Pith. sign in

REVIEW 5 major objections 6 minor 63 references

AGS-Mesh: Adaptive Gaussian Splatting and Meshing with Geometric Priors for Indoor Room Reconstruction Using Smartphones

T0 review · 5 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Filtering phone depth with monocular normals, and meshing with a depth-aware octree, sharpens indoor 3D reconstructions from Gaussian splatting.

desk verdict A sensible adaptive-filtering plug-in for Gaussian splatting that shows real gains, but the DNC normal estimation as written is a reproducibility bug and the meshing claim oversells. read the letter →

arxiv 2411.19271 v2 pith:SLWIU5XT submitted 2024-11-28 cs.CV

Xuqian Ren , Matias Turkulainen , Jiepeng Wang , Otto Seiskari , Iaroslav Melekhov , Juho Kannala , Esa Rahtu This is my paper · ORCID
classification cs.CV
keywords 3DreconstructionGaussiansplattingindoorscenesdepthpriorsnormalregularizationmeshextractionoctreemeshingsmartphoneLiDAR
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

Modern indoor 3D reconstruction from smartphones has a tension: phone LiDAR depth is accurate on flat surfaces but poor at edges and far objects, while monocular normal estimates from pretrained networks are sharper at boundaries but inconsistent across views. This paper argues that both signals can be used safely if they are filtered against each other during Gaussian Splatting optimization, rather than trusted wholesale. It introduces two supervision losses, Depth Normal Consistency and Adaptive Normal Regularization, and reports that they improve both mesh metrics and novel-view synthesis on two indoor room datasets captured with smartphones. It also proposes a scale-aware meshing pipeline combining a depth-adaptive truncated signed distance function with octree isosurface extraction to recover finer detail than standard TSDF or Poisson meshing. If these results hold, existing Gaussian Splatting pipelines can be upgraded with plug-in losses and a post-process mesher for more usable room reconstructions from consumer devices.

What carries the argument

The argument rests on two adaptive filters and one meshing stage. Depth Normal Consistency computes per-pixel normals from back-projected phone depth by K-nearest-neighbor covariance analysis, compares their angle with monocular normals, and zeroes the depth supervision wherever the angle exceeds a threshold of ten degrees by default, using raw depth early in training and filtered depth later. Adaptive Normal Regularization does the reverse for normals: since monocular normals are inconsistent across views, it first supervises rendered normals with them and, after a number of iterations, retains only monocular normals that agree with normals rendered from the current Gaussian geometry. The meshing machinery is a depth-aware truncated signed distance function, where the truncation distance grows linearly with depth so distant, less reliable observations are weighted less, followed by IsoOctree, an octree-based isosurface extraction that subdivides voxels according to back-projected point density and caps detail where the data do not support it. The isofunction also weights each depth map's contribution by normal-to-ray consistency, and the filters change what the optimization is supervised by, while the mesher changes how the optimized field is converted into a surface.

What would settle it

Measure the filter's accuracy directly: on a scene with a ground-truth scan, compute the fraction of depth pixels zeroed by Depth Normal Consistency that are actually within a few centimetres of the true surface; if most discarded pixels are correct, the arbitration is wrong. Alternatively, if removing both filters and using raw depth plus oracle normal supervision produces the same F-scores, the claim that consistency-based filtering causes the improvement is falsified.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes that the quality bottleneck in indoor Gaussian Splatting reconstruction is not the absence of geometric priors but the blind use of them. The central discovery is that comparing priors against each other, normals derived from noisy sensor depth versus normals from a monocular network, and monocular normals versus normals rendered from the optimized Gaussians, yields a self-consistent training signal that improves geometry and rendering simultaneously. Concretely, with a ten-degree angle threshold, the Depth Normal Consistency filter removes depth pixels that would drag Gaussian positions off thin structures and edges, while the Adaptive Normal Regularization term suppresses normal supervision where the prior is uncertain, avoiding baked-in multi-view inconsistencies. The paper further claims that geometry extracted from the optimized Gaussian scene is improved by its depth-aware truncated TSDF with IsoOctree extraction, which adapts voxel resolution to point-cloud density and therefore preserves fine detail while smoothing walls. The reported gains are on real smartphone captures from two indoor room datasets, against baselines including volumetric fusion, implicit SDF methods, DN-Splatter, 2DGS, and 3DGS variants, and the method is presented as a plug-in that works for both 3D and 2D Gaussian representations.

Load-bearing premise

The load-bearing premise is that when a normal computed from the phone depth disagrees with a monocular normal estimate by more than ten degrees, the phone depth is the signal to discard, but monocular normals are themselves estimates with their own errors, so a wrong monocular normal would make the filter throw away correct depth.

Editorial extensions

If this is right

  • Existing 3D Gaussian Splatting and 2D Gaussian Splatting implementations can add both losses on top of their current training loop, and the paper reports better mesh F-scores and Chamfer distances on room-scale smartphone data without sacrificing rendering quality.
  • Mesh extraction from a Gaussian scene no longer requires a global voxel resolution: the depth-adaptive TSDF and IsoOctree pipeline recovers fine detail where point density is high and stays smooth elsewhere, at a much smaller mesh size on the reported room scene.
  • Because the filters depend only on comparing two geometric signals, the method generalizes across 3D and 2D Gaussian variants and is not tied to a particular splatting architecture.
  • Sensor depth remains load-bearing: the ablation shows raw phone depth supervision greatly outperforms monocular depth supervision, so the method's improvements do not eliminate the need for a physical depth sensor.
  • Novel-view synthesis from unseen camera trajectories improves with the filtered priors, indicating that the regularization reduces floaters and multi-view inconsistency, not just surface quality.

