Pith. sign in

REVIEW 4 major objections 5 minor 56 references

MonoSplat: Generalizable 3D Gaussian Splatting from Monocular Depth Foundation Models

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A frozen monocular depth model, lightly adapted, yields state-of-the-art generalizable 3D Gaussian splatting, the paper claims.

desk verdict A clean and credible integration of a frozen monocular depth model into generalizable 3DGS, but the zero-shot DTU numbers hinge on a depth-range protocol ambiguity the authors must clear up. read the letter →

arxiv 2505.15185 v1 pith:6FDBC3W7 submitted 2025-05-21 cs.CV

Yifan Liu , Keyu Fan , Weihao Yu , Chenxin Li , Hao Lu , Yixuan Yuan This is my paper · ORCID
classification cs.CV
keywords 3DGaussianSplattinggeneralizablenovelviewsynthesismonoculardepthfoundationmodelzero-shotcross-datasetgeneralizationplane-sweepcostvolumefeatureadapterAnythingfeed-forwardreconstruction
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper claims that a generalizable 3D Gaussian splatting model can be built by freezing a monocular depth foundation model and adapting its features, instead of training a reconstruction network from scratch. The proposed system, MonoSplat, converts Depth Anything V2 features into multi-view representations with a lightweight attention adapter, then fuses them into the cost volume and refinement network that predicts Gaussian primitives. On RealEstate10K and ACID it reports the best PSNR, SSIM, and LPIPS among compared feed-forward methods, and in zero-shot transfer from RealEstate10K to DTU and ACID it reports larger gains than pixelSplat and MVSplat. If the claim holds, rich monocular depth priors are a reusable, parameter-efficient ingredient for real-time 3D reconstruction of novel scenes without per-scene optimization.

What carries the argument

The load-bearing mechanism is the Mono-Multi Feature Adapter combined with the Integrated Gaussian Prediction module. The Adapter first passes multi-scale encoder features from the frozen Depth Anything V2 backbone through a DPT head to form a unified per-view feature, then runs local-window cross-view attention in the style of Swin Transformer so that each view's features incorporate information from its nearest neighboring views. The Integrated Gaussian Prediction module feeds both these multi-view features and the frozen decoder's monocular features into a plane-sweep cost volume, refines the resulting depth distribution, and decodes Gaussian parameters through depth and appearance heads. The frozen backbone is what preserves the pretrained geometric priors and keeps trainable parameters low, and the paper argues this is why the model transfers across domains.

What would settle it

A decisive test would be to replace the frozen depth encoder's features with randomly initialized features of identical shape, keeping every other component fixed, and then measure zero-shot PSNR on DTU after training on RealEstate10K. If the score does not drop substantially below the reported 15.25 dB, the monocular priors are doing little of the work; if it collapses toward MVSplat's 13.94 dB, the priors are the cause. A second observation: run the same protocol on strongly out-of-distribution imagery such as medical or satellite data, where the depth foundation model has no pretraining; if the advantage disappears there, the generalization claim is bounded by the pretraining distribution.

Watch

Extended reading notes

Core claim

MonoSplat's central claim is that a frozen monocular depth foundation model can be repurposed as the backbone of a generalizable 3D Gaussian splatting system. The paper introduces a Mono-Multi Feature Adapter that fuses multi-scale encoder features through a DPT head and then applies Swin-based cross-view attention, turning view-specific monocular features into geometry-aware multi-view features. These are combined with the decoder's monocular features in an integrated cost volume and in the refinement network that predicts Gaussian positions, opacities, covariances, and colors. On RealEstate10K and ACID, the method reports the best PSNR, SSIM, and LPIPS among the compared feed-forward approaches, and in zero-shot transfer from RealEstate10K to DTU and ACID it reports larger gains than pixelSplat and MVSplat while using only 10.3M trainable parameters out of 30.3M total.

Load-bearing premise

The central load-bearing assumption is that monocular depth features learned on large-scale natural images transfer to arbitrary unseen scenes and genuinely improve cost-volume matching, so that the measured cross-dataset gains come from those priors rather than from extra model capacity or dataset-specific artifacts.

Editorial extensions

If this is right

  • MonoSplat can render novel views of unseen scenes from two posed images without per-scene optimization, with inference times close to those of MVSplat.
  • Zero-shot transfer from indoor RealEstate10K to DTU and ACID improves over prior feed-forward splatting models, and the improvement grows as the domain gap widens.
  • Keeping the depth backbone frozen preserves geometric priors; fine-tuning the backbone drops cross-domain PSNR by 2.61 dB in the paper's ablation.
  • The full model uses 30.3M parameters with only 10.3M trainable, and runs in 0.857 GB of GPU memory, comparable to or better than the cost-volume baseline.
  • Larger frozen Depth Anything V2 variants monotonically improve both in-domain quality and cross-domain generalization.

