Pith. sign in

REVIEW 2 major objections 5 minor 38 references

Joint Optimization of Neural Radiance Fields and Continuous Camera Motion from a Monocular Video

T0 review · 2 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read Modeling camera motion as continuous angular velocity and velocity lets a NeRF recover camera poses and scene geometry from a monocular video without depth priors.

desk verdict Genuinely new continuous-motion parameterization for joint pose-NeRF optimization, with solid static-scene results; the unstated rigid-world assumption is the main gap, but the paper deserves serious review. read the letter →

arxiv 2504.19819 v1 pith:R64QR63R submitted 2025-04-28 cs.CV

Hoang Chuong Nguyen , Wei Mao , Jose M. Alvarez , Miaomiao Liu This is my paper · ORCID
classification cs.CV
keywords neuralradiancefieldscameraposeestimationcontinuousmotionmonocularvideosigneddistancefieldsceneflowjointoptimizationdepth
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

This paper tries to establish that a neural radiance field can be trained from an ordinary monocular video without precomputed camera poses, pose initialization, or depth priors. Instead of optimizing a separate camera pose per frame, the method learns a continuous camera motion, an angular velocity $\omega(t)$ and a velocity $v(t)$, that is integrated over time to give relative motions between frames, with a single time step fixing the world coordinate system. A time-dependent SDF-based NeRF is trained on neighboring frames, and the learned motions are then used to fine-tune a global NeRF that represents the whole scene. If correct, this removes the need for Structure-from-Motion preprocessing and handles large camera rotations that break per-frame pose optimization, with the paper reporting superior pose and depth accuracy on Co3D and ScanNet.

What carries the argument

The load-bearing object is the continuous camera-motion representation: a motion MLP $\phi_v$ that maps time $t$ to angular velocity $\omega(t)\in\mathbb{R}^3$ and velocity $v(t)\in\mathbb{R}^3$, with frame-to-frame transformations $P_{t_1\to t_2}$ obtained by Euler integration of these velocities. The second essential component is a time-dependent SDF-based NeRF, adapted from the NeuS representation, which predicts signed distance $s(x,t)$ and color $c(x,d,t)$ in each local camera coordinate system; training it on neighboring frames gives local geometry sufficient to constrain the motion. The two are coupled by the SDF-flow consistency loss (Eq. 11), which enforces that temporal change in the SDF matches the change expected from rigid camera motion, so scene geometry and camera motion must agree.

What would settle it

Record a monocular video in which the camera moves but a foreground object also moves independently (or the scene contains a large reflective surface), and compare the recovered camera trajectory against a ground-truth trajectory from an inertial measurement unit; if pose error rises sharply or the moving object leaves ghosting artifacts in the reconstructed geometry, the rigidity assumption is what fails.

Watch

Extended reading notes

Core claim

The paper's central claim is that jointly optimizing a continuous-time camera motion and a time-dependent signed-distance-field NeRF, then fine-tuning the NeRF in the world frame with the learned poses held fixed, yields accurate camera poses and scene geometry from monocular video without any geometric priors. The camera motion is represented by an MLP predicting $\omega(t)$ and $v(t)$, and the relative transformation between any two frames is obtained by Euler integration of these velocities; poses are aggregated to a world frame defined at a single middle time step. During joint optimization, the time-dependent NeRF supplies local scene geometry, and three consistency losses tie that geometry to the motion: the SDF-flow loss linking $\partial s/\partial t$ to $(\omega(t)\times x + v(t))^\top n(x)$, a photometric reprojection loss between neighboring frames, and an SDF-consistency loss against the world frame. After the motion is learned, the time-dependent model is fine-tuned as a conventional NeRF at the world time step using all frames. On Co3D and ScanNet, the paper reports that this outperforms prior NeRF-based and 3DGS-based joint-optimization methods in camera pose and depth estimation, with novel-view synthesis comparable to the best baseline.

Load-bearing premise

The load-bearing assumption is that the scene is static and rigid, so every temporal change in the observed images and in the signed distance field is caused only by the camera's own motion.

