Pith. sign in

REVIEW 4 major objections 5 minor 42 references

STD-GS: Exploring Frame-Event Interaction for SpatioTemporal-Disentangled Gaussian Splatting to Reconstruct High-Dynamic Scene

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper claims that separately modeling static background and dynamic objects as Gaussians, guided by frame and event data, reconstructs time-continuous high-dynamic scenes with sharper novel views.

desk verdict Plausible event-guided disentanglement for dynamic scenes, with solid ablations and a useful new dataset, but the paper overstates its 'discovery' of Gaussian-event consistency and should clarify the clustering supervision and report error bars. read the letter →

arxiv 2506.23157 v1 pith:DKBY77IA submitted 2025-06-29 cs.CV

Hanyu Zhou , Haonan Wang , Haoyue Liu , Yuxing Duan , Luxin Yan , Gim Hee Lee This is my paper · ORCID
classification cs.CV
keywords Gaussiansplattingeventcameradynamicscenereconstructionspatiotemporaldisentanglementnovelviewsynthesisappearance-motionclustering4Dframe-eventfusion
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 tries to show that high-dynamic scene reconstruction fails when one Gaussian model is asked to represent both rigid background and deformed moving objects, and that separating the two before rendering fixes it. It claims that frame images supply dense spatial appearance while event cameras supply continuous temporal motion, so the two modalities can be combined to distinguish background from objects by clustering. The method then represents the background as 3D Gaussians and each dynamic object as 4D Gaussians centered on event-tracked points, and uses event data as pseudo-labels to constrain the object's color and deformation. A sympathetic reader would care because the reported numbers, for example 27.12 dB PSNR versus 20.60 dB for the best frame-based competitor on the HD-CED data, suggest that the disentanglement removes the blur and position shift that plague unified dynamic-scene methods.

What carries the argument

The load-bearing object is a two-stage disentanglement. First, dynamic scene disentanglement constructs spatial appearance features via SLIC superpixels and temporal motion features via inner-product correlation volumes of the event stream, then clusters these features with K-Means to produce pseudo-labels; a ClusterFormer trained with a clustering loss $L_{clu}$ recursively updates cluster centers so that background and object features separate in feature space. Second, Gaussian representation fusion uses an extended Kalman filter on event correlation volumes to produce continuous tracking points, places 4D Gaussians on those points, and constrains them with a consistency loss $L_{consis}$ that matches Gaussian color to event brightness and Gaussian deformation to event optical flow. Background 3D Gaussians and object 4D Gaussians are combined by shadow-based weights and rendered with alpha-blending. The mechanism works because it converts the frame-event modality gap into explicit pseudo-labels, then uses event spatiotemporal structure as a training signal for the object's Gaussian parameters.

What would settle it

Capture or render a high-dynamic scene in which the moving object shares the background's color distribution and generates event flow similar to background edges, then check whether appearance-motion clustering still separates its superpixels into the object cluster; if the object Gaussians are not built, the novel view will show the object missing or blurred, and the claimed disentanglement would be falsified.

Watch

Extended reading notes

Core claim

The central claim is that a spatiotemporal-disentangled Gaussian splatting pipeline, called STD-GS, can reconstruct a time-continuous high-dynamic scene by modeling static background and dynamic objects in separate latent Gaussian spaces instead of one unified representation. Scene-level disentanglement uses a learnable appearance-motion clustering network: superpixel patches from frames are the spatial appearance cue, event correlation volumes are the temporal motion cue, K-Means supplies pseudo-labels, and a ClusterFormer refines the separation. Object-level disentanglement represents the object as 4D Gaussians centered on tracking points updated by an extended Kalman filter over event data, and ties Gaussian color and deformation to event brightness and flow through consistency losses. The two families of Gaussians are fused with shadow-aware weights and rendered by alpha-blending. The paper reports that this decomposition-fusion design outperforms unified and event-enhanced baselines on synthetic and real high-dynamic datasets, such as 27.12 dB PSNR versus 20.60 dB on HD-CED.