Reading between the lines

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

  • If the monocular depth priors are the active ingredient, the same recipe could transfer to other feed-forward reconstruction backbones: swapping in a different pretrained depth model should shift the generalization boundary accordingly.
  • The paper's ablation removes the monocular branch together with the fusion subnetwork, so part of the reported drop could reflect reduced capacity; a test that keeps capacity constant would isolate the priors themselves.
  • The frozen-encoder design points to a practical pattern for other 3D tasks: adapt a large pretrained 2D backbone with a small cross-view layer rather than fine-tuning it, preserving broad priors at low parameter cost.
  • The cross-dataset gains are demonstrated only on natural scenes; on imagery far outside the depth model's pretraining distribution, such as medical or satellite data, the advantage may shrink or vanish.
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

4 major / 5 minor

Summary. The paper introduces MonoSplat, a feed-forward generalizable 3D Gaussian splatting model that builds on a frozen monocular depth foundation model (Depth Anything V2). A Mono-Multi Feature Adapter (DPT fusion plus cross-view Swin attention) converts monocular features into multi-view features, and an Integrated Gaussian Prediction module fuses these with a plane-sweep cost volume to predict Gaussian primitives. The model is trained with L_mse + lambda_lpips * L_lpips on RealEstate10K and ACID and evaluated zero-shot on DTU and ACID. The authors report state-of-the-art PSNR/SSIM/LPIPS on both main benchmarks, improved cross-dataset generalization, and efficiency with 10.3M trainable parameters, with code released.

Significance. If the results are reproducible, the paper demonstrates a practical and conceptually clean way to leverage frozen depth foundation models for generalizable Gaussian splatting, with strong cross-domain transfer and a low trainable parameter count. The ablation study comparing DINOv2, UniMatch, and Depth Anything V2 backbones is informative and supports the claim that depth-specific features matter. The main concerns are the ambiguous depth-range protocol in the zero-shot DTU evaluation and several internal inconsistencies, which must be resolved before the state-of-the-art generalization claim can be accepted.

major comments (4)
  1. [Sec. 4.1 and Supp. A.3, Table 2] Section 4.1 fixes the cost-volume depth range as d_near=0.5 and d_far=100 with D=128 planes, but Supplementary A.3 states 'For the DTU dataset, we utilize the depth bounds of 2.125 and 4.525.' Since MonoSplat is never trained on DTU, this sentence can only refer to the evaluation protocol, and the paper does not state whether the baselines in Table 2 (pixelSplat, MVSplat) used the same narrow DTU-specific range or the default 0.5-100 range. Because the plane-sweep cost volume in Eq. (4) is highly sensitive to the sampled depth interval, narrowing it by a factor of about 40 injects knowledge of the target dataset's depth distribution. The claimed 1.31 dB PSNR gain on Re10k-to-DTU (15.25 vs. 13.94) could therefore be an evaluation-protocol artifact rather than evidence for the frozen monocular depth prior. Please specify the exact depth range used for every method in Table 2 and, if the ranges differ, rerun the baselines under identical conditions.
  2. [Sec. 4.3 and Table 3 caption] Section 4.3 states 'all models are trained for 20,000 iterations with a batch size of 14,' while Table 3's caption says 'models trained solely on Re10K with 200k iterations.' These two statements are mutually inconsistent, and the discrepancy directly affects the comparability of the ablations with the main results (300k iterations in Sec. 4.1). Please correct the number and report the actual training budget used for the ablations.
  3. [Related Work and Tables 1-2] The Related Work discusses DepthSplat [48] as a concurrent method that also leverages Depth Anything features for Gaussian splatting, noting that its dual-branch design is parameter-inefficient. However, DepthSplat is absent from all quantitative comparisons in Tables 1 and 2 and from the efficiency analysis. Given the paper's claim of state-of-the-art quality and efficiency, the omission of a direct comparison (or at least a quantitative quote from the DepthSplat paper) leaves the central claim incomplete.
  4. [Table 1 and Abstract] Table 1 shows only small margins over MVSplat (e.g., 26.68 vs. 26.39 PSNR on RealEstate10K; 28.63 vs. 28.25 on ACID), yet no error bars, confidence intervals, or repeated-run statistics are provided. Given the 'convincingly demonstrate' language in the Abstract, please report variance across at least three runs or otherwise justify that the observed differences are statistically significant.
minor comments (5)
  1. [Table 2 caption] The caption contains a typo: 'Quantative' should be 'Quantitative', and 'boldedand underlined' is missing spaces.
  2. [Figure 6 caption] The caption says 'tested on DTU [55]', but reference [55] is the RealEstate10K paper; DTU is [12]. Please correct the citation.
  3. [References] References [12] and [50] appear to have formatting artifacts (a trailing '1' in the page field); please check the bibliography entries.
  4. [Supplementary A.3] The sentence about DTU depth bounds should explicitly state whether these bounds are used during training, evaluation, or both, and for which methods, since the main text only defines the 0.5-100 range.
  5. [Sec. 3.1.2] The phrase 'the decoder's last layer' is ambiguous because Depth Anything V2 has multiple decoder outputs; please specify which output is used (e.g., the feature before the final depth head or the depth head's output).