Editorial extensions

If this is right

  • Training a NeRF from monocular video no longer requires COLMAP, a good pose initialization, or a pre-trained depth network.
  • Large camera rotations and translations become learnable as accumulated small relative motions rather than as large camera-to-world mappings.
  • The time-dependent NeRF stage provides local geometry per time step, which is later merged into a full global scene by fine-tuning with the fixed integrated poses.
  • The paper reports that on Co3D and ScanNet the method reduces pose and depth errors compared to NeRFmm, NoPe-NeRF, and CF3DGS, while matching CF3DGS in novel-view synthesis.

Reading between the lines

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

  • The same velocity-integration strategy could be extended to dynamic scenes by explicitly modeling residual non-rigid flow, so independently moving objects would not be absorbed into camera motion.
  • The method's sensitivity to low-texture and reflective regions suggests a testable modification: weighting the photometric loss by per-pixel uncertainty or adding a normal-smoothness prior could extend it to such surfaces.
  • If the approach scales to longer and more varied videos, it could replace Structure-from-Motion preprocessing in monocular SLAM and dense reconstruction systems, since it outputs both camera trajectories and dense geometry at once.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 5 minor

Summary. The paper proposes a method for jointly optimizing camera poses and a NeRF-based scene representation from a monocular video, without relying on depth priors or SfM initialization. Camera motion is parameterized as continuous angular velocity and velocity functions of time predicted by an MLP; relative poses between frames are obtained by Euler integration of these velocities. A time-dependent SDF-based NeRF is trained with several consistency losses: a flow loss relating the SDF temporal derivative to the camera velocities (Eq. 11), a photometric reprojection loss (Eq. 14), and an SDF consistency loss with a chosen world frame (Eq. 15). After the motion network converges, the poses are fixed and the NeRF is fine-tuned in a conventional way. Experiments on ScanNet and Co3D show consistent improvements over NeRFmm, NoPe-NeRF, and CF3DGS in pose and depth estimation, with competitive novel-view synthesis.

Significance. If the result holds, the method removes the COLMAP and depth-prior requirements for NeRF training on static monocular videos, handles large camera rotations through a continuous motion representation, and provides a principled coupling between scene geometry and camera motion. The paper is clearly written, the ablation study isolates each loss component, and the code is publicly released. The main significance is the demonstration that a continuous-motion parameterization plus a time-dependent NeRF can jointly recover poses and geometry without external priors; however, the claims are currently limited to static scenes and the generality of the method is not tested.

major comments (2)
  1. [Section 3.3, Eq. (11) and Eq. (14)] The flow loss and the photometric consistency loss enforce that every temporal change in the SDF and every pixel color change between frames is explained by the continuous camera motion (ω(t), v(t)). This is mathematically exact only under a static, rigid scene assumption, since any independently moving object or non-rigid deformation would be absorbed into the estimated camera velocities and corrupt both poses and geometry. The paper never states this assumption in the abstract or introduction, and the Limitation paragraph (Section 5) mentions only low-texture and reflective regions, not dynamic content. The evaluation is restricted to ScanNet and Co3D, which are static scenes; consequently the unqualified claim of a 'prior-free pipeline to jointly optimize camera motion and scene geometry from a monocular video' is not established for dynamic or partially dynamic scenes. The authors should either explicitly scope the method to static scenes or add a dynamic-scene experiment (e.g., a video with a moving object) to test the boundary.
  2. [Section 4.2, Tables 1-3] All metrics are reported from a single training run with no error bars or multiple seeds. Given the strong claim of 'superior' pose and depth estimation, and the fact that some ScanNet pose margins are modest (e.g., RPEt of 0.401 vs 0.455 for NoPe-NeRF on scene 0418), the reader cannot assess whether the improvements are statistically meaningful. At minimum, the authors should report results over multiple runs (e.g., 3 seeds) or state the variance observed; this is particularly important in a neural optimization setting where results are known to vary with initialization and ray sampling.
minor comments (5)
  1. [Section 3.1] The heading 'NeRF with SDF Presentation' contains a typo; it should be 'SDF Representation'.
  2. [Section 3.3] The hyperparameter N (the set of intervals to neighboring frames used in Lphoto, Eq. 14) is not specified in the main text; it is never stated which values are used in the experiments, so readers cannot reproduce the photometric consistency loss setting.
  3. [Table 4] The ablation study reports results 'on the Co3D dataset' but does not specify which scene(s) the numbers correspond to; if it is an average over all five Co3D scenes, this should be stated, and if it is a single scene, the scene name and the rationale for choosing it should be given.
  4. [Section 3.2, Eq. (5)-(6)] The frame convention for the angular velocity and velocity (whether they are expressed in the camera frame or the world frame) is not explicitly stated, which makes the integration formulas in Eq. (5)-(6) ambiguous; please clarify the convention.
  5. [Section 4.1] The paper claims that the method requires no priors, but the ray sampling requires a pre-defined depth range and the camera intrinsics are assumed known; a brief note acknowledging these mild assumptions would be helpful for precision.