Load-bearing premise

The pipeline assumes that K-Means clustering of superpixel appearance and event correlation volumes produces reliable background-object pseudo-labels that a ClusterFormer can learn; if the clustering is wrong, the 3D/4D Gaussian decomposition is built on incorrect regions and the disentanglement benefit collapses.

Editorial extensions

If this is right

  • A single model need not carry both rigid and deformable motion: separating background and object Gaussians before fusion is enough to recover sharper novel views in high-dynamic scenes.
  • Event data can serve as pseudo-labels for Gaussian spatiotemporal structure, not just as a deblurring or interpolation aid, if the color-brightness and deformation-flow correspondences hold.
  • Time-continuous rendering of dynamic scenes is achievable by tracking objects in event streams and placing object Gaussians on the tracked points, replacing frame interpolation.
  • The joint reconstruction, clustering, and consistency objectives are trainable end-to-end, with the inference model reduced to a ClusterFormer and a Gaussian representation.
  • On the datasets used, the frame-plus-event input outperforms either modality alone, and the gap is larger in high-dynamic than in low-dynamic motion.

Reading between the lines

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

  • Editorial inference: the same appearance-motion clustering could be reused as a self-supervised segmentation step for other event-frame dynamic reconstruction and video editing pipelines, since it produces an explicit background-object mask rather than a latent-only split.
  • Editorial inference: the reported failure on radial z-axis motion suggests that a depth modality, or a stronger motion model, would be needed to extend the disentanglement to arbitrary 3D trajectories; the authors note this and propose LiDAR.
  • Editorial inference: if the clustering pseudo-labels are noisy, the consistency losses may reinforce the wrong object region; a testable extension would be to add uncertainty weighting or iterative label refinement during training.
  • Editorial inference: the method's benefit should transfer to deblurring and frame-interpolation benchmarks because the event-derived tracking directly addresses temporal discontinuity, though this is not demonstrated in the paper.
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 proposes STD-GS, a multimodal Gaussian splatting framework for high-dynamic scene reconstruction that combines frame images and event streams. The method first disentangles the scene into static background and dynamic objects using a learnable appearance-motion clustering pipeline (Section 3.2), then represents the background as 3D Gaussians and dynamic objects as 4D Gaussians with EKF-based temporal tracking, and finally fuses the two sets through shadow-based weights (Section 3.3). The framework is trained with reconstruction, clustering, and spatiotemporal-consistency losses (Eqs. 1, 2, 5, 6). The authors evaluate on Event-HyperNeRF, HD-CED, and a newly captured CoFED dataset, reporting large gains over existing baselines, e.g., PSNR 27.12 versus 20.60 on HD-CED (Table 1).

Significance. If validated, the paper would make a useful contribution by showing that disentangling static and dynamic components in Gaussian splatting, with event data providing temporal cues, can improve reconstruction of high-dynamic scenes. The proposed CoFED dataset is a potentially valuable resource, and the ablations in Tables 3 to 6 are internally consistent with the contribution of each proposed loss and component. However, the central mechanism of the method—the reliability of the unsupervised background/object pseudo-labels—is not quantitatively verified, and parts of the reported superiority may be attributable to uncontrolled factors such as event-based deblurring. The paper also presents the event-Gaussian consistency as a discovered characteristic when it is partly enforced by the training loss. These issues leave the core claim supported only conditionally, pending additional validation experiments.