Reading between the lines

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

  • Editorial inference: the same mutual-consistency filtering could be applied to RGB-only pipelines by using monocular depth as a stand-in for sensor depth, but the paper's own ablation suggests this would be a large quality drop unless the monocular depth is itself calibrated to metric scale.
  • Editorial inference: the ten-degree thresholds are fixed, and a testable extension would be to make them per-pixel or spatially adaptive, since the paper's own figures show errors concentrate at edges and far surfaces.
  • Editorial inference: the IsoOctree subdivision heuristic, which expands a voxel when it contains at least fifty back-projected points, ties detail to view coverage; in scenes with highly uneven camera trajectories this might over-smooth rarely seen but geometrically important structures.
  • Editorial inference: if the Depth Normal Consistency arbitration were wrong on some scenes, the method could still improve on average while discarding correct depth there, so measuring the filter's precision and recall against ground-truth scans would quantify that risk directly.
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 / 6 minor

Summary. The paper proposes AGS-Mesh, a plug-in supervision strategy for Gaussian Splatting (3DGS and 2DGS) that adaptively filters noisy iPhone depth maps and monocular normal priors during optimization, plus a depth-adaptive TSDF/IsoOctree meshing post-process. Depth Normal Consistency (DNC) compares normals derived from local depth patches with Omnidata monocular normals to suppress unreliable depth, and Adaptive Normal Regularization (ANR) suppresses normal supervision where rendered normals disagree with monocular predictions. Experiments on MuSHRoom (6 scenes) and ScanNet++ (2 scenes) report improved mesh reconstruction metrics and novel-view synthesis over several baselines, including DN-Splatter, 2DGS, and Splatfacto. The main contribution is the adaptive filtering scheme; the meshing contribution is a separate post-processing strategy.

Significance. If the filtering behavior is implemented as intended, the method is a practical and potentially useful plug-in for improving geometric quality of Gaussian Splatting reconstructions from noisy smartphone captures, and the release of code is a strength. The paper also shows that adding the regularizers reduces Gaussian count while improving or maintaining rendering quality, which is a nontrivial and useful observation. However, the significance is currently conditional because the written description of the core filtering mechanism contains technical inconsistencies that prevent reproduction, the ScanNet++ evaluation is limited to two scenes with no variance reporting, and the authors' own limitation statement (Appendix E) undercuts the abstract's meshing claim. These issues need to be resolved before the central claims can be accepted at face value.

major comments (5)
  1. [4.1, Eq. (5)] The normal estimator Nd is described as 'the maximum eigenvector corresponding to the maximum eigenvalue' of the covariance of KNN points. For a locally planar depth patch, the surface normal is the eigenvector of the smallest eigenvalue; the maximum eigenvector lies in the tangent plane. With τd = 10° (Supp. A.1), the angle θd between a tangent-direction eigenvector and a monocular normal would be near 90° on planar regions, so DNC would mask essentially all depths after step Td = 7k, making the large gains in Table 3 impossible under the stated formulation. If the released code actually uses the minimum eigenvector, the text must be corrected; if it uses the maximum eigenvector, the method cannot work as described. This is load-bearing because DNC is one of the two core contributions.
  2. [4.1, Eq. (4) and 4.2, Eq. (8)] Equations (4) and (8) set the filtered depth and normal targets to 0, while the losses in Eqs. (6) and (7) are L1 losses. Under an L1 loss of the form ∥D̂ - 0∥1 or ∥N̂ - 0∥1, masked pixels are pulled toward zero rather than excluded from supervision, which would corrupt rendered depth and normal maps at boundaries and edges—exactly the regions the method claims to filter. The manuscript must specify that a binary mask drops the loss at filtered pixels (or equivalently that the loss is defined only on valid pixels). As written, the mechanism does not implement filtering and is not reproducible.
  3. [Table 5 and Section 5.1] The ScanNet++ evaluation in Table 5 is averaged over only two scenes, and no error bars or significance tests are reported for any of the tables. The abstract claims 'significant improvements' on 'challenging indoor room datasets'; with n = 2 and no variance or per-scene breakdown, the reader cannot judge whether the reported improvements are meaningful. Per-scene results for ScanNet++ and, ideally, error bars or significance tests for MuSHRoom should be provided, or the claim should be tempered to reflect the limited evaluation.
  4. [Appendix E] Appendix E states: 'The IsoOctree meshing technique we propose focuses on reducing the number of vertices and faces in the mesh while smoothing the surface. However, it does not consistently enhance the overall quality of 3D reconstructions.' This directly qualifies the abstract's claim that the meshing strategy 'recovers finer details from Gaussian models compared to other commonly used open-source meshing tools' and the main-text assertion of 'superior surface creation quality' in Section 5.1. The meshing claim should be narrowed to what the data actually support: the IsoOctree method reduces mesh complexity and can smooth surfaces, but its geometric accuracy is not consistently better than existing TSDF extraction.
  5. [4.2, Eq. (8)] The ANR filter uses normals rendered from the current Gaussian geometry (N̂) to decide which monocular normals (Np) to trust. This is a self-referential loop: early geometry errors can cause the filter to reject correct monocular normals or retain incorrect ones, potentially entrenching the errors. The paper does not provide a diagnostic for this behavior. I would like to see a quantitative report of the fraction of pixels filtered by ANR over training (and, if available, a validation of the filter decisions against the ground-truth normals that ScanNet++ provides), to show that the loop does not simply reinforce the initial Gaussian geometry.
minor comments (6)
  1. [Section 2] The word 'inconsitencies' should be 'inconsistencies'.
  2. [Section 4.2] In the sentence following Eq. (8), the symbol θi should be θn to match the definition in Eq. (5) and the notation θn used earlier.
  3. [Figure 8 caption] The caption says 'Depth Normal Consistency (DNR)' but the method is abbreviated DNC; this typo should be fixed.
  4. [Supp. A.1] The text 'We enable our DNR and ANR optimization terms' appears to use 'DNR' where 'DNC' is intended.
  5. [Section 4.1] The KNN-based normal estimation uses k = 200 world-coordinate neighbors, but no details are given about how the neighborhood is weighted or how noise in the depth map affects the covariance estimate; a brief clarification of the robust estimation would aid reproducibility.
  6. [Table 1] The table reports Gaussian counts for a single scene ('vr room'); stating which scene and why this particular count is representative would help the reader interpret the 'reduced Gaussian count' claim.