Circularity Check

1 steps flagged · score 4.0 of 10

Pose/depth claims are externally validated; the novel-view synthesis metric is partially circular because test poses are fitted to the scored images.

  1. fitted input called prediction [Section 4.2, Novel-view Synthesis paragraph]
    "To obtain the camera poses corresponding to the test images, we keep the NeRF model fixed and learn these poses by minimizing the color rendering lossLrgb, as in [2]."

    The test pose is the fitted parameter and the test image is both the fitting target (Lrgb) and the evaluation image for PSNR/SSIM/LPIPS. With a free per-view pose optimized on the scored image, photometric error on that image can be driven down even when the fixed NeRF's geometry is imperfect; the reported novel-view synthesis quality therefore measures pose-fitting plus reconstruction, not prediction of an unseen view. This does not affect the pose/depth results, which use the learned motion network without seeing the test poses, but it makes the novel-view synthesis claim partially forced by construction.

full rationale

The central pose and depth contributions are not circular: the continuous camera motion ω(t), v(t) and the time-dependent SDF are optimized only against RGB frames via Lrgb, Lflow, Lphoto, and Lsdf, and the pose/depth evaluations compare the resulting fixed trajectory and rendered depth against external ScanNet/Co3D ground truth. No target pose or depth is used as supervision, so the main claim is independently testable. The constraint in Eq. 11 is attributed to [18], which shares two authors with this paper; however, the SDF-flow identity ∂s/∂t + (ω×x+v)^T∇s = 0 is a standard mathematical relation for a rigidly moving implicit surface and is externally verifiable, so the self-citation is not a circular load-bearing step. The static-scene rigidity premise and the low-texture/reflective limitation are assumptions, not circularity. The one genuinely circular element is the novel-view synthesis evaluation: the paper optimizes each test image's camera pose by minimizing Lrgb on that same test image and then reports PSNR/SSIM/LPIPS for that image (Sec. 4.2). By construction, the reported NVS quality is a best-fit reconstruction rather than a prediction of an unseen view. Because the pose and depth claims are unaffected, the overall circularity is partial, not systemic.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The method is an empirical optimization pipeline, not a physical derivation. It introduces no new physical entities, but it depends on a stated static-scene assumption, a cited SDF-flow relation, known camera intrinsics, and several hand-set hyperparameters (loss weights, integration steps, sampling counts, schedules) that are not fully specified in the main text.