major comments (4)
  1. [Section 3.2, Eq. (1), Fig. 7] The pseudo-labeling step that drives the whole scene disentanglement is not validated against any ground-truth segmentation. K-Means clustering of superpixel appearance features and event correlation volumes is used to generate labels for training ClusterFormer via Eq. (1), but no IoU, accuracy, or other quantitative metric is reported against actual background/object regions. The t-SNE visualization in Fig. 7 is circular evidence because K-Means always produces separated clusters. Since the downstream 3D/4D Gaussian assignment, EKF tracking, and fusion in Eqs. (3)-(4) all depend on these masks, the manuscript must either report pseudo-label accuracy on datasets where ground-truth masks are available (e.g., synthetic Event-HyperNeRF), or ablate with ground-truth masks to show that the gains in Table 3 are not an artifact of incorrect segmentation. The number of clusters K is also never specified, and it is unclear how the fixed K=2 assumption handles scenes with multiple dynamic objects.
  2. [Section 3.2, sentence on manual candidate regions] The sentence 'we manually divide the regions of background and objects as candidate regions to match' is ambiguous and load-bearing for the claim that the disentanglement is automatic. If manually drawn region masks are used to extract or match the spatial and temporal features that feed the K-Means clustering, then the method is not fully automatic, and the comparison with fully automatic baselines in Tables 1 and 2 is not fair. If the manual division is used only for illustrative visualization, that should be stated explicitly and the entire pipeline should be shown to work without any manual input. As written, the reader cannot determine whether the reported gains come from the proposed appearance-motion clustering or from human annotations.
  3. [Section 3.3, Eq. (2), Fig. 4] The claim that 'we discover that Gaussian representations and event data share the consistent spatiotemporal characteristic' is not supported independently of the proposed loss. Eq. (2) explicitly enforces consistency between Gaussian color/deformation and event brightness/flow by minimizing the l1 difference, so any observed consistency after training is a consequence of this constraint, not a discovered prior. Fig. 4 illustrates the consistency but is presumably produced after or with the same loss. To support the conceptual claim in the abstract, the authors should provide evidence computed before optimization or on a held-out set without the consistency loss, or alternatively reframe the statement as a designed inductive bias rather than an empirical discovery.
  4. [Tables 1-2, Section 3.2] The comparison with frame-based baselines is not controlled for the effect of event-based deblurring. The proposed method first uses event data to restore potential blurry textures (Section 3.2, 'Referring to GEM'), whereas the 'w/ Deblur' baselines use a generic DeblurGAN without event information. The large margins in Tables 1 and 2 may therefore partly reflect the advantage of event-guided deblurring rather than the proposed spatiotemporal disentanglement. The authors should include a comparison with an event-based deblurring baseline, or ablate the event-deblur step within their own framework, to isolate the contribution of the disentanglement and fusion modules.
minor comments (5)
  1. [Table 3-6] All ablation studies report single-run metrics without error bars or significance tests; reporting mean and standard deviation over multiple seeds would strengthen the claims, especially given the large reported margins.
  2. [Section 3.4 and Eq. (1)] Several implementation details are missing: the number of K-Means clusters K, the architecture and hidden size of the shadow-overlap MLP in Eq. (3), the EKF noise parameters, and the exact form of the superpixel and correlation-volume features. These omissions hinder reproducibility.
  3. [Section 4.1, Event-HyperNeRF and HD-CED] The use of blur synthesis and frame extraction to simulate high-dynamic scenes from lower-dynamic data should be discussed more carefully, since the realism of this synthetic degradation directly affects the external validity of the evaluations.
  4. [References] Reference [32] contains a typo ('Clusterfomer' should be 'ClusterFormer'), and the repeated spelling 'TiNeuV ox' in Table 1 and the caption should be corrected.
  5. [Section 4.4, Limitation] The stated limitation about radial motion along the z-axis is honest, but it should be reflected in the abstract and conclusion, which currently claim general 'high-dynamic scene reconstruction' without noting this failure mode.

Circularity Check

2 steps flagged · score 2.0 of 10

External benchmarks keep the central reconstruction claim independent; two supporting 'discovery' statements are self-confirming but not load-bearing.

  1. self definitional [Abstract; Section 3.3, Eq. (2); Fig. 4]
    "As for dynamic object, we discover that Gaussian representations and event data share the consistent spatiotemporal characteristic, which could serve as a prior to guide the spatiotemporal disentanglement of object Gaussians. ... Motivated by this, we treat the spatiotemporal characteristics of event data as pseudo-labels, and construct a spatiotemporal consistency loss between event data and Gaussian feature representation to constrain the learning of the spatiotemporal characteristics of dynamic objects: Lconsis = sum ||Cd − Lev||1 + ||F (Gd) − Fev||1"

    The claimed 'discovery' of a consistent spatiotemporal characteristic between Gaussians and events is not an independent observation: it is enforced by Lconsis in Eq. (2), which optimizes Gaussian color Cd and deformation F(Gd) to match event-derived Lev and Fev. Fig. 4 then shows agreement that the training objective was explicitly constructed to produce, so the consistency prior is self-confirming rather than independently validated. The final novel-view metrics remain external, so this step does not make the reconstruction claim itself circular.