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: MonoSplat's predictions are produced by an end-to-end model trained on held-out RGB views, with no claim reducing to its inputs by construction.

full rationale

Walking the claimed derivation chain: MonoSplat's mapping f_theta (Eq. 1) is trained end-to-end on RGB reconstruction losses (Eq. 8) against held-out target views; the monocular features from the frozen Depth Anything encoder enter only as inputs to the cost-volume integration (Eqs. 4-5) and are never regressed to the reported PSNR/SSIM/LPIPS values. The components in Table 3 are architectural ablations, not fitted parameters renamed as predictions, and the cross-dataset numbers in Table 2 are produced by a fixed model without any DTU or ACID parameter fitting. No load-bearing step invokes a uniqueness theorem or prior work by these authors; the cited foundations (Depth Anything, DPT, MVSplat, pixelSplat) are external, code-dependent pieces of evidence. The supplementary depth-bounds statement for DTU (A.3: 'For the DTU dataset, we utilize the depth bounds of 2.125 and 4.525') is a potential evaluation-protocol confound for the zero-shot comparison, but it is a correctness and fairness concern, not a circular reduction: the bounds are not fitted model parameters, and the predicted renders are not defined in terms of the benchmark scores. Therefore no step reduces to its own inputs, and the circularity score is 0.

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

The central claim rests on a pretrained depth backbone whose features are assumed transferable, on a discrete plane-sweep cost volume with hand-set depth bounds (0.5 to 100, 128 planes), on a loss mixing MSE and LPIPS with lambda=0.05, and on a progressive training schedule. These are reasonable engineering choices, but none of them are derived from first principles: they are empirical hyperparameters.

free parameters (4)
  • LPIPS weight lambda_lpips = 0.05
    Set by hand to balance MSE and perceptual loss in Eq. (8); affects the learned trade-off between pixel accuracy and perceptual quality.
  • Depth sampling planes = 128 planes over [0.5, 100] for RE10K/ACID
    The cost volume (Sec. 3.2.1) samples 128 discrete depths between near=0.5 and far=100. The range and number of planes bound the depth resolution and the scene scale the model can handle.
  • Gaussian scale constraints = [0.5, 15.0]
    Gaussian scales are clamped to [0.5, 15.0] to balance detail and stable optimization (Sec. A.2); this directly shapes the expressiveness of the splats.
  • Progressive view spacing schedule = min distance 25 to 45, max distance 45 to 192 over first 150k steps
    Training curriculum gradually increases baseline between input views; the final performance may depend on this schedule (Sec. A.1).
assumptions (4)
  • domain assumption The Depth Anything V2 pretrained model encodes reliable geometric priors that transfer to novel scenes.
    The entire method relies on this; validated only via the reported experiments, and the ablation 'w/o MF' is confounded.
  • domain assumption Plane-sweep stereo with 128 discrete depth planes provides sufficient geometric matching for general scenes.
    The integrated cost volume (Sec. 3.2.1) uses this discretization; it limits depth resolution and may fail for scenes outside the depth range.
  • standard math The differentiable 3D Gaussian splatting rendering model is an appropriate forward model for photometric training.
    This is standard in the field and is taken as given from prior work (Kerbl et al., 2023).
  • domain assumption Input camera poses are accurate and known.
    The method requires posed images and assumes the projection matrices Pi are correct, as in Eq. (1).

how reviews work

0 comments
Cite this review

Pith. "Pith review of MonoSplat: Generalizable 3D Gaussian Splatting from Monocular Depth Foundation Models." pith.science (2026). https://pith.science/paper/6FDBC3W7

@misc{pith2026250515185,
  author       = {Pith},
  title        = {Pith review of: MonoSplat: Generalizable 3D Gaussian Splatting from Monocular Depth Foundation Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6FDBC3W7}},
  note         = {Machine review of arXiv:2505.15185}
}
read the original abstract

