Pith. sign in

REVIEW 3 major objections 5 minor 38 references

Time of the Flight of the Gaussians: Optimizing Depth Indirectly in Dynamic Radiance Fields

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read A single continuous-wave time-of-flight camera can reconstruct dynamic scenes about 100 times faster than neural volumetric baselines, with similar or better depth accuracy.

desk verdict Solid engineering with a real 100x speedup and honest limitations, but the accuracy claim is softer than the abstract's 'similar or better' on exactly the fast, nonlinear motions the paper highlights. read the letter →

arxiv 2505.05356 v1 pith:O6RTUEYI submitted 2025-05-08 cs.GR cs.AIcs.CV

Runfeng Li , Mikhail Okunev , Zixuan Guo , Anh Ha Duong , Christian Richardt , Matthew O'Toole , James Tompkin This is my paper · ORCID
classification cs.GRcs.AIcs.CV
keywords continuous-wavetime-of-flightGaussiansplattingdynamicscenereconstructionmonocular4Drawsensorquadsradiancefieldsdepthfromphaseoptical-flowweaksupervision
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 make monocular dynamic 3D reconstruction from a continuous-wave time-of-flight (C-ToF) camera practical. C-ToF cameras return four raw intensity samples per pixel, and depth must be inferred from their phase, so the property the optimization cares about—depth—is only an indirect byproduct of fitting those samples. The authors show that replacing the slow neural volumes of prior work with 3D Gaussian splatting makes this indirect optimization brittle, and they add two simple optimization biases—slowing the reflectivity learning rate and initializing reflectivity low—that keep Gaussians placed on true surfaces. On synthetic and real-world scenes, the method matches or improves depth accuracy relative to the neural baselines while cutting optimization and rendering time by about 100x, including on fast motion like a swinging baseball bat. If correct, it turns a single commodity C-ToF camera into a near-real-time dynamic 3D capture system.

What carries the argument

The load-bearing object is a differentiable C-ToF Gaussian rasterizer that turns each Gaussian's position, opacity, and spherical-harmonic reflectivity into the four raw sensor samples, rather than into a color image. The key identity is the raw-quad formation model $q(x) = q_{bg}T_N + \sum_k \frac{s r_k}{d_k^2}\varphi(d_k) o_k G^{2D}_k(x) T_k^2$, where $\varphi(d_k)$ is the sine/cosine phase vector and $T_k$ is squared transmittance; fitting this quantity is what makes depth indirect. A deformation MLP maps canonical Gaussian positions to per-timestep offsets, and between integer timesteps positions are linearly interpolated, enforcing the local-linearity that matches the asynchronous capture of the four quads. Two heuristics carry the optimization: a tenfold-lower reflectivity learning rate (occupancy bias) and a low reflectivity initialization, which together replace a depth-distortion loss and keep Gaussians concentrated at the true surface.

What would settle it

Render a synthetic scene with ground-truth depth and a known nonlinear trajectory—say a fan blade rotating with constant angular acceleration—through the raw-quad model, optimize with this method, and compare mean rendered depth to ground truth. If depth error remains at the level the paper reports for the curved and spinning test scenes and does not shrink with more iterations or smaller time steps, the piecewise-linear motion assumption is the binding limitation of the claimed accuracy.

Watch

Extended reading notes

Core claim

The paper's central claim is that raw C-ToF samples—not recovered depth maps—are the right supervision signal for a Gaussian splatting reconstruction, provided the optimization is conditioned so that Gaussians behave like an opaque surface. The sensor model renders each Gaussian's returned-light contribution, including the inverse-square falloff and the sinusoidal phase modulation, into the four raw quad images; depth is then read out as the mean Gaussian depth along each ray. Fitting these quads is shown to be under-constrained: many multi-peak density arrangements reproduce the measurements while producing wrong depth, and Gaussian splatting lacks the implicit low-entropy bias of MLP volumetric fields. The paper's fix is two-fold: lower the reflectivity learning rate tenfold so the optimizer prefers moving, adding, or removing Gaussians over changing their reflectivity (occupancy bias), and initialize reflectivity near 0.1 so low-reflectivity surfaces are not pushed to incorrect far depths. With these changes, plus a piecewise-linear deformation model and optical-flow weak supervision, the method reports depth errors below both neural volumetric baselines on five of seven synthetic scenes, competitive results on the remaining two, and 40–60 minute optimizations rather than 60–72 hours.