  2. fitted input called prediction [Section 3.2, Eq. (1); Section 4.4, Fig. 7]
    "we apply K-Means algorithm to cluster these spatiotemporal features corresponding to static background and dynamic objects, as pseudo-labels to guide subsequent network learning. ... After introducing appearance-motion clustering, the spatiotemporal features of background and objects are pushed away. This demonstrates that appearance-motion clustering effectively disentangles the spatiotemporal features of static background and dynamic objects."

    ClusterFormer is trained with Lclu in Eq. (1) using K-Means labels yi computed from the same frame/event features the network is meant to learn. The post-training t-SNE separation shown in Fig. 7 is therefore a consequence of the fitting objective, and K-Means will separate any feature set into k clusters. The visualization cannot independently establish that the clusters correspond to background versus object. This is supporting evidence only, and the rendering evaluation is independent, so the circularity is partial.

full rationale

The paper's quantitative core is independent of its motivating 'discoveries': the reported gains (e.g., PSNR 27.12 vs 20.60 on HD-CED) are measured on novel-view synthesis against ground-truth frames, with competing methods trained under comparable input conditions, so the central reconstruction claim is not reduced to a fitted input. The circularity concerns are confined to two supporting statements. First, the claimed consistent spatiotemporal characteristic between Gaussians and events is not an independent observation: it is enforced by Lconsis in Eq. (2), which optimizes Gaussian color and deformation to match event brightness and flow, so comparing the trained Gaussians to the same event quantities in Fig. 4 is self-confirming. Second, the demonstration that appearance-motion clustering effectively disentangles background and objects uses K-Means pseudo-labels as training targets in Eq. (1) and then displays t-SNE separation, which K-Means would produce for any k-partition of the input features. Neither step invalidates the externally benchmarked reconstruction numbers, so a low partial score is warranted.

Assumptions & free parameters 4 free parameters · 3 assumptions · 1 invented entities

The central claim depends on several hand-chosen hyperparameters and on domain assumptions about scene structure and event-Gaussian consistency. Most free parameters and dataset details are not published, so the contribution cannot be independently reconstructed from the preprint.

free parameters (4)
  • Loss weights lambda1, lambda2, lambda3 = not reported
    Eq. 6 controls the relative importance of reconstruction, clustering, and consistency losses; values are chosen by hand and not stated.
  • Number of K-Means clusters K = not reported
    K defines the background/object decomposition in Section 3.2; no value or sensitivity analysis is given.
  • SLIC superpixel size = not reported
    Superpixel patches are the spatial appearance features in Section 3.2; the scale is a free choice that affects clustering.
  • EKF noise parameters = not reported
    Extended Kalman filter tracking in Section 3.3 requires process and measurement noise settings that are not specified.
assumptions (3)
  • domain assumption High-dynamic scenes decompose into exactly one rigid static background and deformable foreground objects.
    Used throughout Sections 1 and 3.2 to justify a two-way disentanglement; real scenes may contain multiple moving objects or non-rigid background elements.
  • domain assumption Event brightness and flow are consistent with Gaussian color and deformation for the same object.
    Section 3.3, Eq. 2 enforces this consistency as a loss, so it is assumed rather than independently established.
  • domain assumption EKF with a single correlation-volume template can track dynamic objects across time.
    Section 3.3 relies on EKF tracking to produce continuous points; severe deformation or radial motion violates this, as acknowledged in the Limitation section.
invented entities (1)
  • Shadow-based overlap probability rho (MLP)
    purpose: Weights the fusion of background and object Gaussians in Eq. 3 and Eq. 4.
    The MLP is introduced to model occlusion or shadow between background and objects; no external validation of this quantity is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of STD-GS: Exploring Frame-Event Interaction for SpatioTemporal-Disentangled Gaussian Splatting to Reconstruct High-Dynamic Scene." pith.science (2026). https://pith.science/paper/DKBY77IA