free parameters (7)
  • lambda_1, lambda_2, lambda_3, lambda_4 (loss weights in Eq. 16)
    The four loss terms are combined with weights lambda_i that are chosen by hand; values are not given in the main text, so the balance of losses is a free choice that affects all results.
  • U (Euler sub-intervals per frame) = 10
    Number of sub-intervals used to integrate angular velocity and velocity between consecutive frames (Sec. 3.2). Larger or smaller U changes integration accuracy and cost.
  • Neighboring frame interval set N for Lphoto
    Defines which neighboring frames are used in the photometric consistency loss (Eq. 14); not specified in the main text.
  • Depth range for ray sampling
    Training samples 128 points per ray within a pre-defined depth range (Implementation Details); the range is not given and depends on scene scale.
  • First-stage training length
    The joint motion and time-dependent NeRF stage has no fixed epoch count; the stopping criterion is only described as 'until convergence'.
  • SDF consistency loss schedule = lambda_4 = 0 for first 200 epochs, then gradually increased
    The ramp schedule for Lsdf is a hand-designed curriculum (Sec. 3.4).
  • gamma (SDF density sharpness in Eq. 1) = learnable
    Inherited from NeuS [29], gamma is a learnable parameter in the sigmoid that converts SDF to density; it is optimized, not fitted to external data.
assumptions (5)
  • domain assumption The scene is static and rigid; all temporal changes in the SDF are due to camera motion.
    Used in Lflow (Eq. 11) and Lphoto (Eq. 14); if false, camera motion estimates are biased. The limitation paragraph acknowledges low-texture and reflective regions.
  • domain assumption The SDF-flow relation from [18] is valid with the sign convention used in Eq. 11.
    The flow loss directly adopts ds/dt + (omega x x + v)^T n = 0 from Neural SDF Flow [18], a self-cited ICLR paper; the paper does not re-derive or validate the sign convention.
  • domain assumption Camera intrinsics are known and fixed.
    Stated in Sec. 3: 'We assume no additional information beyond camera intrinsics and images from the input video.' All projection operations (Eqs. 13 and 14) depend on K.
  • ad hoc to paper A smooth MLP mapping time to omega and v can represent the true camera trajectory, and Euler integration with U=10 sub-intervals is a sufficiently accurate approximation.
    The motion network phi_v (Eq. 3) and the discrete integration (Eqs. 5-6) are central to pose estimation; no guarantee is given for large accelerations or long sequences.
  • ad hoc to paper The time-dependent NeRF trained with neighboring frames provides a locally accurate geometry for estimating camera motion.
    The paper relies on this to make Lflow, Lphoto, and Lsdf meaningful; if the local geometry is wrong, the consistency losses are misaligned.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Joint Optimization of Neural Radiance Fields and Continuous Camera Motion from a Monocular Video." pith.science (2026). https://pith.science/paper/R64QR63R

@misc{pith2026250419819,
  author       = {Pith},
  title        = {Pith review of: Joint Optimization of Neural Radiance Fields and Continuous Camera Motion from a Monocular Video},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/R64QR63R}},
  note         = {Machine review of arXiv:2504.19819}
}
read the original abstract

Neural Radiance Fields (NeRF) has demonstrated its superior capability to represent 3D geometry but require accurately precomputed camera poses during training. To mitigate this requirement, existing methods jointly optimize camera poses and NeRF often relying on good pose initialisation or depth priors. However, these approaches struggle in challenging scenarios, such as large rotations, as they map each camera to a world coordinate system. We propose a novel method that eliminates prior dependencies by modeling continuous camera motions as time-dependent angular velocity and velocity. Relative motions between cameras are learned first via velocity integration, while camera poses can be obtained by aggregating such relative motions up to a world coordinate system defined at a single time step within the video. Specifically, accurate continuous camera movements are learned through a time-dependent NeRF, which captures local scene geometry and motion by training from neighboring frames for each time step. The learned motions enable fine-tuning the NeRF to represent the full scene geometry. Experiments on Co3D and Scannet show our approach achieves superior camera pose and depth estimation and comparable novel-view synthesis performance compared to state-of-the-art methods. Our code is available at https://github.com/HoangChuongNguyen/cope-nerf.

Figures

Figures reproduced from arXiv: 2504.19819 by the authors.

Figure 1
Figure 1. Comparison with previous works [2, 6]. Our method achieves superior performance in camera pose estimation (side), depth estimation (bottom) and novel-view synthesis (top). Abstract Neural Radiance Fields (NeRF) has demonstrated its su￾perior capability to represent 3D geometry but require accurately precomputed camera poses during training. To mitigate this requirement, existing methods jointly optimize camera poses… view at source ↗
Figure 2
Figure 2. Overview of our method. Left: We jointly optimize the camera motion network ϕv to obtain the continuous camera motion represented as angular velocity ω and velocity v, and a time-dependent NeRF (ϕg , ϕc) to obtain the scene geometry (represented as Signed Distance Field) and appearance at different time steps. The camera velocities can be integrated to obtain the camera transformation P between any two frames. Such … view at source ↗
Figure 3
Figure 3. Qualitative results on the Co3D (top) and Scannet (bottom) dataset. Our synthesized images are more photo-realistic compared to the other methods. In terms of geometry, our method produces the most accurate depth maps among all methods. Hydrant (Co3D) Plant (Co3D) Teddybear (Co3D) 0079 (Scannet) [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Camera trajectory visualization. Our poses are better aligned with the ground-truth compared to the others. method consistently outperforms other methods with notable margins across all scenes in both datasets. On average, our method reduces the depth square relative e…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 27 canonical work pages