Recent advances in generalizable 3D Gaussian Splatting have demonstrated promising results in real-time high-fidelity rendering without per-scene optimization, yet existing approaches still struggle to handle unfamiliar visual content during inference on novel scenes due to limited generalizability. To address this challenge, we introduce MonoSplat, a novel framework that leverages rich visual priors from pre-trained monocular depth foundation models for robust Gaussian reconstruction. Our approach consists of two key components: a Mono-Multi Feature Adapter that transforms monocular features into multi-view representations, coupled with an Integrated Gaussian Prediction module that effectively fuses both feature types for precise Gaussian generation. Through the Adapter's lightweight attention mechanism, features are seamlessly aligned and aggregated across views while preserving valuable monocular priors, enabling the Prediction module to generate Gaussian primitives with accurate geometry and appearance. Through extensive experiments on diverse real-world datasets, we convincingly demonstrate that MonoSplat achieves superior reconstruction quality and generalization capability compared to existing methods while maintaining computational efficiency with minimal trainable parameters. Codes are available at https://github.com/CUHK-AIM-Group/MonoSplat.

Figures

Figures reproduced from arXiv: 2505.15185 by the authors.

Figure 1
Figure 1. Sketch of the proposed MonoSplat framework. We repurpose the depth foundation model to generalizable Gaussian reconstruction framework. (NeRF) [21], light field networks (LFN) [35], and most re￾cently 3D Gaussian Splatting (3DGS) [14] - practical de￾ployment still faces critical barriers. These include substan￾tial computational overhead during rendering [42, 51] and time-consuming per-scene optimization procedures … view at source ↗
Figure 2
Figure 2. Illustration of the proposed MonoSplat framework. It consists of a Mono-Multi Feature Adapter to convert monocular features to multi-view features with cross-view awareness, and Integrated Gaussian Prediction to integrate monocular features into the Gaussian primitives prediction. thing Model [50]. However, their dual-branch network ar￾chitecture introduces redundant model parameters. In con￾trast, our framework ach… view at source ↗
Figure 3
Figure 3. Visual comparisons of cross-dataset generalization. Using a model trained on RealEstate10k [55], we directly render scenes from DTU [12] datasets. MonoSplat shows superior gener￾alization ability compared to previous methods. outdoor scenes and DTU object-centric captures. As evi￾denced in [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Visual comparisons of novel view synthesis on RealEstate10k [55]. Models are trained with a collection of scenes from RealEstate10k, and tested on novel scenes from the test splits. MonoSplat shows superior rendering quality in challenging regions. disparity between ou…
Figure 5
Figure 5. Figure 5: Visual comparisons of predicted 3D Gaussians (top) and depth maps (bottom). We examine the quality of geometric reconstruction by presenting comparative visualizations of 3D Gaussian primitives generated by pixelSplat, MVSplat, and our MonoSplat, alongside depth maps r…
Figure 6
Figure 6. Figure 6: Visual comparisons on Re10k→ DTU with different designs. Models are trained on RealEstate10k [55] and directly tested on DTU [55]. Color-coded error distributions are illustrated with the pixel-wise differences between synthesized outputs and ground truth images. drama…
Figure 7
Figure 7. Figure 7: Additional visual comparisons of geometry reconstruction on RealEstate10k [55]. All models were trained on a diverse collection of RealEstate10k scenes and evaluated on previously unseen scenes from the test split. For reference, we provide two rendered depth maps from…
Figure 8
Figure 8. Figure 8: More visual comparisons of cross-dataset generalization from RealEstate10k [55] to DTU [55]. Models are trained solely on RealEstate10k, and tested on novel scenes from DTU. Our method shows superior rendering quality compared to previous methods [PITH_FULL_IMAGE:figu…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

56 extracted references · 29 canonical work pages

  1. [48]

    Depthsplat: Connecting gaussian splatting and depth.arXiv preprint arXiv:2410.13862, 2024

    Haofei Xu, Songyou Peng, Fangjinhua Wang, Hermann Blum, Daniel Barath, Andreas Geiger, and Marc Pollefeys. Depthsplat: Connecting gaussian splatting and depth.arXiv preprint arXiv:2410.13862, 2024. 2

    arXiv 2024

  2. [1]

    pixelsplat: 3d gaussian splats from image pairs for scalable generalizable 3d reconstruction

    David Charatan, Sizhe Li, Andrea Tagliasacchi, and Vincent Sitzmann. pixelsplat: 3d gaussian splats from image pairs for scalable generalizable 3d reconstruction. InCVPR, 2024. 1, 2, 5, 6, 7

    work page 2024

  3. [2]

    Mvsnerf: Fast general- izable radiance field reconstruction from multi-view stereo