@misc{pith2026250623157,
  author       = {Pith},
  title        = {Pith review of: STD-GS: Exploring Frame-Event Interaction for SpatioTemporal-Disentangled Gaussian Splatting to Reconstruct High-Dynamic Scene},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DKBY77IA}},
  note         = {Machine review of arXiv:2506.23157}
}
read the original abstract

High-dynamic scene reconstruction aims to represent static background with rigid spatial features and dynamic objects with deformed continuous spatiotemporal features. Typically, existing methods adopt unified representation model (e.g., Gaussian) to directly match the spatiotemporal features of dynamic scene from frame camera. However, this unified paradigm fails in the potential discontinuous temporal features of objects due to frame imaging and the heterogeneous spatial features between background and objects. To address this issue, we disentangle the spatiotemporal features into various latent representations to alleviate the spatiotemporal mismatching between background and objects. In this work, we introduce event camera to compensate for frame camera, and propose a spatiotemporal-disentangled Gaussian splatting framework for high-dynamic scene reconstruction. As for dynamic scene, we figure out that background and objects have appearance discrepancy in frame-based spatial features and motion discrepancy in event-based temporal features, which motivates us to distinguish the spatiotemporal features between background and objects via clustering. As for dynamic object, we discover that Gaussian representations and event data share the consistent spatiotemporal characteristic, which could serve as a prior to guide the spatiotemporal disentanglement of object Gaussians. Within Gaussian splatting framework, the cumulative scene-object disentanglement can improve the spatiotemporal discrimination between background and objects to render the time-continuous dynamic scene. Extensive experiments have been performed to verify the superiority of the proposed method.

Figures

Figures reproduced from arXiv: 2506.23157 by the authors.

Figure 1
Figure 1. Illustration of Gaussian splatting paradigms for scene [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The architecture of the STD-GS mainly contains dynamic scene disentanglement (DSD) and Gaussian representation fusion (GRF). [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Spatiotemporal discrepancy between the background and [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Spatiotemporal consistency between Gaussians and [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Visual comparison of novel view synthesis on Event-HyperNeRF and HD-CED datasets. “F” denotes frame and “E” denotes event. [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Visual comparison of novel view synthesis on the proposed CoFED dataset. “F” denotes frame and “E” denotes event. [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: t-SNE visualization of spatiotemporal features. Frame [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

42 extracted references · 36 canonical work pages

  1. [1]

    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. Communications of the ACM, 65(1):99–106, 2021. 1, 2

    2021

  2. [2]

    3d gaussian splatting for real-time radiance field rendering

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimkühler, and George Drettakis. 3d gaussian splatting for real-time radiance field rendering. ACM Trans. Graph., 42(4):139–1, 2023. 1, 2, 5, 6

    work page 2023

  3. [3]

    Fast dynamic radiance fields with time-aware neural vox- els

    Jiemin Fang, Taoran Yi, Xinggang Wang, Lingxi Xie, Xi- aopeng Zhang, Wenyu Liu, Matthias Nießner, and Qi Tian. Fast dynamic radiance fields with time-aware neural vox- els. In SIGGRAPH Asia 2022 Conference Papers, pages 1–9,

    2022

  4. [4]

    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. In Int. Conf. Learn. Represent., 2024

    work page 2024

  5. [5]

    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 IEEE Conf. Comput. Vis. Pattern Recog., pages 21136–21145,

    work page

  6. [6]

    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 IEEE Conf. Comput. Vis. Pattern Recog. , pages 20310– 20320, 2024. 3, 6

    work page 2024

  7. [7]

    Shape of motion: 4d reconstruc- tion from a single video

    Qianqian Wang, Vickie Ye, Hang Gao, Jake Austin, Zhengqi Li, and Angjoo Kanazawa. Shape of motion: 4d reconstruc- tion from a single video. arXiv preprint arXiv:2407.13764,

    arXiv

  8. [8]