Circularity Check

1 steps flagged · score 3.0 of 10

ANR's filtered normal target is defined by the very Gaussian geometry it supervises; otherwise the paper's comparisons are external and not circular.

  1. self definitional [Section 4.2, Eqs. (7)-(9)]
    "where Nf are filtered normals calculated by comparing the angle difference between ˆN and Np with a threshold τN . Nf = ( 0 if θn > τN Np otherwise (8) where θi is the angle difference between ˆN and Np like Eq. (5)."

    Eq. (8) defines the filtered normal target Nf as a function of ˆN, the normal rendered from the Gaussian scene that is being optimized (Eq. 9 uses rendered depth to compute ˆN). Eq. (7) then trains ˆN to match Nf. Thus the normal 'prior' is not independent: the model selects which monocular normals to trust based on its own current geometry, and the loss fits that same geometry to the selected targets. A wrong but self-consistent early geometry will cause correct Np values that disagree to be zeroed, and the L1 loss is minimized without correcting the error. The Table 3 improvement attributed to ANR is therefore partly a self-reinforcement effect rather than confirmation from an external prior; the loop is not a first-principles derivation.

full rationale

The paper's final evaluation is against external benchmarks (MuSHRoom, ScanNet++) with ground-truth scans, so the overall improvement claims are externally testable and largely not circular. The single clear circular step is ANR: its filtered normal target Nf is defined via Eq. (8) as a function of the current rendered normal ˆN, which is then supervised by that same target in Eq. (7). This is a feedback loop that can entrench early geometry errors, and the paper provides no ground-truth validation of the arbitration between monocular and rendered normals. DNC, by contrast, filters sensor depth using only external inputs (sensor depth and Omnidata normals), so it is not circular; the potential issue there is correctness/reproducibility (the text calls the maximum PCA eigenvector the normal, which is the tangent direction for a planar patch, and Eqs. (4)/(8) set filtered targets to 0 rather than masking the loss), and I do not score those as circularity. The paper also builds on the authors' DN-Splatter and MuSHRoom, but these are open baselines/datasets and the comparisons are against them, so no load-bearing self-citation chain is present. Appendix E's limitation that IsoOctree 'does not consistently enhance the overall quality of 3D reconstructions' tempers the meshing claim but is an honest limitation, not circularity. Overall score 3: one central self-referential mechanism, but the pipeline still contains independent external supervision and external evaluation.

Assumptions & free parameters 9 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new physical entities, particles, or forces. Its contribution rests on hand-set thresholds and heuristics plus domain assumptions about sensor and monocular prior reliability. The most fragile assumptions are the arbitrating role of Omnidata normals in DNC and the self-referential filtering in ANR. The meshing method adds further heuristics (point-density subdivision, depth-relative truncation) that are plausible but untested for sensitivity.