Load-bearing premise

The scene's motion is approximately piecewise linear between integer timesteps, with Gaussian positions at fractional times obtained by linear interpolation; strongly curved or rotational motion, which the curved and spinning test scenes exhibit, breaks this assumption and degrades reconstruction.

Editorial extensions

If this is right

  • A static monocular C-ToF camera can reconstruct dynamic scenes—including fast motion such as a swinging baseball bat—that previously required either a moving camera, extra RGB constraints, or tens of hours of neural volumetric optimization.
  • Optimization time drops from 60–72 hours to 40–60 minutes per sequence, and rendering runs in real time above 100 Hz, making iterative capture-and-analysis loops practical.
  • Fitting raw C-ToF quads directly, without a depth-distortion loss, yields depth estimates that are better on five of seven synthetic scenes and competitive on the remaining two.
  • The two heuristics are necessary: ablations with neither heuristic fail catastrophically, and adding the depth-distortion loss on top of the heuristics reintroduces oversmoothed depth and instability.
  • Depth wrapping in static-camera C-ToF backgrounds remains an ill-posed ambiguity that no tested method resolves.

Reading between the lines

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

  • Editorial inference: The same 'optimize the raw measurement, bias the optimizer toward occupancy' recipe could transfer to other indirect-imaging modalities—structured light, sonar, X-ray, or snapshot lidar—where the desired property is a byproduct of fitting the sensor signal; the paper does not claim this transfer.
  • Editorial inference: Replacing the piecewise-linear interpolation with a higher-order or per-Gaussian trajectory model, possibly regularized by rigidity, is the most direct route to the nonlinear-motion failures the paper reports; this is a testable extension, not a result of the paper.
  • Editorial inference: If the speed and accuracy hold at scale, a single C-ToF camera becomes a viable low-cost dynamic capture input for telepresence and measurement, but practical deployment would need to address the depth-wrapping ambiguity caused by a static camera.
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

3 major / 5 minor

Summary. This paper proposes a dynamic 3D Gaussian splatting method for monocular continuous-wave time-of-flight (C-ToF) cameras, operating directly on raw asynchronous sensor quads rather than derived depth. The scene is represented by canonical Gaussians deformed by an MLP, with intermediate raw-frame motion assumed piecewise linear, and the optimization is stabilized by two heuristics: an occupancy bias implemented as a 10x lower reflectivity learning rate, and a low initial reflectivity bias. The method is evaluated on the T"oRF real-world dataset, the F-T"oRF real-world dataset, and the F-T"oRF synthetic dataset, reporting depth errors competitive with or better than the neural volumetric baselines while being roughly two orders of magnitude faster to optimize and render in real time.

Significance. If the claims are substantiated, this is a practically important result: it would make single-camera dynamic C-ToF reconstruction feasible at near-interactive optimization times and real-time rendering, which no prior method achieves. The paper also gives a clear and honest analysis of why Gaussian-splatting optimization is brittle when depth is only indirectly supervised, and it identifies two simple, reproducible biases that substantially improve stability. The ablations in the supplement are valuable, documenting that the method without the heuristics fails dramatically and that the depth distortion loss is unstable in this setting. However, the central accuracy claim currently rests on a single run per scene, and the two heuristics are explicitly acknowledged to have no universal guarantees and appear to have been selected largely on the basis of the same benchmark scenes used for evaluation. These issues bound the strength of the claims as currently stated.