    Anpei Chen, Zexiang Xu, Fuqiang Zhao, Xiaoshuai Zhang, Fanbo Xiang, Jingyi Yu, and Hao Su. Mvsnerf: Fast general- izable radiance field reconstruction from multi-view stereo. InICCV, 2021. 2

    work page 2021

  4. [3]

    Tensorf: Tensorial radiance fields

    Anpei Chen, Zexiang Xu, Andreas Geiger, Jingyi Yu, and Hao Su. Tensorf: Tensorial radiance fields. InEuropean con- ference on computer vision, pages 333–350. Springer, 2022. 2

    2022

  5. [4]

    A survey on 3d gaussian splatting.arXiv, 2024

    Guikun Chen and Wenguan Wang. A survey on 3d gaussian splatting.arXiv, 2024. 2

    work page 2024

  6. [5]

    Explicit correspondence matching for generalizable neural radiance fields

    Yuedong Chen, Haofei Xu, Qianyi Wu, Chuanxia Zheng, Tat-Jen Cham, and Jianfei Cai. Explicit correspondence matching for generalizable neural radiance fields. InarXiv,

    work page

  7. [6]

    Mvsplat: Efficient 3d gaussian splatting from sparse multi-view images.arXiv preprint arXiv:2403.14627, 2024

    Yuedong Chen, Haofei Xu, Chuanxia Zheng, Bohan Zhuang, Marc Pollefeys, Andreas Geiger, Tat-Jen Cham, and Jianfei Cai. Mvsplat: Efficient 3d gaussian splatting from sparse multi-view images.arXiv preprint arXiv:2403.14627, 2024. 1, 2, 5, 6, 7

    arXiv 2024

  8. [7]

    Stereo radiance fields (srf): Learning view syn- thesis for sparse views of novel scenes

    Julian Chibane, Aayush Bansal, Verica Lazova, and Gerard Pons-Moll. Stereo radiance fields (srf): Learning view syn- thesis for sparse views of novel scenes. InCVPR, 2021. 2

    work page 2021

Show all 56 references
  1. [8]

    Learning to render novel views from wide-baseline stereo pairs

    Yilun Du, Cameron Smith, Ayush Tewari, and Vincent Sitz- mann. Learning to render novel views from wide-baseline stereo pairs. InCVPR, 2023. 2, 6

    2023

  2. [9]

    Pixelgaussian: Gen- eralizable 3d gaussian reconstruction from arbitrary views

    Xin Fei, Wenzhao Zheng, Yueqi Duan, Wei Zhan, Masayoshi Tomizuka, Kurt Keutzer, and Jiwen Lu. Pixelgaussian: Gen- eralizable 3d gaussian reconstruction from arbitrary views. arXiv preprint arXiv:2410.18979, 2024. 2

    2024 arXiv

  3. [10]

    Plenoxels: Radiance fields without neural networks

    Sara Fridovich-Keil, Alex Yu, Matthew Tancik, Qinhong Chen, Benjamin Recht, and Angjoo Kanazawa. Plenoxels: Radiance fields without neural networks. InCVPR, 2022. 2

    2022

  4. [11]

    Unsupervised learning of 3d object cate- gories from videos in the wild

    Philipp Henzler, Jeremy Reizenstein, Patrick Labatut, Ro- man Shapovalov, Tobias Ritschel, Andrea Vedaldi, and David Novotny. Unsupervised learning of 3d object cate- gories from videos in the wild. InCVPR, 2021. 2

    2021

  5. [12]

    Large scale multi-view stereopsis evalu- ation

    Rasmus Jensen, Anders Dahl, George V ogiatzis, Engin Tola, and Henrik Aanæs. Large scale multi-view stereopsis evalu- ation. InCVPR, 2014. 5, 6, 8, 1

    2014

  6. [13]

    Geonerf: Generalizing nerf with geometry priors

    Mohammad Mahdi Johari, Yann Lepoittevin, and Franc ¸ois Fleuret. Geonerf: Generalizing nerf with geometry priors. InCVPR, 2022. 2

    2022

  7. [14]

    3d gaussian splatting for real-time radiance field rendering.TOG, 42(4), 2023

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

    2023

  8. [15]

    Infinite nature: Perpetual view generation of natural scenes from a single image

    Andrew Liu, Richard Tucker, Varun Jampani, Ameesh Makadia, Noah Snavely, and Angjoo Kanazawa. Infinite nature: Perpetual view generation of natural scenes from a single image. InICCV, 2021. 5, 6

    2021

  9. [16]

    Neural sparse voxel fields.Advances in Neural Information Processing Systems, 33:15651–15663,