free parameters (9)
  • DNC angle threshold tau_d = 10 degrees
    Used in Eq. (4) to discard sensor depth points whose derived normal deviates from monocular normal; hand-set, no sensitivity analysis.
  • ANR angle threshold tau_N = 10 degrees
    Used in Eq. (8) to drop monocular normals that disagree with rendered normals; hand-set, no sensitivity analysis.
  • Depth supervision weight lambda_d = 0.2
    Weight for the depth loss in Eq. (10), chosen by hand.
  • Normal supervision weight lambda_n = 0.1
    Weight for the normal loss in Eq. (10), chosen by hand.
  • Iteration schedule T_d, T_n = T_d = 7k, T_n = 15k
    Iterations at which DNC and ANR filtering are enabled; Appendix A.1 also reads inconsistently as '7k iterations' for normal regularization.
  • KNN neighbor count k = 200
    Number of neighbors used to estimate normals from depth in Section 4.1; hand-selected.
  • IsoOctree subdivision threshold N_e = 50
    Point count within a voxel that triggers octree subdivision in Section 4.4; user-defined heuristic.
  • Relative TSDF truncation distance tau = 0.05
    Depth-relative truncation distance in Appendix A.2; no sensitivity study reported.
  • Maximum octree depth = 10
    Maximum subdivision depth for IsoOctree in Appendix A.2.
assumptions (5)
  • domain assumption Low-resolution phone depth is accurate on planar and smooth surfaces but unreliable for edges, thin structures, and far surfaces.
    Motivates filtering in Section 4.1 and is the premise for why DNC helps; supported qualitatively by Fig. 2, not by quantitative per-scene analysis.
  • domain assumption Omnidata monocular normals have clear object boundaries and can be used as a reference for filtering depth.
    Relied on by DNC in Section 4.1; no independent validation on these datasets shows that Omnidata boundaries are reliable enough for arbitration.
  • ad hoc to paper Rendered normals from the current Gaussian scene are a reliable arbiter for accepting or rejecting monocular normal priors.
    ANR Eq. (8) compares rendered normals with Np; this is a self-consistency criterion that can entrench errors and is not externally grounded.
  • ad hoc to paper The appropriate level of detail in the extracted mesh is proportional to backprojected rendered depth point-cloud density.
    Used as IsoOctree subdivision criterion in Section 4.4; plausible heuristic but not derived or validated.
  • ad hoc to paper Depth uncertainty scales roughly linearly with depth, so TSDF truncation distance should grow with depth.
    Depth-aware truncation in Section 4.4 and Appendix A.2; chosen with tau = 0.05, no sensitivity study.

how reviews work

0 comments
Cite this review

Pith. "Pith review of AGS-Mesh: Adaptive Gaussian Splatting and Meshing with Geometric Priors for Indoor Room Reconstruction Using Smartphones." pith.science (2026). https://pith.science/paper/SLWIU5XT

@misc{pith2026241119271,
  author       = {Pith},
  title        = {Pith review of: AGS-Mesh: Adaptive Gaussian Splatting and Meshing with Geometric Priors for Indoor Room Reconstruction Using Smartphones},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SLWIU5XT}},
  note         = {Machine review of arXiv:2411.19271}
}
read the original abstract

Geometric priors are often used to enhance 3D reconstruction. With many smartphones featuring low-resolution depth sensors and the prevalence of off-the-shelf monocular geometry estimators, incorporating geometric priors as regularization signals has become common in 3D vision tasks. However, the accuracy of depth estimates from mobile devices is typically poor for highly detailed geometry, and monocular estimators often suffer from poor multi-view consistency and precision. In this work, we propose an approach for joint surface depth and normal refinement of Gaussian Splatting methods for accurate 3D reconstruction of indoor scenes. We develop supervision strategies that adaptively filters low-quality depth and normal estimates by comparing the consistency of the priors during optimization. We mitigate regularization in regions where prior estimates have high uncertainty or ambiguities. Our filtering strategy and optimization design demonstrate significant improvements in both mesh estimation and novel-view synthesis for both 3D and 2D Gaussian Splatting-based methods on challenging indoor room datasets. Furthermore, we explore the use of alternative meshing strategies for finer geometry extraction. We develop a scale-aware meshing strategy inspired by TSDF and octree-based isosurface extraction, which recovers finer details from Gaussian models compared to other commonly used open-source meshing tools. Our code is released in https://xuqianren.github.io/ags_mesh_website/.

Figures

Figures reproduced from arXiv: 2411.19271 by the authors.

Figure 1
Figure 1. We present AGS-Mesh, a method that adaptively integrates geometric priors into Gaussian Splatting for indoor room reconstruction [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Demonstration of iPhone and Kinect sensor depths. The iPhone struggles to capture accurate depth values for (a) objects at a far distance, and (b) small objects, and edges. Instead, the Kinect sensor [1] (c) is able to filter unconfident depth values (ToF) sensor with a 1024 × 1024 resolution with a per-pixel error range from 1.4mm to 12mm [27]. Depths are denoised through post-processing by considering systematic a… view at source ↗
Figure 3
Figure 3. Pipeline Overview. Our approach leverages geometric consistency between normals derived from raw sensor depths and those predicted by a pretrained model to filter out noisy sensor depth data. Likewise, we compare rendered normals from a Gaussian scene with pseudo-normal estimates to dynamically filter uncertainties in normal supervision during optimization. Our adaptive depth and normal regularization terms assist v… view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: We demonstrate our method with two Gaussian-based methods DN-Splatter [ [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Novel view synthesis comparisons on the MuSHRoom dataset. From left to right: 2DGS [ [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Qualitative comparison with our optimization strat￾egy. IsoOctree mesh extraction can efficiently generate a smoother surface compared to the TSDF [61] baseline. Furthermore, our adaptive regularization terms help mitigate misleading reconstruc￾tion caused by inaccurat…
Figure 7
Figure 7. Figure 7: Qualitative mesh comparison for the ”8b5caf3398” scene from ScanNet++ dataset. [PITH_FULL_IMAGE:figures/full_fig_p015_7.png]
Figure 8
Figure 8. Figure 8: Qualitative visuals of our Depth Normal Consistency (DNR) and Adaptive Normal Regularization (ANR) terms. We visualize [PITH_FULL_IMAGE:figures/full_fig_p016_8.png]
Figure 9
Figure 9. Figure 9: Novel view synthesis comparisons on the MuSHRoom dataset. From left to right: 2DGS [ [PITH_FULL_IMAGE:figures/full_fig_p017_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