major comments (3)
  1. [Section 5.1, Table 1] All quantitative depth errors are reported for a single optimization run per scene, with no error bars or repeated-seed statistics. This is load-bearing because the paper's own motivation is that the C-ToF optimization is under-constrained and brittle, and because the discussion invokes scene-specific optimization failures, e.g., the local optimum on one cube in Orthogonal Speed Test. A single run cannot distinguish a robust method from one that happens to converge on these seven synthetic scenes. I ask the authors to report means and standard deviations over at least three seeds, and to include the per-seed spread for the scenes where the claim of 'similar or better accuracy' is made.
  2. [Section 4, Eq. (7); Section 1; Section 6] The deformation model forces Gaussian positions at fractional raw-frame times to be linear interpolations between integer-time positions. The paper's own results show that this assumption is violated on exactly the scenes with nonlinear motion: in Table 1, our method is worse than F-T"oRF in dToF on Arcing Cube (1.060 vs 0.470) and 3 Chairs Speed Test (0.639 vs 0.324), and the limitations section states that strongly nonlinear motion, as in Fan and partly Arcing Cube, causes the model to struggle. Yet the abstract and introduction advertise accurate reconstruction for fast motions like swinging baseball bats, which are predominantly rotational. This is not a disagreement with current consensus but a mismatch between the method's modeling assumption and the scope of the central claim. The claim should either be restricted to approximately piecewise-linear motions, or the paper should include an experiment or discussion quantifying when the piecewise-linear assumption is adequate (e.g., the real Baseball sequence appears to work well despite the bat's rotation, so the paper should explain why the synthetic rotational scenes fail while the real bat scene does not).
  3. [Section 3.2, Heuristic 1 and Heuristic 2; Section 6, No universal guarantees] The two heuristics are introduced as general remedies for the under-constrained C-ToF optimization, but the paper explicitly acknowledges that they are based on empirical observations and may not be suitable for different scenes. Because the heuristics and hyperparameters (initial reflectivity 0.1, 10x reflectivity learning rate reduction, random background range [-1,1], loss weights alpha and beta) were selected using the same benchmark suites that are used for the final evaluation, the quantitative result overstates the expected accuracy on new scenes. This is a form of evaluation-set adaptation. I would like to see at least one held-out generalization check, e.g., a new capture or a synthetic scene with different object materials and motion types, along with a small sensitivity analysis over the two key heuristic parameters. Without this, the claim that the method 'produces accurate reconstructions under constrained C-ToF sensing conditions' remains specific to the tested datasets rather than to the general setting claimed in the title and abstract.
minor comments (5)
  1. [Section 4, Eq. (8)] The symbol x is used both for a 3D world position in MLP(x, i) and for a pixel coordinate in G_k^{2D}(x), which makes Eq. (8) unnecessarily hard to read; please use distinct notations, e.g., p for 3D points and u for pixel positions.
  2. [Section 2, Eq. (1)] The denominator in the arctan formula appears as Q_{\pi/2}-Q_{3\pi/2} without subscripts being clearly visually separated in the typeset equation; a minor formatting fix would improve readability.
  3. [Section 4, Random background] Randomizing the background quad uniformly in [-1,1] at every iteration is an unusual and aggressive regularizer, and the paper does not discuss whether it biases the recovered background phase or amplitude; a short justification or ablation would help.
  4. [Section 5.1, Metrics] For the real-world datasets, only qualitative results are shown; a quantitative real-world evaluation would strengthen the paper, though I understand that ground-truth dynamic depth is not available for these captures.
  5. [Section 3.1, Eq. (6)] The explanation of why Jensen's inequality does not lead to equality is phrased loosely ('sine and cosine functions oscillate between convex and concave regions'); consider replacing this with a one-sentence algebraic statement that W(d_k) is not affine in d_k, so the finite sum of W(d_k) does not equal W of the mean depth.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the depth metric d is independent of the fitted quad loss, and self-citations are not load-bearing.

full rationale

The derivation chain is self-contained with respect to the input measurements. The image-formation model (Eqs. 3-4) is a direct phasor/quad extension of standard C-ToF physics, not a restatement of the depth values being evaluated. The training objective (Eq. 9) fits raw quads and optional color, and the two reported depth metrics are distinct: d (Eq. 5) is a volume-rendered mean Gaussian depth that is not a deterministic function of the quad fit, and the paper explicitly demonstrates the failure mode where quad fitting succeeds but d is wrong (Fig. 2, Sec. 3.1). Thus the d results, which carry the main accuracy claim, are not forced by the loss. The dToF metric is derived from reconstructed quads, but it is transparently labeled as such in Table 1 and is a standard reconstruction-fidelity metric rather than a claimed out-of-sample prediction. Self-citations to TöRF and F-TöRF provide the benchmark datasets, baselines, and the phase-aware interpolation idea (Eq. 7), but none is used as a uniqueness theorem or as the sole justification for a load-bearing premise; the motion-model assumption is stated as an assumption, and the paper's own limitations section concedes its failure on nonlinear motion and admits 'no universal guarantees.' No equation reduces to its inputs by construction.