    Lingjie Liu, Jiatao Gu, Kyaw Zaw Lin, Tat-Seng Chua, and Christian Theobalt. Neural sparse voxel fields.Advances in Neural Information Processing Systems, 33:15651–15663,

  10. [17]

    Fast generalizable gaussian splatting reconstruction from multi-view stereo.arXiv preprint arXiv:2405.12218,

    Tianqi Liu, Guangcong Wang, Shoukang Hu, Liao Shen, Xinyi Ye, Yuhang Zang, Zhiguo Cao, Wei Li, and Ziwei Liu. Fast generalizable gaussian splatting reconstruction from multi-view stereo.arXiv preprint arXiv:2405.12218,

    arXiv

  11. [18]

    Neural rays for occlusion-aware image-based render- ing

    Yuan Liu, Sida Peng, Lingjie Liu, Qianqian Wang, Peng Wang, Christian Theobalt, Xiaowei Zhou, and Wenping Wang. Neural rays for occlusion-aware image-based render- ing. InCVPR, 2022. 2

    2022

  12. [19]

    Swin transformer: Hierarchical vision transformer using shifted windows

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. In ICCV, 2021. 4

    2021

  13. [20]

    Neural vol- umes: Learning dynamic renderable volumes from images

    Stephen Lombardi, Tomas Simon, Jason Saragih, Gabriel Schwartz, Andreas Lehrmann, and Yaser Sheikh. Neural vol- umes: Learning dynamic renderable volumes from images. arXiv preprint arXiv:1906.07751, 2019. 2

    1906 arXiv

  14. [21]

    Nerf: Representing scenes as neural radiance fields for view syn- thesis

    Ben Mildenhall, Pratul P Srinivasan, Matthew Tancik, Jonathan T Barron, Ravi Ramamoorthi, and Ren Ng. Nerf: Representing scenes as neural radiance fields for view syn- thesis. InECCV, 2020. 1, 2

    2020

  15. [22]

    Epipolar-free 3d gaussian splatting for generalizable novel view synthesis.arXiv preprint arXiv:2410.22817, 2024

    Zhiyuan Min, Yawei Luo, Jianwen Sun, and Yi Yang. Epipolar-free 3d gaussian splatting for generalizable novel view synthesis.arXiv preprint arXiv:2410.22817, 2024. 1, 2, 6

    2024 arXiv

  16. [23]

    Gta: A geometry-aware attention mechanism for multi-view transformers

    Takeru Miyato, Bernhard Jaeger, Max Welling, and Andreas Geiger. Gta: A geometry-aware attention mechanism for multi-view transformers. InICLR, 2024. 2

    2024

  17. [24]

    Instant neural graphics primitives with a mul- tiresolution hash encoding.ACM transactions on graphics (TOG), 41(4):1–15, 2022

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

    2022

  18. [25]

    Rendernet: A deep convolutional network for differentiable rendering from 3d shapes.Advances in neural information processing systems, 31, 2018

    Thu H Nguyen-Phuoc, Chuan Li, Stephen Balaban, and Yongliang Yang. Rendernet: A deep convolutional network for differentiable rendering from 3d shapes.Advances in neural information processing systems, 31, 2018. 2

    2018

  19. [26]

    Reg- nerf: Regularizing neural radiance fields for view synthesis from sparse inputs

    Michael Niemeyer, Jonathan T Barron, Ben Mildenhall, Mehdi SM Sajjadi, Andreas Geiger, and Noha Radwan. Reg- nerf: Regularizing neural radiance fields for view synthesis from sparse inputs. InCVPR, 2022. 1

    2022

  20. [27]

    Dreamfusion: Text-to-3d using 2d diffusion.arXiv preprint arXiv:2209.14988, 2022

    Ben Poole, Ajay Jain, Jonathan T Barron, and Ben Milden- hall. Dreamfusion: Text-to-3d using 2d diffusion.arXiv preprint arXiv:2209.14988, 2022. 1

    2022 arXiv

  21. [28]

    Ren ´e Ranftl, Katrin Lasinger, David Hafner, Konrad Schindler, and Vladlen Koltun. Towards robust monocular depth estimation: Mixing datasets for zero-shot cross-dataset transfer.IEEE transactions on pattern analysis and machine intelligence, 44(3):1623–1637, 2020. 2

    2020

  22. [29]

    Vi- sion transformers for dense prediction

    Ren ´e Ranftl, Alexey Bochkovskiy, and Vladlen Koltun. Vi- sion transformers for dense prediction. InProceedings of the IEEE/CVF international conference on computer vision, pages 12179–12188, 2021. 4

    2021

  23. [30]

    Com- mon objects in 3d: Large-scale learning and evaluation of real-life 3d category reconstruction