63 extracted references · 50 canonical work pages

  1. [1]

    https:// learn.microsoft.com/en- us/azure/kinect- dk/depth-camera

    Azure kinect dk depth camera documentation. https:// learn.microsoft.com/en- us/azure/kinect- dk/depth-camera. 3

    work page

  2. [2]

    Neural rgb-d surface reconstruction

    Dejan Azinovi´c, Ricardo Martin-Brualla, Dan B Goldman, Matthias Nießner, and Justus Thies. Neural rgb-d surface reconstruction. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6290– 6301, 2022. 2

    work page 2022

  3. [3]

    Mip-nerf 360: Unbounded anti-aliased neural radiance fields, 2022

    Jonathan T Barron, Ben Mildenhall, Dor Verbin, Pratul P Srinivasan, and Peter Hedman. Mip-nerf 360: Unbounded anti-aliased neural radiance fields, 2022. 2

    work page 2022

  4. [4]

    Zoedepth: Zero-shot transfer by com- bining relative and metric depth, 2023

    Shariq Farooq Bhat, Reiner Birkl, Diana Wofk, Peter Wonka, and Matthias M¨uller. Zoedepth: Zero-shot transfer by com- bining relative and metric depth, 2023. 3, 8

    work page 2023

  5. [5]

    Apple iphone 13 pro lidar accu- racy assessment for engineering applications

    Payton Chase, Kianna Clarke, Audrey Hawkes, Shabnam Jabari, and Jakov Jakus. Apple iphone 13 pro lidar accu- racy assessment for engineering applications. Transforming Construction with Reality Capture Technologies, 2022. 3

    work page 2022

  6. [6]

    Vcr-gaus: View consistent depth- normal regularizer for gaussian surface reconstruction

    Hanlin Chen, Fangyin Wei, Chen Li, Tianxin Huang, Yunsong Wang, and Gim Hee Lee. Vcr-gaus: View consistent depth- normal regularizer for gaussian surface reconstruction. arXiv preprint arXiv:2406.05774, 2024. 3

    arXiv 2024

  7. [7]

    Tmo: Textured mesh acquisition of objects with a mobile device by using differentiable rendering

    Jaehoon Choi, Dongki Jung, Taejae Lee, Sangwook Kim, Youngdong Jung, Dinesh Manocha, and Donghwan Lee. Tmo: Textured mesh acquisition of objects with a mobile device by using differentiable rendering. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 16674–16684, 2023. 3

    work page 2023

  8. [8]

    Nearest neighbor pattern classification

    Thomas Cover and Peter Hart. Nearest neighbor pattern classification. IEEE transactions on information theory, 13 (1):21–27, 1967. 4

    work page 1967

Show all 63 references
  1. [9]

    A volumetric method for building complex models from range images

    Brian Curless and Marc Levoy. A volumetric method for building complex models from range images. In Proceedings of the 23rd annual conference on Computer graphics and interactive techniques, pages 303–312, 1996. 2, 6, 7

    1996

  2. [10]

    High-quality surface reconstruction using gaussian surfels

    Pinxuan Dai, Jiamin Xu, Wenxiang Xie, Xinguo Liu, Huamin Wang, and Weiwei Xu. High-quality surface reconstruction using gaussian surfels. In ACM SIGGRAPH 2024 Conference Papers, pages 1–11, 2024. 3, 2

    2024

  3. [11]

    Depth-supervised nerf: Fewer views and faster training for free

    Kangle Deng, Andrew Liu, Jun-Yan Zhu, and Deva Ramanan. Depth-supervised nerf: Fewer views and faster training for free. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition, pages 12882–12891,

  4. [12]

    Depth-supervised NeRF: Fewer views and faster training for free

    Kangle Deng, Andrew Liu, Jun-Yan Zhu, and Deva Ramanan. Depth-supervised NeRF: Fewer views and faster training for free. In CVPR, 2022. 6

    2022

  5. [13]

    Transmvs- net: Global context-aware multi-view stereo network with transformers

    Yikang Ding, Wentao Yuan, Qingtian Zhu, Haotian Zhang, Xiangyue Liu, Yuanjiang Wang, and Xiao Liu. Transmvs- net: Global context-aware multi-view stereo network with transformers. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8585–8594,

  6. [14]

    Omnidata: A scalable pipeline for making multi-task mid-level vision datasets from 3d scans

    Ainaz Eftekhar, Alexander Sax, Jitendra Malik, and Amir Zamir. Omnidata: A scalable pipeline for making multi-task mid-level vision datasets from 3d scans. In ICCV, pages 10786–10796, 2021. 4, 1, 5

    2021

  7. [15]

    Multi-view stereo for com- munity photo collections

    Michael Goesele, Noah Snavely, Brian Curless, Hugues Hoppe, and Steven M Seitz. Multi-view stereo for com- munity photo collections. In 2007 IEEE 11th International Conference on Computer Vision, pages 1–8. IEEE, 2007. 2

    2007

  8. [16]

    Sugar: Surface-aligned gaussian splatting for efficient 3d mesh reconstruction and high-quality mesh rendering, 2023

    Antoine Gu´edon and Vincent Lepetit. Sugar: Surface-aligned gaussian splatting for efficient 3d mesh reconstruction and high-quality mesh rendering, 2023. 2, 3, 5, 6, 7

    2023

  9. [17]

    Rgb-d mapping: Using kinect-style depth cameras for dense 3d modeling of indoor environments