Assumptions & free parameters 6 free parameters · 6 assumptions · 0 invented entities

The method is an empirical optimization pipeline. The main free parameters are optimization choices (initial reflectivity, learning rates, loss weights, warm-up length, background randomization) selected by hand on the authors' benchmark scenes; the ablation table shows depth errors blow up without them. The domain assumptions are the standard C-ToF volumetric model and the piecewise-linear motion model. No new physical entities are introduced.

free parameters (6)
  • Initial reflectivity = 0.1
    Low initial reflectivity (H2) is chosen empirically to keep low-reflectivity Gaussians from being pushed far away by the inverse-square falloff term (Sec. 3.1).
  • Reflectivity learning rate ratio = 10x lower than default (0.00016)
    Occupancy bias (H1) is implemented by lowering the reflectivity learning rate so position and opacity dominate; the factor 10 is chosen by hand (Sec. 3.1).
  • Loss weight alpha = 5 (1 for synthetic)
    Balances raw quad loss and optional color loss; chosen experimentally (Sec. 4, Hyperparameters).
  • Optical flow loss weight beta = 0.0008
    Weak supervision strength for motion offsets; tuned empirically (Sec. 4).
  • Warm-up iterations = 2000
    Static warm-up stage to initialize canonical Gaussian positions; selected by hand (Sec. 4).
  • Random background range = Uniform [-1, 1]
    Random background signal each iteration avoids low-contributing Gaussians in low-reflectivity areas; an ad hoc regularizer (Sec. 4).
assumptions (6)
  • domain assumption C-ToF signal follows the volumetric transmission model of Eq. (4): incoherent sinusoid modulation, co-located emitter, inverse-square falloff, and squared transmittance.
    Sec. 3, Eqs. (3)-(4): the method and baselines all assume this image formation model; multi-path is only modeled via volumetric superposition, not full wave optics.
  • domain assumption Motion is piecewise linear between integer timesteps; positions at fractional times are linearly interpolated (Eq. 7).
    Sec. 4: required to render asynchronous raw quads; the Fan and Arcing Cube results show failure when this assumption is violated.
  • domain assumption RAFT optical flow is accurate enough to weakly supervise 3D Gaussian motion offsets (Eq. 8).
    Sec. 4: flow errors propagate into the deformation MLP; the authors acknowledge this in the assumptions list.
  • ad hoc to paper The two heuristics (occupancy bias and low-reflectivity bias) suffice to make the under-constrained optimization converge to accurate depth.
    Sec. 3.1 and Sec. 6: the authors state these are empirical observations and give no universal guarantee; ablation Table 2 shows they are essential for the reported results.
  • ad hoc to paper Randomizing the background quad each iteration in [-1,1] is a valid regularizer that does not bias the final reconstruction.
    Sec. 4: used to prevent low-contributing Gaussians; no theoretical justification beyond avoiding a local minimum.
  • domain assumption Accurate camera poses are available and align scene scale with C-ToF measurements.
    Sec. 4: the pipeline assumes known poses; for the ToRF dataset, poses come from the ToRF model.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Time of the Flight of the Gaussians: Optimizing Depth Indirectly in Dynamic Radiance Fields." pith.science (2026). https://pith.science/paper/O6RTUEYI

@misc{pith2026250505356,
  author       = {Pith},
  title        = {Pith review of: Time of the Flight of the Gaussians: Optimizing Depth Indirectly in Dynamic Radiance Fields},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/O6RTUEYI}},
  note         = {Machine review of arXiv:2505.05356}
}
read the original abstract

