Pith. sign in

REVIEW 4 major objections 6 minor 65 references

TwoSquared: 4D Generation from 2D Image Pairs

T0 review · 4 major / 6 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 A full 4D motion sequence can be generated from just two 2D photographs: one of the object at the start of an action and one at the end.

desk verdict A genuinely new two-image-to-4D task with a clean, honest pipeline, but the quantitative evidence is thin and the central claim is conditional on the 3D backbone. read the letter →

arxiv 2504.12825 v1 pith:UTN4RNXQ submitted 2025-04-17 cs.CV

Lu Sang , Zehranaz Canfes , Dongliang Cao , Riccardo Marin , Florian Bernard , Daniel Cremers This is my paper · ORCID
classification cs.CV
keywords 4Dgenerationimagepairvelocityfieldshapedeformationfunctionalmapsdiffusionfeaturestextureconsistencyphysicallyplausiblemotion
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

TwoSquared claims that a complete 4D motion sequence — a continuously deforming 3D object — can be generated from just two ordinary 2D photographs of the same object, one at the start and one at the end of an action. The paper's method avoids generating intermediate images or relying on category templates: it lifts each input image to a 3D mesh with an existing image-to-3D model, finds correspondences between the two meshes, and learns a time-continuous velocity field that physically deforms the first mesh into the second. The authors report that the resulting sequences stay texture-consistent and geometry-consistent on real scanned human motions, and that the approach beats image-morphing baselines, the 4Deform baseline, and even direct reconstruction of ground-truth intermediate frames. A sympathetic reader would care because this turns 4D content creation into a two-image problem with no class-specific priors and arbitrary frame-rate output.

What carries the argument

The load-bearing mechanism is a time-continuous velocity field $V: \mathbb{R}^3 \times [0,T] \to \mathbb{R}^3$, implemented as an 8-layer MLP (the Velocity Net), which maps each point of the source mesh to its trajectory so that intermediate shapes are obtained by Euler integration. The field is trained against six losses: a kinetic-energy smoothing term, an overlapping Chamfer loss pulling the endpoint of the trajectory onto the target mesh, a normal-alignment loss using the deformation gradient, a stretching penalty, a confidence-weighted matching loss, and a distortion term. The correspondences that feed the matching loss come from per-vertex diffusion features (Diff3F) refined by a functional-map optimization and a bidirectional close-loop consistency check, which filters unreliable matches.

What would settle it

Take a pair of images for which the image-to-3D backbone produces one mesh with a missing limb or large hole, run TwoSquared, and check the intermediate meshes: if the artifact appears in every intermediate frame and Chamfer distance to ground-truth scans grows accordingly, the load-bearing premise — that two good keyframe meshes are required — is violated.

Watch

Extended reading notes

Core claim

The central discovery is that 4D generation from two RGB images is best approached by decomposition rather than end-to-end synthesis: first hallucinate the two endpoint geometries in 3D, then solve for a physically plausible path between them. On the 4D-DRESS real-scanned human dataset, TwoSquared reports lower Chamfer and Hausdorff distances than per-frame reconstruction from morphing baselines and than 4Deform, and it also outperforms lifting ground-truth intermediate images with the same backbone, which the paper attributes to the inconsistency of independent per-frame generation. The deformation is trained with losses for kinetic-energy smoothness, endpoint overlap, normal alignment, stretching, confidence-weighted matching, and distortion, so that the interpolated meshes preserve surface properties rather than merely blending shapes.

Load-bearing premise

The entire 4D sequence inherits the quality of the two keyframe meshes produced by the external image-to-3D model: if either mesh contains significant artifacts, the deformation stage will faithfully interpolate that error rather than repair it.

Editorial extensions

If this is right

  • Since the velocity field is continuous in time, the same trained model outputs 4D sequences at any frame rate without retraining; the paper recommends $T' < 2T$ for high frame rates.
  • The pipeline needs no templates or object-class priors, so it applies to humans, animals, sketches, and everyday photographs, and even to pairs of different subjects of the same species for pose transfer.
  • The closed-loop correspondence check and confidence weighting let the method keep intermediate shapes physically plausible even when registration fails completely in some regions, such as a leg with no matched correspondences.
  • The method is cheaper than morphing-based pipelines at inference: after about 17 minutes per pair, generating a mesh takes under one second, whereas baselines regenerate images and meshes per frame.
  • Deforming one consistent source mesh outperforms reconstructing each ground-truth intermediate image with the same image-to-3D backbone, indicating that per-frame lifting is the main source of texture and geometry inconsistency.

