Yuan, W. Zhang, J. Pang, G. Cheng, K. Chen, Z. Liu, and C. C. Loy, “Transformer-based visual segmentation: A survey,” IEEE Transactions on Pattern Analysis and Machine Intelligence , 2024

    2024

  9. [17]

    Layered rendering diffusion model for zero-shot guided image synthesis,

    Z. Qi, G. Huang, Z. Huang, Q. Guo, J. Chen, J. Han, J. Wang, G. Zhang, L. Liu, E. Ding et al. , “Layered rendering diffusion model for zero-shot guided image synthesis,” arXiv preprint arXiv:2311.18435 , 2023

    2023 arXiv

  10. [18]

    Panoptic nerf: 3d-to-2d label transfer for panoptic urban scene segmentation,

    X. Fu, S. Zhang, T. Chen, Y . Lu, L. Zhu, X. Zhou, A. Geiger, and Y . Liao, “Panoptic nerf: 3d-to-2d label transfer for panoptic urban scene segmentation,” arXiv preprint arXiv:2203.15224, 2022

    2022 arXiv

  11. [19]

    In-place scene labelling and understanding with implicit scene representation,

    S. Zhi, T. Laidlow, S. Leutenegger, and A. J. Davison, 14 “In-place scene labelling and understanding with implicit scene representation,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2021, pp. 15 838–15 847

    2021

  12. [20]

    Implicit ray transformers for multiview remote sensing image segmentation,

    Z. Qi, H. Chen, C. Liu, Z. Shi, and Z. Zou, “Implicit ray transformers for multiview remote sensing image segmentation,” IEEE Transactions on Geoscience and Remote Sensing , vol. 61, pp. 1–15, 2023

    2023

  13. [21]

    Nerf: Representing scenes as neural radiance fields for view synthesis,

    B. Mildenhall, P. P. Srinivasan, M. Tancik, J. T. Barron, R. Ramamoorthi, and R. Ng, “Nerf: Representing scenes as neural radiance fields for view synthesis,” Communi- cations of the ACM , vol. 65, no. 1, pp. 99–106, 2021

    2021

  14. [22]

    Nerf++: Analyzing and improving neural radiance fields,

    K. Zhang, G. Riegler, N. Snavely, and V . Koltun, “Nerf++: Analyzing and improving neural radiance fields,” arXiv preprint arXiv:2010.07492 , 2020

    2010 arXiv

  15. [23]

    Mip-nerf 360: Unbounded anti-aliased neural radiance fields,

    J. T. Barron, B. Mildenhall, D. Verbin, P. P. Srinivasan, and P. Hedman, “Mip-nerf 360: Unbounded anti-aliased neural radiance fields,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 5470–5479

    2022

  16. [24]

    Sam 2: Segment anything in images and videos,

    N. Ravi, V . Gabeur, Y .-T. Hu, R. Hu, C. Ryali, T. Ma, H. Khedr, R. R ¨adle, C. Rolland, L. Gustafson et al. , “Sam 2: Segment anything in images and videos,” arXiv preprint arXiv:2408.00714, 2024

    2024 arXiv

  17. [25]

    3d gaussian splatting for real-time radiance field rendering,

    B. Kerbl, G. Kopanas, T. Leimk ¨uhler, and G. Drettakis, “3d gaussian splatting for real-time radiance field rendering,” 2023. [Online]. Available: https://arxiv.org/ abs/2308.04079

    2023 arXiv

  18. [26]

    Carla: An open urban driving simulator,

    A. Dosovitskiy, G. Ros, F. Codevilla, A. Lopez, and V . Koltun, “Carla: An open urban driving simulator,” in Conference on robot learning . PMLR, 2017, pp. 1–16

    2017

  19. [27]

    U-net: Convolutional networks for biomedical image segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-net: Convolutional networks for biomedical image segmentation,” 2015. [Online]. Available: https://arxiv.org/abs/1505.04597

    2015 arXiv

  20. [28]

    Multi-scale context aggregation by dilated convolutions,

    F. Yu and V . Koltun, “Multi-scale context aggregation by dilated convolutions,” 2016. [Online]. Available: https://arxiv.org/abs/1511.07122

    2016 arXiv

  21. [29]

    Enhanced feature pyramid network for semantic seg- mentation,

    M. Ye, J. Ouyang, G. Chen, J. Zhang, and X. Yu, “Enhanced feature pyramid network for semantic seg- mentation,” in 2020 25th International Conference on Pattern Recognition (ICPR) . IEEE, 2021, pp. 3209– 3216