    Jeremy Reizenstein, Roman Shapovalov, Philipp Henzler, Luca Sbordone, Patrick Labatut, and David Novotny. Com- mon objects in 3d: Large-scale learning and evaluation of real-life 3d category reconstruction. InICCV, 2021. 2

    2021

  24. [31]

    Scene representation transformer: Geometry-free novel view syn- thesis through set-latent scene representations

    Mehdi SM Sajjadi, Henning Meyer, Etienne Pot, Urs Bergmann, Klaus Greff, Noha Radwan, Suhani V ora, Mario Luˇci´c, Daniel Duckworth, Alexey Dosovitskiy, et al. Scene representation transformer: Geometry-free novel view syn- thesis through set-latent scene representations. InCVPR,

  25. [32]

    Deep- voxels: Learning persistent 3d feature embeddings

    Vincent Sitzmann, Justus Thies, Felix Heide, Matthias Nießner, Gordon Wetzstein, and Michael Zollhofer. Deep- voxels: Learning persistent 3d feature embeddings. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2437–2446, 2019. 2

    2019

  26. [33]

    Scene representation networks: Continuous 3d- structure-aware neural scene representations.NeurIPS,

    Vincent Sitzmann, Michael Zollh ¨ofer, and Gordon Wet- zstein. Scene representation networks: Continuous 3d- structure-aware neural scene representations.NeurIPS,

  27. [34]

    Implicit neural representa- tions with periodic activation functions.Advances in neural information processing systems, 33:7462–7473, 2020

    Vincent Sitzmann, Julien Martel, Alexander Bergman, David Lindell, and Gordon Wetzstein. Implicit neural representa- tions with periodic activation functions.Advances in neural information processing systems, 33:7462–7473, 2020. 2

    2020

  28. [35]

    Light field networks: Neu- ral scene representations with single-evaluation rendering

    Vincent Sitzmann, Semon Rezchikov, Bill Freeman, Josh Tenenbaum, and Fredo Durand. Light field networks: Neu- ral scene representations with single-evaluation rendering. NeurIPS, 2021. 1

    2021

  29. [36]

    Generalizable patch-based neural render- ing

    Mohammed Suhail, Carlos Esteves, Leonid Sigal, and Ameesh Makadia. Generalizable patch-based neural render- ing. InECCV, 2022. 6

    2022

  30. [37]

    Flash3d: Feed-forward gener- alisable 3d scene reconstruction from a single image.arxiv,

    Stanislaw Szymanowicz, Eldar Insafutdinov, Chuanxia Zheng, Dylan Campbell, Joao Henriques, Christian Rup- precht, and Andrea Vedaldi. Flash3d: Feed-forward gener- alisable 3d scene reconstruction from a single image.arxiv,

  31. [38]

    Splatter image: Ultra-fast single-view 3d recon- struction

    Stanislaw Szymanowicz, Christian Rupprecht, and Andrea Vedaldi. Splatter image: Ultra-fast single-view 3d recon- struction. InCVPR, 2024. 2

    2024

  32. [39]

    Block-nerf: Scalable large scene neural view synthesis

    Matthew Tancik, Vincent Casser, Xinchen Yan, Sabeek Prad- han, Ben Mildenhall, Pratul P Srinivasan, Jonathan T Barron, and Henrik Kretzschmar. Block-nerf: Scalable large scene neural view synthesis. InProceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Reco...

    2022

  33. [40]

    Hisplat: Hierarchical 3d gaus- sian splatting for generalizable sparse-view reconstruction

    Shengji Tang, Weicai Ye, Peng Ye, Weihao Lin, Yang Zhou, Tao Chen, and Wanli Ouyang. Hisplat: Hierarchical 3d gaus- sian splatting for generalizable sparse-view reconstruction. arXiv preprint arXiv:2410.06245, 2024. 2

    2024 arXiv

  34. [41]

    Advances in neural rendering

    Ayush Tewari, Justus Thies, Ben Mildenhall, Pratul Srini- vasan, Edgar Tretschk, Wang Yifan, Christoph Lassner, Vin- cent Sitzmann, Ricardo Martin-Brualla, Stephen Lombardi, et al. Advances in neural rendering. InComputer Graphics Forum, pages 703–735. Wiley Online Library, 2022. 2

    2022

  35. [42]

    Ibr- net: Learning multi-view image-based rendering

    Qianqian Wang, Zhicheng Wang, Kyle Genova, Pratul P Srinivasan, Howard Zhou, Jonathan T Barron, Ricardo Martin-Brualla, Noah Snavely, and Thomas Funkhouser. Ibr- net: Learning multi-view image-based rendering. InCVPR,