    Orchard, Chiara Bar- tolozzi, Brian Taba, Andrea Censi, Stefan Leutenegger, An- drew J

    Guillermo Gallego, Tobi Delbrück, G. Orchard, Chiara Bar- tolozzi, Brian Taba, Andrea Censi, Stefan Leutenegger, An- drew J. Davison, Jörg Conradt, Kostas Daniilidis, and Davide Scaramuzza. Event-based vision: A survey. IEEE Trans. Pattern Anal. Mach. Intell., 44:154–180, 2019. 2

    work page 2019

Show all 42 references
  1. [9]

    Slic superpixels compared to state-of-the-art superpixel methods

    Radhakrishna Achanta, Appu Shaji, Kevin Smith, Aurelien Lucchi, Pascal Fua, and Sabine Süsstrunk. Slic superpixels compared to state-of-the-art superpixel methods. 34(11):2274– 2282, 2012. 2

    2012

  2. [10]

    Robust extended kalman filtering

    Garry A Einicke and Langford B White. Robust extended kalman filtering. IEEE transactions on signal processing , 47(9):2596–2599, 1999. 2, 5

    1999

  3. [11]

    Photo tourism: exploring photo collections in 3d

    Noah Snavely, Steven M Seitz, and Richard Szeliski. Photo tourism: exploring photo collections in 3d. In ACM siggraph 2006 papers, pages 835–846. 2006. 2

    2006

  4. [12]

    Orb-slam: a versatile and accurate monocular slam system

    Raul Mur-Artal, Jose Maria Martinez Montiel, and Juan D Tardos. Orb-slam: a versatile and accurate monocular slam system. IEEE transactions on robotics , 31(5):1147–1163,

  5. [13]

    Neural point-based graphics

    Kara-Ali Aliev, Artem Sevastopolsky, Maria Kolos, Dmitry Ulyanov, and Victor Lempitsky. Neural point-based graphics. In Eur. Conf. Comput. Vis., pages 696–712. Springer, 2020. 2

    2020

  6. [14]

    Point-based neural rendering with per- view optimization

    Georgios Kopanas, Julien Philip, Thomas Leimkühler, and George Drettakis. Point-based neural rendering with per- view optimization. In Computer Graphics Forum, volume 40, pages 29–43. Wiley Online Library, 2021

    2021

  7. [15]

    Adop: Approximate differentiable one-pixel point rendering

    Darius Rückert, Linus Franke, and Marc Stamminger. Adop: Approximate differentiable one-pixel point rendering. ACM Trans. Graph., 41(4):1–14, 2022. 2

    2022

  8. [16]

    The re- lightables: V olumetric performance capture of humans with realistic relighting

    Kaiwen Guo, Peter Lincoln, Philip Davidson, Jay Busch, Xueming Yu, Matt Whalen, Geoff Harvey, Sergio Orts- Escolano, Rohit Pandey, Jason Dourgarian, et al. The re- lightables: V olumetric performance capture of humans with realistic relighting. ACM Trans. Graph., 38(6):1–19, 2019. 2

    2019

  9. [17]

    Robust 3d human motion reconstruction via dynamic template con- struction

    Zhong Li, Yu Ji, Wei Yang, Jinwei Ye, and Jingyi Yu. Robust 3d human motion reconstruction via dynamic template con- struction. In International Conference on 3D Vision, pages 496–505. IEEE, 2017. 2

    2017

  10. [18]

    Immersive light field video with a layered mesh representation

    Michael Broxton, John Flynn, Ryan Overbeck, Daniel Er- ickson, Peter Hedman, Matthew Duvall, Jason Dourgarian, Jay Busch, Matt Whalen, and Paul Debevec. Immersive light field video with a layered mesh representation. ACM Trans. Graph., 39(4):86–1, 2020. 2

    2020

  11. [19]