    2020

  22. [30]

    Panoptic feature pyramid networks,

    A. Kirillov, R. Girshick, K. He, and P. Doll ´ar, “Panoptic feature pyramid networks,” 2019. [Online]. Available: https://arxiv.org/abs/1901.02446

    2019 arXiv

  23. [31]

    Feature pyramid network for multi-class land segmen- tation,

    S. Seferbekov, V . Iglovikov, A. Buslaev, and A. Shvets, “Feature pyramid network for multi-class land segmen- tation,” in Proceedings of the IEEE conference on com- puter vision and pattern recognition workshops , 2018, pp. 272–275

    2018

  24. [32]

    Not all noises are created equally: Diffusion noise selection and optimiza- tion,

    Z. Qi, L. Bai, H. Xiong, and Z. Xie, “Not all noises are created equally: Diffusion noise selection and optimiza- tion,” arXiv preprint arXiv:2407.14041 , 2024

    2024 arXiv

  25. [33]

    Dual attention network for scene segmentation,

    J. Fu, J. Liu, H. Tian, Y . Li, Y . Bao, Z. Fang, and H. Lu, “Dual attention network for scene segmentation,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2019, pp. 3146–3154

    2019

  26. [34]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is all you need,” Advances in neural informa- tion processing systems , vol. 30, 2017

    2017

  27. [35]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weis- senborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Min- derer, G. Heigold, S. Gelly et al. , “An image is worth 16x16 words: Transformers for image recognition at scale,” arXiv preprint arXiv:2010.11929 , 2020

    2010 arXiv

  28. [36]

    Segformer: Simple and efficient design for semantic segmentation with transformers,

    E. Xie, W. Wang, Z. Yu, A. Anandkumar, J. M. Alvarez, and P. Luo, “Segformer: Simple and efficient design for semantic segmentation with transformers,” Advances in neural information processing systems , vol. 34, pp. 12 077–12 090, 2021

    2021

  29. [37]

    Semsegdepth: A combined model for semantic segmentation and depth completion,

    J. P. Lagos and E. Rahtu, “Semsegdepth: A combined model for semantic segmentation and depth completion,” arXiv preprint arXiv:2209.00381 , 2022

    2022 arXiv

  30. [38]

    Hybridnet for depth estimation and semantic seg- mentation,

    D. Sanchez-Escobedo, X. Lin, J. R. Casas, and M. Par- das, “Hybridnet for depth estimation and semantic seg- mentation,” in 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) . IEEE, 2018, pp. 1563–1567

    2018

  31. [39]

    Remote sensing image segmentation based on implicit 3d scene represen- tation,

    Z. Qi, Z. Zou, H. Chen, and Z. Shi, “Remote sensing image segmentation based on implicit 3d scene represen- tation,” IEEE Geoscience and Remote Sensing Letters , 2022

    2022

  32. [40]

    Spin-nerf: Multiview segmentation and perceptual inpainting with neural radiance fields,

    A. Mirzaei, T. Aumentado-Armstrong, K. G. Derpa- nis, J. Kelly, M. A. Brubaker, I. Gilitschenski, and A. Levinshtein, “Spin-nerf: Multiview segmentation and perceptual inpainting with neural radiance fields,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pat...

    2023

  33. [41]

    Opennerf: Open set 3d neural scene segmentation with pixel-wise features and rendered novel views,

    F. Engelmann, F. Manhardt, M. Niemeyer, K. Tateno, M. Pollefeys, and F. Tombari, “Opennerf: Open set 3d neural scene segmentation with pixel-wise features and rendered novel views,” arXiv preprint arXiv:2404.03650, 2024

    2024 arXiv

  34. [42]

    3d reconstruction of remote sensing mountain areas with tsdf-based neural networks,

    Z. Qi, Z. Zou, H. Chen, and Z. Shi, “3d reconstruction of remote sensing mountain areas with tsdf-based neural networks,” Remote Sensing , vol. 14, no. 17, p. 4333, 2022

    2022

  35. [43]

    Deep learning for 3d point clouds: A survey,

    Y . Guo, H. Wang, Q. Hu, H. Liu, L. Liu, and M. Ben- namoun, “Deep learning for 3d point clouds: A survey,” IEEE transactions on pattern analysis and machine in- telligence, vol. 43, no. 12, pp. 4338–4364, 2020