    Peter Henry, Michael Krainin, Evan Herbst, Xiaofeng Ren, and Dieter Fox. Rgb-d mapping: Using kinect-style depth cameras for dense 3d modeling of indoor environments. The international journal of Robotics Research, 31(5):647–663,

  10. [18]

    Metric3d v2: A versatile monocular geomet- ric foundation model for zero-shot metric depth and surface normal estimation

    Mu Hu, Wei Yin, Chi Zhang, Zhipeng Cai, Xiaoxiao Long, Hao Chen, Kaixuan Wang, Gang Yu, Chunhua Shen, and Shaojie Shen. Metric3d v2: A versatile monocular geomet- ric foundation model for zero-shot metric depth and surface normal estimation. arXiv preprint arXiv:2404.15506, 2024. 3

    2024 arXiv

  11. [19]

    2d gaussian splatting for geometrically accu- rate radiance fields

    Binbin Huang, Zehao Yu, Anpei Chen, Andreas Geiger, and Shenghua Gao. 2d gaussian splatting for geometrically accu- rate radiance fields. In SIGGRAPH 2024 Conference Papers. Association for Computing Machinery, 2024. 1, 2, 3, 4, 5, 6, 7, 8

    2024

  12. [20]

    Newcombe, Pushmeet Kohli, Jamie Shotton, Steve Hodges, Dustin Freeman, Andrew J

    Shahram Izadi, David Kim, Otmar Hilliges, David Molyneaux, Richard A. Newcombe, Pushmeet Kohli, Jamie Shotton, Steve Hodges, Dustin Freeman, Andrew J. Davison, and Andrew William Fitzgibbon. Kinectfusion: real-time 3d reconstruction and interaction using a moving depth camera....

    2011

  13. [21]

    James T. Kajiya. The rendering equation. In Proceedings of the 13th Annual Conference on Computer Graphics and Interactive Techniques, page 143–150, New York, NY , USA,

  14. [22]

    Screened poisson surface reconstruction

    Michael Kazhdan and Hugues Hoppe. Screened poisson surface reconstruction. ACM Trans. Graph., 32(3), 2013. 2

    2013

  15. [23]

    Pois- son surface reconstruction

    Michael Kazhdan, Matthew Bolitho, and Hugues Hoppe. Pois- son surface reconstruction. In Proceedings of the fourth Eu- rographics symposium on Geometry processing, 2006. 3

    2006

  16. [24]

    Unconstrained isosurface extraction on arbitrary oc- trees

    Michael Kazhdan, Allison Klein, Ketan Dalal, and Hugues Hoppe. Unconstrained isosurface extraction on arbitrary oc- trees. In Proceedings of the Fifth Eurographics Symposium on Geometry Processing, page 125–133, Goslar, DEU, 2007. Eurographics Association. 2, 6, 1

    2007

  17. [25]

    Unconstrained isosurface extraction on arbitrary oc- trees

    Michael Kazhdan, Allison Klein, Ketan Dalal, and Hugues Hoppe. Unconstrained isosurface extraction on arbitrary oc- trees. In Symposium on Geometry Processing, 2007. 3, 5

    2007

  18. [26]

    3d gaussian splatting for real-time radiance field rendering

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimk¨uhler, and George Drettakis. 3d gaussian splatting for real-time radiance field rendering. ACM TOG, 42(4), 2023. 1, 2, 3, 4, 5, 6, 7

    2023

  19. [27]

    Accuracy and resolution of kinect depth data for indoor mapping appli- cations

    Kourosh Khoshelham and Sander Oude Elberink. Accuracy and resolution of kinect depth data for indoor mapping appli- cations. sensors, 12(2):1437–1454, 2012. 3

    2012

  20. [28]

    Marching cubes: A high resolution 3d surface construction algorithm

    William E Lorensen and Harvey E Cline. Marching cubes: A high resolution 3d surface construction algorithm. In Seminal graphics: pioneering efforts that shaped the field, pages 347–

  21. [29]

    Melekhov, J

    I. Melekhov, J. Kannala, and E. Rahtu. Image patch matching using convolutional descriptors with euclidean distance. In Proc. ACCVW, 2016. 2

    2016

  22. [30]

    Real-time visibility-based fusion of depth maps

    Paul Merrell, Amir Akbarzadeh, Liang Wang, Philippos Mor- dohai, Jan-Michael Frahm, Ruigang Yang, David Nist´er, and Marc Pollefeys. Real-time visibility-based fusion of depth maps. In 2007 IEEE 11th International Conference on Com- puter Vision, pages 1–8. Ieee, 2007. 2

    2007

  23. [31]

    Srinivasan, Matthew Tancik, Jonathan T

    Ben Mildenhall, Pratul P. Srinivasan, Matthew Tancik, Jonathan T. Barron, Ravi Ramamoorthi, and Ren Ng. Nerf: Representing scenes as neural radiance fields for view synthe- sis. In ECCV, 2020. 2

    2020

  24. [32]

    Instant neural graphics primitives with a multires- olution hash encoding

    Thomas M¨uller, Alex Evans, Christoph Schied, and Alexan- der Keller. Instant neural graphics primitives with a multires- olution hash encoding. ACM Transactions on Graphics, 41 (4):1–15, 2022. 2

    2022

  25. [33]

    H2O-SDF: Two-phase learning for 3d indoor reconstruction using object surface fields, 2024