Reading between the lines

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

  • A testable prediction of the two-stage design is that its advantage over per-frame reconstruction grows with pose change and with the inconsistency of the image-to-3D backbone across unrelated viewpoints; this is not directly measured in the paper.
  • The velocity-field scaffold is, in principle, agnostic to the correspondence source and the 3D backbone; the paper states this flexibility but only demonstrates Hunyuan3D with Diff3F features.
  • The 'physically plausible' label comes from an optimization objective, not from a physics simulation; a stricter test would compare predicted trajectories against motion capture or optical-flow-derived motion rather than mesh distances.
  • Symmetry and featureless regions remain a likely failure zone, as the paper itself flags hair as an open problem where thousands of simultaneous small deformations exceed the tracking quality available.
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

4 major / 6 minor

Summary. The paper introduces TwoSquared, a method for generating a 4D (dynamic 3D) sequence from only two RGB images that depict the initial and final states of an object. The pipeline decomposes the problem into three components: (i) an image-to-3D generation block based on Hunyuan3D that produces keyframe meshes; (ii) a vertex registration block that extracts per-vertex features with Diff3F, refines them via functional-map optimization, and enforces bidirectional loop-closure consistency to obtain sparse correspondences with confidence weights; and (iii) a shape deformation block that learns a time-continuous velocity field parameterized by a small MLP, trained with a combination of smoothness, overlapping, normal, stretching, confidence-guided matching, and distortion losses. The authors report quantitative results on two sequences from 4D-DRESS, comparing against image-morphing baselines (DiffMorpher, DreamMover) and the 3D-to-4D method 4Deform, and qualitative results on web images, sketches, and daily photographs. The central claims are that the method is template-free, class-agnostic, physically plausible, and that it outperforms prior approaches and even reconstruction from ground-truth intermediate images on the tested sequences.

Significance. If the claims hold, TwoSquared offers a simple and practical formulation for 4D generation from minimal input, with the appealing property of arbitrary-frame-rate inference at near-real-time cost after a short per-pair optimization. The decomposition into a mature image-to-3D backbone and a physics-inspired deformation module is conceptually clean, and the authors provide a code link, use a purely JAX implementation, and include an interesting control baseline that reconstructs the shape from ground-truth intermediate images. The physical loss terms (overlapping, normal, stretching, confidence-guided matching) are reasonable in intent, and the qualitative results on in-the-wild images are visually compelling. However, the evidence currently supports a much narrower claim than the abstract and introduction suggest: the quantitative evaluation is limited to two human-motion sequences, the method's success is explicitly conditional on the quality of the external Hunyuan3D backbone, and several implementation-critical hyperparameters are undisclosed.