  1. [18]

    Neural sdf flow for 3d reconstruction of dynamic scenes

    Wei Mao, Richard Hartley, Mathieu Salzmann, and Miaomiao Liu. Neural sdf flow for 3d reconstruction of dynamic scenes. In The Twelfth International Conference on Learning Representations. 2, 3, 5

    work page

  2. [1]

    Surf: Speeded up robust features

    Herbert Bay, Tinne Tuytelaars, and Luc Van Gool. Surf: Speeded up robust features. In Computer Vision–ECCV 2006: 9th European Conference on Computer Vision, Graz, Austria, May 7-13, 2006. Proceedings, Part I 9 , pages 404–417. Springer, 2006. 1

    work page 2006

  3. [2]

    Nope-nerf: Optimising neural radiance field with no pose prior

    Wenjing Bian, Zirui Wang, Kejie Li, Jia-Wang Bian, and Victor Adrian Prisacariu. Nope-nerf: Optimising neural radiance field with no pose prior. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4160–4169, 2023. 1, 2, 3, 4, 5, 6, 7, 8

    work page 2023

  4. [3]

    Tensorf: Tensorial radiance fields

    Anpei Chen, Zexiang Xu, Andreas Geiger, Jingyi Yu, and Hao Su. Tensorf: Tensorial radiance fields. In European conference on computer vision, pages 333–350. Springer,

    work page

  5. [4]

    Gaussian activated neural radiance fields for high fidelity reconstruction and pose estima- tion

    Shin-Fang Chng, Sameera Ramasinghe, Jamie Sherrah, and Simon Lucey. Gaussian activated neural radiance fields for high fidelity reconstruction and pose estima- tion. In European Conference on Computer Vision, pages 264–280. Springer, 2022. 3

    work page 2022

  6. [5]

    Scannet: Richly-annotated 3d reconstructions of indoor scenes

    Angela Dai, Angel X Chang, Manolis Savva, Ma- ciej Halber, Thomas Funkhouser, and Matthias Nießner. Scannet: Richly-annotated 3d reconstructions of indoor scenes. In Proceedings of the IEEE conference on com- puter vision and pattern recognition , pages 5828–5839,

    work page

  7. [6]

    Efros, and Xiaolong Wang

    Yang Fu, Sifei Liu, Amey Kulkarni, Jan Kautz, Alexei A. Efros, and Xiaolong Wang. Colmap-free 3d gaussian splatting. 2023. 1, 2, 3, 5, 6, 7, 8

    work page 2023

  8. [7]

    Dynamic view synthesis from dynamic monoc- ular video

    Chen Gao, Ayush Saraf, Johannes Kopf, and Jia-Bin Huang. Dynamic view synthesis from dynamic monoc- ular video. In Proceedings of the IEEE/CVF Inter- national Conference on Computer Vision , pages 5712– 5721, 2021. 2, 3

    work page 2021

Show all 38 references
  1. [8]

    Implicit geometric regularization for learning shapes

    Amos Gropp, Lior Yariv, Niv Haim, Matan Atzmon, and Yaron Lipman. Implicit geometric regularization for learning shapes. arXiv preprint arXiv:2002.10099, 2020. 5

    2002 arXiv

  2. [9]

    Sugar: Surface- aligned gaussian splatting for efficient 3d mesh recon- struction and high-quality mesh rendering

    Antoine Gu ´edon and Vincent Lepetit. Sugar: Surface- aligned gaussian splatting for efficient 3d mesh recon- struction and high-quality mesh rendering. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5354–5363, 2024. 2

    2024

  3. [10]

    2d gaussian splatting for geometri- cally accurate radiance fields

    Binbin Huang, Zehao Yu, Anpei Chen, Andreas Geiger, and Shenghua Gao. 2d gaussian splatting for geometri- cally accurate radiance fields. In ACM SIGGRAPH 2024 Conference Papers, pages 1–11, 2024. 2

    2024

  4. [11]

    Splatam: Splat track & map 3d gaussians for dense rgb-d slam

    Nikhil Keetha, Jay Karhade, Krishna Murthy Jataval- labhula, Gengshan Yang, Sebastian Scherer, Deva Ra- manan, and Jonathon Luiten. Splatam: Splat track & map 3d gaussians for dense rgb-d slam. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognitio...