We present a method to reconstruct dynamic scenes from monocular continuous-wave time-of-flight (C-ToF) cameras using raw sensor samples that achieves similar or better accuracy than neural volumetric approaches and is 100x faster. Quickly achieving high-fidelity dynamic 3D reconstruction from a single viewpoint is a significant challenge in computer vision. In C-ToF radiance field reconstruction, the property of interest-depth-is not directly measured, causing an additional challenge. This problem has a large and underappreciated impact upon the optimization when using a fast primitive-based scene representation like 3D Gaussian splatting, which is commonly used with multi-view data to produce satisfactory results and is brittle in its optimization otherwise. We incorporate two heuristics into the optimization to improve the accuracy of scene geometry represented by Gaussians. Experimental results show that our approach produces accurate reconstructions under constrained C-ToF sensing conditions, including for fast motions like swinging baseball bats. https://visual.cs.brown.edu/gftorf

Figures

Figures reproduced from arXiv: 2505.05356 by the authors.

Figure 1
Figure 1. Measuring the depth of a fast-moving object is challenging [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Fitting C-ToF images ̸= fitting depth. Top left: Camera-derived depth from C-ToF. Top right: Rendering a GS scene reconstruction into C-ToF raw image samples, then deriving depth. As this is similar to the camera-derived depth to the left, the reconstruction objective was met. Bottom left: Rendered mean scene depth from Gaussians, which is highly inaccurate. Bottom right: Depth distortion error [10], which measures … view at source ↗
Figure 3
Figure 3. Pipeline. Left: We capture input raw quads (or phasors, not shown) from a continuous-wave time-of-flight camera with optional color camera.Right to Left: From randomly initialized Gaussians, the warm-up stage estimates canonical scene geometry for a static scene. Then, given time t, the MLP predicts offsets (δxk) that reposition the canonical Gaussians. Then, we render the C-ToF and color images and compute losses. … view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Our approach is competitive or better in terms of accuracy against the state of the art on synthetic scenes, while being two orders of magnitude faster. We use the synthetic dataset from F-ToRF to demonstrate this. All images show rendered volumetric depth ¨ d. State-o…
Figure 5
Figure 5. Figure 5: F-ToRF real scenes; rendered scene depth ¨ d. As it models temporal dynamics and constrains geometry appropriately during training, our model produces comparable quality reconstructions with F-ToRF while being consistently better than other baselines. Our method ¨ tend…
Figure 6
Figure 6. Figure 6: Our approach is competitive with other baselines on ToRF dataset. ¨ Comparison of our method with ToRF and DeformableGS ¨ (with and without C-ToF depth prior). Images are novel views of rendered volumetric depth d and color, rendered along a spiral path around the trai…
Figure 7
Figure 7. Figure 7: Ablations. In the first column, the even rows show the reflectivity map, which is computed as input amplitude multiplied with the square of input depth (light falloff), and can be understood as the expected of Gaussian reflectivity at the corresponding surface. For vis…
Figure 8
Figure 8. Figure 8: Results on all F-ToRF synthetic scenes ¨ . DeformableGS is given C-ToF-derived depth as an additional input. Inset on ground truth: corresponding RGB image. Full RGB images can be found in [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]
Figure 9
Figure 9. Figure 9: Results on all F-ToRF real-world scenes ¨ . DeformableGS is given C-ToF-derived depth as an additional input. 15 [PITH_FULL_IMAGE:figures/full_fig_p015_9.png]
Figure 10
Figure 10. Figure 10: More results on ToRF real-world scenes ¨ . DeformableGS is given C-ToF-derived depth as an additional input. 16 [PITH_FULL_IMAGE:figures/full_fig_p016_10.png]
Figure 11
Figure 11. Figure 11: Corresponding RGB color images for the synthetic scenes for Fig. (4) in the main paper and [PITH_FULL_IMAGE:figures/full_fig_p017_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 36 canonical work pages

  1. [1]

    T¨oRF: Time-of-flight radiance fields for dynamic scene view synthesis

    Benjamin Attal, Eliot Laidlaw, Aaron Gokaslan, Changil Kim, Christian Richardt, James Tompkin, and Matthew O’Toole. T¨oRF: Time-of-flight radiance fields for dynamic scene view synthesis. In NeurIPS, 2021. 1, 3, 5, 6, 7, 8, 14, 15, 16

    work page 2021

  2. [2]