    High-quality streamable free-viewpoint video

    Alvaro Collet, Ming Chuang, Pat Sweeney, Don Gillett, Den- nis Evseev, David Calabrese, Hugues Hoppe, Adam Kirk, and Steve Sullivan. High-quality streamable free-viewpoint video. ACM Trans. Graph., 34(4):1–13, 2015

    2015

  12. [20]

    Robust non-rigid motion tracking and surface recon- struction using l0 regularization

    Kaiwen Guo, Feng Xu, Yangang Wang, Yebin Liu, and Qiong- hai Dai. Robust non-rigid motion tracking and surface recon- struction using l0 regularization. In Int. Conf. Comput. Vis., pages 3083–3091, 2015

    2015

  13. [21]

    Robustfusion: Human volumetric capture with data- driven visual cues using a rgbd camera

    Zhuo Su, Lan Xu, Zerong Zheng, Tao Yu, Yebin Liu, and Lu Fang. Robustfusion: Human volumetric capture with data- driven visual cues using a rgbd camera. In Eur. Conf. Comput. Vis., pages 246–264. Springer, 2020. 2

    2020

  14. [22]

    Hypernerf: A higher- dimensional representation for topologically varying neural radiance fields

    Keunhong Park, Utkarsh Sinha, Peter Hedman, Jonathan T Barron, Sofien Bouaziz, Dan B Goldman, Ricardo Martin- Brualla, and Steven M Seitz. Hypernerf: A higher- dimensional representation for topologically varying neural radiance fields. arXiv preprint arXiv:2106.13228, 2021. 2, 6

    2021 arXiv

  15. [23]

    Emernerf: Emergent spatial-temporal scene decomposition via self-supervision

    Jiawei Yang, Boris Ivanovic, Or Litany, Xinshuo Weng, Se- ung Wook Kim, Boyi Li, Tong Che, Danfei Xu, Sanja Fidler, Marco Pavone, et al. Emernerf: Emergent spatial-temporal scene decomposition via self-supervision. In Int. Conf. Learn. Represent., 2024. 3

    2024

  16. [24]

    Spacetime gaus- sian feature splatting for real-time dynamic view synthesis

    Zhan Li, Zhang Chen, Zhong Li, and Yi Xu. Spacetime gaus- sian feature splatting for real-time dynamic view synthesis. In 9 IEEE Conf. Comput. Vis. Pattern Recog., pages 8508–8520,

  17. [25]

    Motiongs: Exploring explicit motion guidance for deformable 3d gaussian splatting

    Ruijie Zhu, Yanzhe Liang, Hanzhi Chang, Jiacheng Deng, Ji- ahao Lu, Wenfei Yang, Tianzhu Zhang, and Yongdong Zhang. Motiongs: Exploring explicit motion guidance for deformable 3d gaussian splatting. arXiv preprint arXiv:2410.07707 , 2024

    arXiv

  18. [26]

    Eventnerf: Neural radiance fields from a single colour event camera

    Viktor Rudnev, Mohamed Elgharib, Christian Theobalt, and Vladislav Golyanik. Eventnerf: Neural radiance fields from a single colour event camera. In IEEE Conf. Comput. Vis. Pattern Recog., pages 4992–5002, 2023. 6

    2023

  19. [27]

    E2nerf: Event enhanced neural radiance fields from blurry images

    Yunshan Qi, Lin Zhu, Yu Zhang, and Jia Li. E2nerf: Event enhanced neural radiance fields from blurry images. In Int. Conf. Comput. Vis., pages 13254–13264, 2023. 6

    2023

  20. [28]

    Deformable neural radiance fields using rgb and event cameras

    Qi Ma, Danda Pani Paudel, Ajad Chhatkuli, and Luc Van Gool. Deformable neural radiance fields using rgb and event cameras. In Int. Conf. Comput. Vis., pages 3590–3600,

  21. [29]

    Event3dgs: Event-based 3d gaussian splatting for high-speed robot egomotion

    Tianyi Xiong, Jiayi Wu, Botao He, Cornelia Fermuller, Yiannis Aloimonos, Heng Huang, and Christopher Metzler. Event3dgs: Event-based 3d gaussian splatting for high-speed robot egomotion. In 8th Annual Conference on Robot Learn- ing, 2024. 3