    2024

  5. [12]

    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 Transactions on Graphics, 42(4), 2023. 2, 3

    2023

  6. [13]

    3d gaussian splatting as markov chain monte carlo

    Shakiba Kheradmand, Daniel Rebain, Gopal Sharma, Weiwei Sun, Jeff Tseng, Hossam Isack, Abhishek Kar, Andrea Tagliasacchi, and Kwang Moo Yi. 3d gaussian splatting as markov chain monte carlo. arXiv preprint arXiv:2404.09591, 2024. 3

    2024 arXiv

  7. [14]

    Kingma and Jimmy Ba

    Diederik P. Kingma and Jimmy Ba. Adam: A method for stochastic optimization, 2017. 6

    2017

  8. [15]

    Tanks and temples: Benchmarking large-scale scene reconstruction

    Arno Knapitsch, Jaesik Park, Qian-Yi Zhou, and Vladlen Koltun. Tanks and temples: Benchmarking large-scale scene reconstruction. ACM Transactions on Graphics (ToG), 36(4):1–13, 2017. 5

    2017

  9. [16]

    Barf: Bundle-adjusting neural radiance fields

    Chen-Hsuan Lin, Wei-Chiu Ma, Antonio Torralba, and Simon Lucey. Barf: Bundle-adjusting neural radiance fields. In Proceedings of the IEEE/CVF international conference on computer vision, pages 5741–5751, 2021. 2, 3, 4

    2021

  10. [17]

    Distinctive image features from scale- invariant keypoints

    David G Lowe. Distinctive image features from scale- invariant keypoints. International journal of computer vision, 60:91–110, 2004. 1

    2004

  11. [19]

    Hidenobu Matsuki, Riku Murai, Paul H. J. Kelly, and An- drew J. Davison. Gaussian Splatting SLAM. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024. 2, 3

    2024

  12. [20]

    Nerf: Representing scenes as neural radiance fields for view synthesis

    Ben Mildenhall, Pratul P Srinivasan, Matthew Tancik, Jonathan T Barron, Ravi Ramamoorthi, and Ren Ng. Nerf: Representing scenes as neural radiance fields for view synthesis. Communications of the ACM, 65(1):99– 106, 2021. 1, 2, 5, 6

    2021

  13. [21]

    Instant neural graphics primitives with a multiresolution hash encoding

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

    2022

  14. [22]

    Au- tomatic differentiation in pytorch

    Adam Paszke, Sam Gross, Soumith Chintala, Gregory Chanan, Edward Yang, Zachary DeVito, Zeming Lin, Alban Desmaison, Luca Antiga, and Adam Lerer. Au- tomatic differentiation in pytorch. 2017. 3

    2017

  15. [23]

    Vision transformers for dense prediction

    Ren ´e Ranftl, Alexey Bochkovskiy, and Vladlen Koltun. Vision transformers for dense prediction. InProceedings of the IEEE/CVF international conference on computer vision, pages 12179–12188, 2021. 3

    2021

  16. [24]

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

    Jeremy Reizenstein, Roman Shapovalov, Philipp Hen- zler, Luca Sbordone, Patrick Labatut, and David Novotny. Common objects in 3d: Large-scale learning and evaluation of real-life 3d category reconstruction. In Proceedings of the IEEE/CVF international conference on computer vi...