    2020

  36. [44]

    Flashsplat: 2d to 3d gaussian splatting segmentation solved optimally,

    Q. Shen, X. Yang, and X. Wang, “Flashsplat: 2d to 3d gaussian splatting segmentation solved optimally,” in European Conference on Computer Vision . Springer, 2025, pp. 456–472

    2025

  37. [45]

    Gaussian grouping: Segment and edit anything in 3d scenes,

    M. Ye, M. Danelljan, F. Yu, and L. Ke, “Gaussian grouping: Segment and edit anything in 3d scenes,” in European Conference on Computer Vision . Springer, 2025, pp. 162–179

    2025

  38. [46]

    Click- gaussian: Interactive segmentation to any 3d gaussians,

    S. Choi, H. Song, J. Kim, T. Kim, and H. Do, “Click- gaussian: Interactive segmentation to any 3d gaussians,” in European Conference on Computer Vision . Springer, 2025, pp. 289–305

    2025

  39. [47]

    Segment anything,

    A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, 15 L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.- Y . Lo et al. , “Segment anything,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 4015–4026

    2023

  40. [48]

    Segment anything in high quality,

    L. Ke, M. Ye, M. Danelljan, Y .-W. Tai, C.-K. Tang, F. Yu et al. , “Segment anything in high quality,” Advances in Neural Information Processing Systems , vol. 36, 2024

    2024

  41. [49]

    Segment anything model for medical image analysis: an experimental study,

    M. A. Mazurowski, H. Dong, H. Gu, J. Yang, N. Konz, and Y . Zhang, “Segment anything model for medical image analysis: an experimental study,” Medical Image Analysis, vol. 89, p. 102918, 2023

    2023

  42. [50]

    Segment anything in 3d with nerfs,

    J. Cen, Z. Zhou, J. Fang, W. Shen, L. Xie, D. Jiang, X. Zhang, Q. Tian et al. , “Segment anything in 3d with nerfs,” Advances in Neural Information Processing Systems, vol. 36, pp. 25 971–25 990, 2023

    2023

  43. [51]

    Multi-view remote sensing image segmentation with sam priors,

    Z. Qi, C. Liu, Z. Liu, H. Chen, Y . Wu, Z. Zou, and Z. Shi, “Multi-view remote sensing image segmentation with sam priors,” in IGARSS 2024-2024 IEEE International Geoscience and Remote Sensing Symposium . IEEE, 2024, pp. 8446–8449

    2024

  44. [52]

    Structure-from- motion revisited,

    J. L. Schonberger and J.-M. Frahm, “Structure-from- motion revisited,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2016, pp. 4104–4113

    2016

  45. [53]

    Application of 3d gaussian splatting for cinematic anatomy on consumer class devices,

    S. Niedermayr, C. Neuhauser, K. Petkov, K. Engel, and R. Westermann, “Application of 3d gaussian splatting for cinematic anatomy on consumer class devices,” 2024

    2024

  46. [54]

    Gsedit: Efficient text-guided editing of 3d objects via gaussian splatting,

    F. Palandra, A. Sanchietti, D. Baieri, and E. Rodol `a, “Gsedit: Efficient text-guided editing of 3d objects via gaussian splatting,” arXiv preprint arXiv:2403.05154 , 2024

    2024 arXiv

  47. [55]

    Text-to-3d us- ing gaussian splatting,

    Z. Chen, F. Wang, Y . Wang, and H. Liu, “Text-to-3d us- ing gaussian splatting,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 21 401–21 412

    2024

  48. [56]

    Seg- net: A deep convolutional encoder-decoder architecture for image segmentation,

    V . Badrinarayanan, A. Kendall, and R. Cipolla, “Seg- net: A deep convolutional encoder-decoder architecture for image segmentation,” IEEE transactions on pattern analysis and machine intelligence , vol. 39, no. 12, pp. 2481–2495, 2017

    2017

  49. [57]

    U-net: Con- volutional networks for biomedical image segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-net: Con- volutional networks for biomedical image segmentation,” in International Conference on Medical image computing and computer-assisted intervention . Springer, 2015, pp. 234–241

    2015

  50. [58]

    Rethinking atrous convolution for semantic image seg- mentation,

    L.-C. Chen, G. Papandreou, F. Schroff, and H. Adam, “Rethinking atrous convolution for semantic image seg- mentation,” arXiv preprint arXiv:1706.05587 , 2017

    2017 arXiv

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.