    2024

  22. [30]

    E2gs: Event enhanced gaussian splatting

    Hiroyuki Deguchi, Mana Masuda, Takuya Nakabayashi, and Hideo Saito. E2gs: Event enhanced gaussian splatting. In 2024 IEEE International Conference on Image Processing (ICIP), pages 1676–1682. IEEE, 2024. 3, 6

    2024

  23. [31]

    Generalizing event-based motion deblurring in real-world scenarios

    Xiang Zhang, Lei Yu, Wen Yang, Jianzhuang Liu, and Gui- Song Xia. Generalizing event-based motion deblurring in real-world scenarios. In Int. Conf. Comput. Vis., pages 10734– 10744, 2023. 4

    2023

  24. [32]

    Clusterfomer: clustering as a universal visual learner

    James Liang, Yiming Cui, Qifan Wang, Tong Geng, Wenguan Wang, and Dongfang Liu. Clusterfomer: clustering as a universal visual learner. Adv. Neural Inform. Process. Syst., 36, 2024. 4

    2024

  25. [33]

    Differentiable surface splatting for point-based geometry processing

    Wang Yifan, Felice Serena, Shihao Wu, Cengiz Öztireli, and Olga Sorkine-Hornung. Differentiable surface splatting for point-based geometry processing. ACM Trans. Graph., 38(6):1–14, 2019. 5

    2019

  26. [34]

    Pulsar: Efficient sphere-based neural rendering

    Christoph Lassner and Michael Zollhofer. Pulsar: Efficient sphere-based neural rendering. In IEEE Conf. Comput. Vis. Pattern Recog., pages 1440–1449, 2021. 5

    2021

  27. [35]

    Pytorch: An imperative style, high-performance deep learning library

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. Pytorch: An imperative style, high-performance deep learning library. Adv. Neural Inform. Process. Syst., 32, 2019. 5

    2019

  28. [36]

    v2e: From video frames to realistic dvs events

    Yuhuang Hu, Shih-Chii Liu, and Tobi Delbruck. v2e: From video frames to realistic dvs events. In IEEE Conf. Comput. Vis. Pattern Recog. Worksh., pages 1312–1321, 2021. 6

    2021

  29. [37]

    Ced: Color event camera dataset

    Cedric Scheerlinck, Henri Rebecq, Timo Stoffregen, Nick Barnes, Robert Mahony, and Davide Scaramuzza. Ced: Color event camera dataset. In IEEE Conf. Comput. Vis. Pattern Recog. Worksh., pages 0–0, 2019. 6

    2019

  30. [38]

    Deblurgan: Blind motion deblurring using conditional adversarial networks

    Orest Kupyn, V olodymyr Budzan, Mykola Mykhailych, Dmytro Mishkin, and Jiˇrí Matas. Deblurgan: Blind motion deblurring using conditional adversarial networks. In IEEE Conf. Comput. Vis. Pattern Recog., pages 8183–8192, 2018. 7

    2018

  31. [39]

    Image quality metrics: Psnr vs

    Alain Hore and Djemel Ziou. Image quality metrics: Psnr vs. ssim. In 2010 20th international conference on pattern recognition, pages 2366–2369. IEEE, 2010. 7

    2010

  32. [40]

    Image quality assessment: from error visibility to structural similarity

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Simoncelli. Image quality assessment: from error visibility to structural similarity. IEEE Trans. Image Process., 13(4):600– 612, 2004. 7

    2004

  33. [41]

    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. In IEEE Conf. Comput. Vis. Pattern Recog., pages 586–595, 2018. 7

    2018

  34. [42]

    Time lens: Event-based video frame interpolation

    Stepan Tulyakov, Daniel Gehrig, Stamatios Georgoulis, Julius Erbach, Mathias Gehrig, Yuanyou Li, and Davide Scaramuzza. Time lens: Event-based video frame interpolation. In IEEE Conf. Comput. Vis. Pattern Recog., pages 16155–16164, 2021. 8 10

    2021

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.