    Minyoung Park, Mirae Do, YeonJae Shin, Jaeseok Yoo, Jongkwang Hong, Joongrock Kim, and Chul Lee. H2O-SDF: Two-phase learning for 3d indoor reconstruction using object surface fields, 2024. 3

    2024

  26. [34]

    Infinitam v3: A framework for large- scale 3d reconstruction with loop closure

    Victor Adrian Prisacariu, Olaf K ¨ahler, Stuart Golodetz, Michael Sapienza, Tommaso Cavallari, Philip HS Torr, and David W Murray. Infinitam v3: A framework for large- scale 3d reconstruction with loop closure. arXiv preprint arXiv:1708.00783, 2017. 2

    2017 arXiv

  27. [35]

    Stopthe- pop: Sorted gaussian splatting for view-consistent real-time rendering, 2024

    Lukas Radl, Michael Steiner, Mathias Parger, Alexander Weinrauch, Bernhard Kerbl, and Markus Steinberger. Stopthe- pop: Sorted gaussian splatting for view-consistent real-time rendering, 2024. 4

    2024

  28. [36]

    Nerfmeshing: Distilling neural radiance fields into geometrically-accurate 3d meshes

    Marie-Julie Rakotosaona, Fabian Manhardt, Diego Martin Arroyo, Michael Niemeyer, Abhijit Kundu, and Federico Tombari. Nerfmeshing: Distilling neural radiance fields into geometrically-accurate 3d meshes. In Proc. of the Interna- tional Conf. on 3D Vision (3DV), 2024. 2

    2024

  29. [37]

    Mushroom: Multi-sensor hybrid room dataset for joint 3d reconstruction and novel view synthesis

    Xuqian Ren, Wenjia Wang, Dingding Cai, Tuuli Tuominen, Juho Kannala, and Esa Rahtu. Mushroom: Multi-sensor hybrid room dataset for joint 3d reconstruction and novel view synthesis. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 4508–...

    2024

  30. [38]

    Octnetfusion: Learning depth fusion from data

    Gernot Riegler, Ali Osman Ulusoy, Horst Bischof, and An- dreas Geiger. Octnetfusion: Learning depth fusion from data. In 2017 International Conference on 3D Vision (3DV), pages 57–66. IEEE, 2017. 2

    2017

  31. [39]

    Dense depth priors for neural radiance fields from sparse input views

    Barbara Roessle, Jonathan T Barron, Ben Mildenhall, Pratul P Srinivasan, and Matthias Nießner. Dense depth priors for neural radiance fields from sparse input views. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 12892–12901, 2022. 2

    2022

  32. [40]

    Deep marching tetrahedra: a hybrid represen- tation for high-resolution 3d shape synthesis

    Tianchang Shen, Jun Gao, Kangxue Yin, Ming-Yu Liu, and Sanja Fidler. Deep marching tetrahedra: a hybrid represen- tation for high-resolution 3d shape synthesis. Advances in Neural Information Processing Systems, 34:6087–6101, 2021. 3, 2

    2021

  33. [41]

    Nerfstudio: A modular framework for neural radiance field development

    Matthew Tancik, Ethan Weber, Evonne Ng, Ruilong Li, Brent Yi, Justin Kerr, Terrance Wang, Alexander Kristoffersen, Jake Austin, Kamyar Salahi, Abhik Ahuja, David McAllister, and Angjoo Kanazawa. Nerfstudio: A modular framework for neural radiance field development. In ACM SIGG...

    2023

  34. [42]

    Delicate textured mesh recovery from nerf via adaptive surface refinement

    Jiaxiang Tang, Hang Zhou, Xiaokang Chen, Tianshu Hu, Er- rui Ding, Jingdong Wang, and Gang Zeng. Delicate textured mesh recovery from nerf via adaptive surface refinement. arXiv preprint arXiv:2303.02091, 2022. 2

    2022 arXiv

  35. [43]

    Dn-splatter: Depth and normal priors for gaussian splatting and meshing

    Matias Turkulainen, Xuqian Ren, Iaroslav Melekhov, Otto Seiskari, Esa Rahtu, and Juho Kannala. Dn-splatter: Depth and normal priors for gaussian splatting and meshing. arXiv preprint arXiv:2403.17822, 2024. 1, 2, 3, 4, 6, 7

    2024 arXiv

  36. [44]

    High accuracy and visibility-consistent dense multiview stereo.IEEE transactions on pattern analysis and machine intelligence, 34(5):889–901, 2011

    Hoang-Hiep Vu, Patrick Labatut, Jean-Philippe Pons, and Renaud Keriven. High accuracy and visibility-consistent dense multiview stereo.IEEE transactions on pattern analysis and machine intelligence, 34(5):889–901, 2011. 2

    2011

  37. [45]

    Go- surf: Neural feature grid optimization for fast, high-fidelity rgb-d surface reconstruction

    Jingwen Wang, Tymoteusz Bleja, and Lourdes Agapito. Go- surf: Neural feature grid optimization for fast, high-fidelity rgb-d surface reconstruction. In 2022 International Confer- ence on 3D Vision (3DV) , pages 433–442. IEEE, 2022. 2, 6

    2022

  38. [46]

    Neuris: Neural reconstruction of indoor scenes using normal priors

    Jiepeng Wang, Peng Wang, Xiaoxiao Long, Christian Theobalt, Taku Komura, Lingjie Liu, and Wenping Wang. Neuris: Neural reconstruction of indoor scenes using normal priors. In European Conference on Computer Vision, pages 139–155. Springer, 2022. 3