    2021

  17. [25]

    Splat-slam: Globally opti- mized rgb-only slam with 3d gaussians

    Erik Sandstr ¨om, Keisuke Tateno, Michael Oechsle, Michael Niemeyer, Luc Van Gool, Martin R Oswald, and Federico Tombari. Splat-slam: Globally opti- mized rgb-only slam with 3d gaussians. arXiv preprint arXiv:2405.16544, 2024. 2, 3

    2024 arXiv

  18. [26]

    Structure-from-motion revisited

    Johannes L Schonberger and Jan-Michael Frahm. Structure-from-motion revisited. In Proceedings of the IEEE conference on computer vision and pattern recog- nition, pages 4104–4113, 2016. 1, 2

    2016

  19. [27]

    Block- nerf: Scalable large scene neural view synthesis

    Matthew Tancik, Vincent Casser, Xinchen Yan, Sabeek Pradhan, Ben Mildenhall, Pratul P Srinivasan, Jonathan T Barron, and Henrik Kretzschmar. Block- nerf: Scalable large scene neural view synthesis. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogn...

    2022

  20. [28]

    Least-squares estimation of transfor- mation parameters between two point patterns

    Shinji Umeyama. Least-squares estimation of transfor- mation parameters between two point patterns. IEEE Transactions on Pattern Analysis & Machine Intelli- gence, 13(04):376–380, 1991. 6

    1991

  21. [29]

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

    Peng Wang, Lingjie Liu, Yuan Liu, Christian Theobalt, Taku Komura, and Wenping Wang. Neus: Learning neu- ral implicit surfaces by volume rendering for multi-view reconstruction. arXiv preprint arXiv:2106.10689, 2021. 1, 2, 3, 6, 8

    2021 arXiv

  22. [30]

    Image quality assessment: from error visibil- ity to structural similarity

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Simoncelli. Image quality assessment: from error visibil- ity to structural similarity. IEEE transactions on image processing, 13(4):600–612, 2004. 6

    2004

  23. [31]

    Nerf–: Neural radiance fields without known camera parameters

    Zirui Wang, Shangzhe Wu, Weidi Xie, Min Chen, and Victor Adrian Prisacariu. Nerf–: Neural radiance fields without known camera parameters. arXiv preprint arXiv:2102.07064, 2021. 2, 3, 6, 7, 8

    2021 arXiv

  24. [32]

    Gs-slam: Dense visual slam with 3d gaussian splatting

    Chi Yan, Delin Qu, Dan Xu, Bin Zhao, Zhigang Wang, Dong Wang, and Xuelong Li. Gs-slam: Dense visual slam with 3d gaussian splatting. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 19595–19604, 2024. 2, 3

    2024

  25. [33]

    V olume rendering of neural implicit surfaces

    Lior Yariv, Jiatao Gu, Yoni Kasten, and Yaron Lip- man. V olume rendering of neural implicit surfaces. Ad- vances in Neural Information Processing Systems , 34: 4805–4815, 2021. 1, 2

    2021

  26. [34]

    inerf: Inverting neural radiance fields for pose estimation

    Lin Yen-Chen, Pete Florence, Jonathan T Barron, Al- berto Rodriguez, Phillip Isola, and Tsung-Yi Lin. inerf: Inverting neural radiance fields for pose estimation. in 2021 ieee. In RSJ International Conference on Intelli- gent Robots and Systems (IROS), pages 1323–1330. 3

    2021

  27. [35]

    Gaus- sian opacity fields: Efficient and compact surface re- construction in unbounded scenes

    Zehao Yu, Torsten Sattler, and Andreas Geiger. Gaus- sian opacity fields: Efficient and compact surface re- construction in unbounded scenes. arXiv preprint arXiv:2404.10772, 2024. 2, 3

    2024 arXiv

  28. [36]

    Rade-gs: Ras- terizing depth in gaussian splatting

    Baowen Zhang, Chuan Fang, Rakesh Shrestha, Yixun Liang, Xiaoxiao Long, and Ping Tan. Rade-gs: Ras- terizing depth in gaussian splatting. arXiv preprint arXiv:2406.01467, 2024. 2

    2024 arXiv

  29. [37]

    The unreasonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shecht- man, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 586–595, 2018. 6

    2018

  30. [38]

    Unsupervised learning of depth and ego- motion from video

    Tinghui Zhou, Matthew Brown, Noah Snavely, and David G Lowe. Unsupervised learning of depth and ego- motion from video. In Proceedings of the IEEE confer- ence on computer vision and pattern recognition , pages 1851–1858, 2017. 6

    2017

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.