  36. [43]

    Freesplat: Generalizable 3d gaussian splatting to- wards free-view synthesis of indoor scenes.arXiv preprint arXiv:2405.17958, 2024

    Yunsong Wang, Tianxin Huang, Hanlin Chen, and Gim Hee Lee. Freesplat: Generalizable 3d gaussian splatting to- wards free-view synthesis of indoor scenes.arXiv preprint arXiv:2405.17958, 2024. 1, 2

    2024 arXiv

  37. [44]

    Image quality assessment: from error visibility to structural similarity.TIP, 13(4), 2004

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Simoncelli. Image quality assessment: from error visibility to structural similarity.TIP, 13(4), 2004. 5

    2004

  38. [45]

    latentsplat: Autoencoding variational gaussians for fast generalizable 3d reconstruction.arXiv preprint arXiv:2403.16292, 2024

    Christopher Wewer, Kevin Raj, Eddy Ilg, Bernt Schiele, and Jan Eric Lenssen. latentsplat: Autoencoding variational gaussians for fast generalizable 3d reconstruction.arXiv preprint arXiv:2403.16292, 2024. 1, 2, 6

    2024 arXiv

  39. [46]

    Neural fields in visual computing and beyond

    Yiheng Xie, Towaki Takikawa, Shunsuke Saito, Or Litany, Shiqin Yan, Numair Khan, Federico Tombari, James Tomp- kin, Vincent Sitzmann, and Srinath Sridhar. Neural fields in visual computing and beyond. InComputer Graphics Forum, pages 641–676. Wiley Online Library, 2022. 2

    2022

  40. [47]

    Murf: Multi-baseline radiance fields

    Haofei Xu, Anpei Chen, Yuedong Chen, Christos Sakaridis, Yulun Zhang, Marc Pollefeys, Andreas Geiger, and Fisher Yu. Murf: Multi-baseline radiance fields. InCVPR, 2024. 2, 6

    2024

  41. [49]

    Vr-nerf: High- fidelity virtualized walkable spaces

    Linning Xu, Vasu Agrawal, William Laney, Tony Garcia, Aayush Bansal, Changil Kim, Samuel Rota Bul `o, Lorenzo Porzi, Peter Kontschieder, Aljaˇz Boˇziˇc, et al. Vr-nerf: High- fidelity virtualized walkable spaces. InSIGGRAPH Asia 2023 Conference Papers, pages 1–12, 2023. 1

    2023

  42. [50]

    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. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 10371–10381, 2024. 2, 3, 5, 1

    2024

  43. [51]

    pixelnerf: Neural radiance fields from one or few images

    Alex Yu, Vickie Ye, Matthew Tancik, and Angjoo Kanazawa. pixelnerf: Neural radiance fields from one or few images. In CVPR, 2021. 1, 2, 6

    2021

  44. [52]

    Transplat: Generalizable 3d gaussian splatting from sparse multi-view images with transformers

    Chuanrui Zhang, Yingshuang Zou, Zhuoling Li, Minmin Yi, and Haoqian Wang. Transplat: Generalizable 3d gaussian splatting from sparse multi-view images with transformers. arXiv preprint arXiv:2408.13770, 2024. 2

    2024 arXiv

  45. [53]

    The unreasonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shechtman, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. InCVPR, 2018. 5

    2018

  46. [54]

    Gps- gaussian: Generalizable pixel-wise 3d gaussian splatting for real-time human novel view synthesis

    Shunyuan Zheng, Boyao Zhou, Ruizhi Shao, Boning Liu, Shengping Zhang, Liqiang Nie, and Yebin Liu. Gps- gaussian: Generalizable pixel-wise 3d gaussian splatting for real-time human novel view synthesis. InCVPR, 2024. 2

    2024

  47. [55]

    Stereo magnification: learning view syn- thesis using multiplane images.TOG, page 65, 2018

    Tinghui Zhou, Richard Tucker, John Flynn, Graham Fyffe, and Noah Snavely. Stereo magnification: learning view syn- thesis using multiplane images.TOG, page 65, 2018. 5, 6, 7, 8, 1, 2, 3

    2018

  48. [56]

    Fsgs: Real-time few-shot view synthesis using gaussian splatting.arXiv preprint arXiv:2312.00451, 2023

    Zehao Zhu, Zhiwen Fan, Yifan Jiang, and Zhangyang Wang. Fsgs: Real-time few-shot view synthesis using gaussian splatting.arXiv preprint arXiv:2312.00451, 2023. 1 MonoSplat: Generalizable 3D Gaussian Splatting from Monocular Depth Foundation Models Supplementary Material A. Mor...

    2023 arXiv

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.