    2022

  39. [47]

    Neus: Learning neural implicit surfaces by volume rendering for multi-view reconstruction

    Peng Wang, Lingjie Liu, Yuan Liu, Christian Theobalt, Taku Komura, and Wenping Wang. Neus: Learning neural implicit surfaces by volume rendering for multi-view reconstruction. arXiv preprint arXiv:2106.10689, 2021. 2

    2021 arXiv

  40. [48]

    Neus2: Fast learning of neural implicit surfaces for multi-view reconstruction

    Yiming Wang, Qin Han, Marc Habermann, Kostas Daniilidis, Christian Theobalt, and Lingjie Liu. Neus2: Fast learning of neural implicit surfaces for multi-view reconstruction. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 3295–3306, 2023. 2

    2023

  41. [49]

    Neuralfusion: Online depth fusion in latent space

    Silvan Weder, Johannes L Schonberger, Marc Pollefeys, and Martin R Oswald. Neuralfusion: Online depth fusion in latent space. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3162–3172,

  42. [50]

    Nerfingmvs: Guided optimization of neural radiance fields for indoor multi-view stereo

    Yi Wei, Shaohui Liu, Yongming Rao, Wang Zhao, Jiwen Lu, and Jie Zhou. Nerfingmvs: Guided optimization of neural radiance fields for indoor multi-view stereo. In ICCV, pages 5610–5619, 2021. 2

    2021

  43. [51]

    Debsdf: Delving into the details and bias of neural indoor scene reconstruction

    Yuting Xiao, Jingwei Xu, Zehao Yu, and Shenghua Gao. Debsdf: Delving into the details and bias of neural indoor scene reconstruction. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024. 3

    2024

  44. [52]

    Sparsegs: Real-time 360° sparse view synthesis using gaussian splatting

    Haolin Xiong, Sairisheek Muttukuru, Rishi Upadhyay, Pradyumna Chari, and Achuta Kadambi. Sparsegs: Real-time 360° sparse view synthesis using gaussian splatting. Arxiv,

  45. [53]

    Depth anything: Unleashing the power of large-scale unlabeled data

    Lihe Yang, Bingyi Kang, Zilong Huang, Xiaogang Xu, Jiashi Feng, and Hengshuang Zhao. Depth anything: Unleashing the power of large-scale unlabeled data. arXiv:2401.10891,

    arXiv

  46. [54]

    V ol- ume rendering of neural implicit surfaces.Advances in Neural Information Processing Systems, 34:4805–4815, 2021

    Lior Yariv, Jiatao Gu, Yoni Kasten, and Yaron Lipman. V ol- ume rendering of neural implicit surfaces.Advances in Neural Information Processing Systems, 34:4805–4815, 2021. 2 10

    2021

  47. [55]

    Scannet++: A high-fidelity dataset of 3d indoor scenes

    Chandan Yeshwanth, Yueh-Cheng Liu, Matthias Nießner, and Angela Dai. Scannet++: A high-fidelity dataset of 3d indoor scenes. In ICCV, 2023. 2, 6

    2023

  48. [56]

    Improving neural indoor surface reconstruction with mask-guided adaptive consistency constraints

    Xinyi Yu, Liqin Lu, Jintao Rong, Guangkai Xu, and Linlin Ou. Improving neural indoor surface reconstruction with mask-guided adaptive consistency constraints. arXiv preprint arXiv:2309.09739, 2023. 3

    2023 arXiv

  49. [57]

    Fast-mvsnet: Sparse-to-dense multi-view stereo with learned propagation and gauss-newton refinement

    Zehao Yu and Shenghua Gao. Fast-mvsnet: Sparse-to-dense multi-view stereo with learned propagation and gauss-newton refinement. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 1949–1958,

    1949

  50. [58]

    Monosdf: Exploring monocular geomet- ric cues for neural implicit surface reconstruction

    Zehao Yu, Songyou Peng, Michael Niemeyer, Torsten Sattler, and Andreas Geiger. Monosdf: Exploring monocular geomet- ric cues for neural implicit surface reconstruction. NeurIPS,

  51. [59]

    Gaussian opacity fields: Efficient and compact surface reconstruction in unbounded scenes

    Zehao Yu, Torsten Sattler, and Andreas Geiger. Gaussian opacity fields: Efficient and compact surface reconstruction in unbounded scenes. arXiv preprint arXiv:2404.10772, 2024. 2, 3, 6, 7

    2024 arXiv

  52. [60]

    Dense scene reconstruc- tion with points of interest

    Qian-Yi Zhou and Vladlen Koltun. Dense scene reconstruc- tion with points of interest. ACM Trans. Graph., 32(4), 2013. 1

    2013

  53. [61]

    Open3D: A modern library for 3D data processing

    Qian-Yi Zhou, Jaesik Park, and Vladlen Koltun. Open3D: A modern library for 3D data processing. arXiv:1801.09847,

    arXiv

  54. [1986]

    Association for Computing Machinery. 2

  55. [2018]

    We also give further details about the mesh- ing strategies in Appendix B

    3, 5, 6, 8, 1 11 AGS-Mesh: Adaptive Gaussian Splatting and Meshing with Geometric Priors for Indoor Room Reconstruction Using Smartphones Supplementary Material In this supplementary material, we provide additional details regarding our AGS-Mesh optimization and the pro- posed...

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.