    Radiative Gaussian splatting for efficient X-ray novel view synthesis

    Yuanhao Cai, Yixun Liang, Jiahao Wang, Angtian Wang, Yulun Zhang, Xiaokang Yang, Zongwei Zhou, and Alan Yuille. Radiative Gaussian splatting for efficient X-ray novel view synthesis. In ECCV, 2024. 8

    work page 2024

  3. [3]

    pixelSplat: 3D Gaussian splats from image pairs for scalable generalizable 3D reconstruction

    David Charatan, Sizhe Lester Li, Andrea Tagliasacchi, and Vincent Sitzmann. pixelSplat: 3D Gaussian splats from image pairs for scalable generalizable 3D reconstruction. InCVPR,

    work page

  4. [4]

    Thermal3D-GS: Physics-induced 3D Gaussians for thermal infrared novel-view synthesis

    Qian Chen, Shihao Shu, and Xiangzhi Bai. Thermal3D-GS: Physics-induced 3D Gaussians for thermal infrared novel-view synthesis. In ECCV, 2024. 8

    work page 2024

  5. [5]

    Depth-regularized optimization for 3D Gaussian splatting in few-shot images

    Jaeyoung Chung, Jeongtaek Oh, and Kyoung Mu Lee. Depth-regularized optimization for 3D Gaussian splatting in few-shot images. In CVPR, pages 811–820, 2024. 3, 8

    work page 2024

  6. [6]

    Snapshot lidar: Fourier embedding of amplitude and phase for single-image depth reconstruction

    Sarah Friday, Yunzi Shi, Yaswanth Cherivirala, Vishwanath Saragadam, and Adithya Pediredla. Snapshot lidar: Fourier embedding of amplitude and phase for single-image depth reconstruction. In CVPR, 2024. 7

    work page 2024

  7. [7]

    SuGaR: Surface-aligned Gaussian splatting for efficient 3D mesh reconstruction and high-quality mesh rendering

    Antoine Gu´edon and Vincent Lepetit. SuGaR: Surface-aligned Gaussian splatting for efficient 3D mesh reconstruction and high-quality mesh rendering. In CVPR, 2024. 7

    work page 2024

  8. [8]

    Phasor imaging: A generalization of correlation-based time-of-flight imaging

    Mohit Gupta, Shree K Nayar, Matthias B Hullin, and Jaime Martin. Phasor imaging: A generalization of correlation-based time-of-flight imaging. ACM Transactions on Graphics (ToG), 34(5):1–18, 2015. 2, 3

    work page 2015

Show all 38 references
  1. [9]

    Time-of-Flight Cameras: Principles, Methods and Applications

    Miles Hansard, Seungkyu Lee, Ouk Choi, and Radu Ho- raud. Time-of-Flight Cameras: Principles, Methods and Applications. Springer, 2012. 2

    2012

  2. [10]

    2D Gaussian splatting for geometrically accurate radiance fields

    Binbin Huang, Zehao Yu, Anpei Chen, Andreas Geiger, and Shenghua Gao. 2D Gaussian splatting for geometrically accurate radiance fields. In SIGGRAPH, 2024. 2, 3, 7, 11

    2024

  3. [11]

    Lighting every darkness with 3DGS: Fast training and real-time rendering for HDR view synthesis

    Xin Jin, Pengyi Jiao, Zheng-Peng Duan, Xingchao Yang, Chun-Le Guo, Bo Ren, and Chongyi Li. Lighting every darkness with 3DGS: Fast training and real-time rendering for HDR view synthesis. InNeurIPS, 2024. 8

    2024

  4. [12]

    A compact dynamic 3D Gaussian representation for real-time dynamic view synthesis

    Kai Katsumata, Duc Minh V o, and Hideki Nakayama. A compact dynamic 3D Gaussian representation for real-time dynamic view synthesis. In ECCV, 2024. 8

    2024

  5. [13]

    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. 1, 2, 3, 5, 7, 11

    2023

  6. [14]

    PlatoNeRF: 3D reconstruction in Plato’s cave via single-view two-bounce lidar

    Tzofi Klinghoffer, Xiaoyu Xiang, Siddharth Somasundaram, Yuchen Fan, Christian Richardt, Ramesh Raskar, and Rakesh Ranjan. PlatoNeRF: 3D reconstruction in Plato’s cave via single-view two-bounce lidar. InCVPR, 2024. 7

    2024

  7. [15]

    DynMF: Neural motion factorization for real-time dynamic view synthesis with 3D Gaussian splatting

    Agelos Kratimenos, Jiahui Lei, and Kostas Daniilidis. DynMF: Neural motion factorization for real-time dynamic view synthesis with 3D Gaussian splatting. InECCV, 2024. 8

    2024

  8. [16]

    Light-in-flight for a world-in-motion

    Jongho Lee, Ryan J Suess, and Mohit Gupta. Light-in-flight for a world-in-motion. In European Conference on Computer Vision, pages 204–220. Springer, 2024. 7

    2024

  9. [17]

    DNGaussian: Optimizing sparse-view 3D Gaussian radiance fields with global-local depth normalization

    Jiahe Li, Jiawei Zhang, Xiao Bai, Jin Zheng, Xin Ning, Jun Zhou, and Lin Gu. DNGaussian: Optimizing sparse-view 3D Gaussian radiance fields with global-local depth normalization. In CVPR, 2024. 1, 3

    2024

  10. [18]

    Spacetime Gaussian feature splatting for real-time dynamic view synthesis

    Zhan Li, Zhang Chen, Zhong Li, and Yi Xu. Spacetime Gaussian feature splatting for real-time dynamic view synthesis. In CVPR, 2024. 8

    2024

  11. [19]

    From chaos to clarity: 3DGS in the dark

    Zhihao Li, Yufei Wang, Alex Kot, and Bihan Wen. From chaos to clarity: 3DGS in the dark. InNeurIPS, 2024. 8

    2024

  12. [20]

    Monocular dynamic Gaussian splatting is fast and brittle but smooth motion helps, 2024

    Yiqing Liang, Mikhail Okunev, Mikaela Angelina Uy, Runfeng Li, Leonidas Guibas, James Tompkin, and Adam W Harley. Monocular dynamic Gaussian splatting is fast and brittle but smooth motion helps, 2024. arXiv:2412.04457. 1

    2024 arXiv

  13. [21]

    GauFRe: Gaussian deformation fields for real-time dynamic novel view synthesis

    Yiqing Liang, Numair Khan, Zhengqin Li, Thu Nguyen-Phuoc, Douglas Lanman, James Tompkin, and Lei Xiao. GauFRe: Gaussian deformation fields for real-time dynamic novel view synthesis. In WACV, 2025. 1, 2, 8

    2025

  14. [22]

    Gaussian- Flow: 4D reconstruction with dynamic 3D Gaussian particle

    Youtian Lin, Zuozhuo Dai, Siyu Zhu, and Yao Yao. Gaussian- Flow: 4D reconstruction with dynamic 3D Gaussian particle. In CVPR, 2023. 8

    2023

  15. [23]

    Dynamic 3D Gaussians: Tracking by persistent dynamic view synthesis

    Jonathon Luiten, Georgios Kopanas, Bastian Leibe, and Deva Ramanan. Dynamic 3D Gaussians: Tracking by persistent dynamic view synthesis. In 3DV, 2024. 8

    2024

  16. [24]

    Lindell, and Kiriakos N

    Parsa Mirdehghan, Maxx Wu, Wenzheng Chen, David B. Lindell, and Kiriakos N. Kutulakos. Turbosl: Dense accurate and fast 3d by neural inverse structured light. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 25067–25076, 2024. 7

    2024

  17. [25]

    Softmax splatting for video frame interpolation

    Simon Niklaus and Feng Liu. Softmax splatting for video frame interpolation. In CVPR, 2020. 5

    2020

  18. [26]

    Flowed time of flight radiance fields

    Mikhail Okunev, Marc Mapeke, Benjamin Attal, Christian Richardt, Matthew O’Toole, and James Tompkin. Flowed time of flight radiance fields. InECCV, 2024. 1, 2, 3, 4, 5, 6, 7, 14, 15

    2024

  19. [27]

    Elsevier, 3rd edition, 2016

    Matt Pharr, Wenzel Jakob, and Greg Humphreys.Physically Based Rendering: From Theory to Implementation. Elsevier, 3rd edition, 2016. 5

    2016

  20. [28]

    Z-splat: Z-axis Gaussian splatting for camera-sonar fusion

    Ziyuan Qu, Omkar Vengurlekar, Mohamad Qadri, Kevin Zhang, Michael Kaess, Christopher Metzler, Suren Jayasuriya, and Adithya Pediredla. Z-splat: Z-axis Gaussian splatting for camera-sonar fusion. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024. 8

    2024

  21. [29]

    Neural fields for structured lighting

    Aarrushi Shandilya, Benjamin Attal, Christian Richardt, James Tompkin, and Matthew O’Toole. Neural fields for structured lighting. In ICCV, 2023. 7

    2023

  22. [30]

    Dynamic Gaussian marbles for novel view synthesis of casual monocular videos

    Colton Stearns, Adam W Harley, Mikaela Uy, Florian Dubost, Federico Tombari, Gordon Wetzstein, and Leonidas Guibas. Dynamic Gaussian marbles for novel view synthesis of casual monocular videos. In SIGGRAPH Asia, 2024. 7, 8

    2024

  23. [31]

    RAFT: Recurrent all-pairs field transforms for optical flow

    Zachary Teed and Jia Deng. RAFT: Recurrent all-pairs field transforms for optical flow. InECCV, 2020. 4

    2020

  24. [32]

    Shape of motion: 4D reconstruction from a single video

    Qianqian Wang, Vickie Ye, Hang Gao, Jake Austin, Zhengqi Li, and Angjoo Kanazawa. Shape of motion: 4D reconstruction from a single video. arXiv:2407.13764, 2024. 8

    2024

  25. [33]

    4D Gaussian splatting for real-time dynamic scene rendering

    Guanjun Wu, Taoran Yi, Jiemin Fang, Lingxi Xie, Xiaopeng Zhang, Wei Wei, Wenyu Liu, Qi Tian, and Xinggang Wang. 4D Gaussian splatting for real-time dynamic scene rendering. In CVPR, 2024. 8

    2024

  26. [34]

    PhysGaussian: Physics-integrated 3D Gaussians for generative dynamics

    Tianyi Xie, Zeshun Zong, Yuxing Qiu, Xuan Li, Yutao Feng, Yin Yang, and Chenfanfu Jiang. PhysGaussian: Physics-integrated 3D Gaussians for generative dynamics. In CVPR, 2024. 7 9

    2024

  27. [35]

    GRM: Large Gaussian reconstruction model for efficient 3D reconstruction and generation

    Yinghao Xu, Zifan Shi, Wang Yifan, Hansheng Chen, Ceyuan Yang, Sida Peng, Yujun Shen, and Gordon Wetzstein. GRM: Large Gaussian reconstruction model for efficient 3D reconstruction and generation. In ECCV, 2024. 1

    2024

  28. [36]

    Deformable 3D Gaussians for high-fidelity monocular dynamic scene reconstruction

    Ziyi Yang, Xinyu Gao, Wen Zhou, Shaohui Jiao, Yuqing Zhang, and Xiaogang Jin. Deformable 3D Gaussians for high-fidelity monocular dynamic scene reconstruction. In CVPR, 2024. 1, 2, 4, 5, 6, 7, 8, 14, 15, 16

    2024

  29. [37]

    Real-time photorealistic dynamic scene representation and rendering with 4D Gaussian splatting

    Zeyu Yang, Hongye Yang, Zijie Pan, and Li Zhang. Real-time photorealistic dynamic scene representation and rendering with 4D Gaussian splatting. InICLR, 2024. 8

    2024

  30. [38]

    No bias” refers to no consideration of either of our heuristics. “DD

    Zehao Zhu, Zhiwen Fan, Yifan Jiang, and Zhangyang Wang. FSGS: Real-time few-shot view synthesis using Gaussian splatting. In ECCV, 2024. 1 10 Time of the Flight of the Gaussians: Optimizing Depth Indirectly in Dynamic Radiance Fields Supplementary Material A. Supplemental Vide...

    2024

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.