major comments (4)
  1. [§4.2, Table 1] The quantitative evaluation is restricted to two sequences (Take22 and Take7) from 4D-DRESS, both capturing human motions from the same dataset. No error bars, repeated runs, or statistical tests are reported, and the differences between methods in several rows are small (e.g., CD values around 0.074–0.126 in Take7). This is load-bearing for the claim that TwoSquared achieves 'lower error rates compared to other approaches' and 'even outperforms results generated directly from ground truth intermediate images.' Please either expand the evaluation to more sequences and object categories, report variance over multiple runs with different seeds, or explicitly scope the claim to the two tested sequences.
  2. [§3.4, Eq. (15); §4.1] The total loss in Eq. (15) depends on six scalar weights λv, λo, λn, λs, λm, λd, and the correspondence filter in Eq. (1) depends on a distance threshold δd. None of these values is reported in the main paper or the supplementary material. Since the method trains a small network per image pair, these hyperparameters are essential for reproducibility, and the ablation in Table 2 shows that the final metrics are sensitive to which loss terms are active. Please disclose all hyperparameter values and, ideally, report a sensitivity study.
  3. [§4.4, with §1 and §5] The abstract and introduction assert that TwoSquared generates 'physically plausible' sequences from arbitrary in-the-wild image pairs without object-class priors. However, §4.4 states that plausibility holds 'as long as the starting and ending meshes are generated without significant artifacts,' making the quality of Hunyuan3D a load-bearing premise. There is no quantitative measurement of how often Hunyuan3D produces artifact-free keyframes on the claimed in-the-wild categories, no experiment that replaces Hunyuan3D with another image-to-3D backbone, and the only failure case (Fig. 8) is qualitative. Consequently, the method's generality is verified only for 4D-DRESS human motions. Please either add a quantitative robustness study or explicitly reframe the claims as conditional on the backbone's output quality.
  4. [§3.3.2, Eq. (14)] The distortion loss as written appears dimensionally inconsistent. Both 1/6 Tr(D)² and 1/2 Tr(D·D)² are scalars, so the Frobenius norm is simply the absolute value of their difference; squaring the trace of D·D is unusual, and the expression is likely a typo for something like Tr(D)² - Tr(D·D) or a norm of the matrix D·D. Since this term is part of the total loss and is claimed to follow [38], please correct the formula or clarify the intended expression.
minor comments (6)
  1. [§3.3.1, Eq. (3)–(6)] The text refers to 'Kinect energy'; this should be 'kinetic energy.' Also, the constraint notation in Eq. (4), ∫Ω0 X(0) dx = P0, is confusing because Ω0 is described as a subset of the point domain; please clarify the measure used and the relation between Ω0 and the sampled point cloud.
  2. [§3.2, Eq. (2)] The confidence map Cij = (δd − D_j^i) / max_j(D_j^i) uses the index j both for the target shape and as the variable in the maximum, which is ambiguous. Please rewrite with explicit per-vertex notation, e.g., C(xi) = (δd − D(xi)) / max_{xk} D(xk) for correspondences that pass the loop-closure filter.
  3. [§4.2, Table 1] The text says 'we extract five keyframes from two sequences ... and generate four 4D deformation sequences,' but it is unclear how many intermediate frames are evaluated per sequence and what the temporal spacing is. Please specify the exact keyframe indices and the number of compared intermediate steps.
  4. [Supplementary, S0] The supplementary reports that Diff3F feature extraction takes 5 minutes per frame and the velocity net trains in under 2 minutes, but the total 'roughly 17 minutes to train' does not include the 3 minutes per Hunyuan3D mesh generation; please make the pipeline timing consistent and state the total wall-clock time for one pair.
  5. [§4.1, Dataset] The qualitative evaluation on web images is presented only as images, with no quantitative or user study. Since the paper claims robustness to 'varying input quality,' a simple user study or at least a larger set of diverse inputs would strengthen the claim.
  6. [§4.2, Baselines] The comparison against 4Deform uses the same potentially noisy correspondences as TwoSquared, which is a good ablation of the deformation module, but it is not a comparison of the full methods in 4Deform's intended setting (high-quality correspondences). Please state this explicitly to avoid overgeneralizing the comparison.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the intermediate 4D shapes are produced from an external image-to-3D backbone and optimized velocity fields, then evaluated against external ground-truth meshes, so the derivation does not return its own inputs by construction.

full rationale

The claimed prediction—intermediate 4D meshes from two 2D images—is not equivalent to its inputs by construction. The endpoint meshes come from Hunyuan3D, an external system not controlled by the paper; the correspondences come from Diff3F features and functional-map refinement; and the intermediate shapes come from optimizing a velocity field with the stated losses in Sec. 3.3. The quantitative evaluation in Tab. 1 compares against ground-truth intermediate meshes from 4D-DRESS, and the losses are not designed to reproduce those ground-truth meshes; the overlapping loss only forces the deformed point cloud toward the generated endpoint mesh, not toward the GT intermediate shapes. The paper's strongest self-citations are the velocity-field smoothness term from [37] and the stretching/distortion losses from [38]; these are published prior works by overlapping authors, used transparently as building blocks, and the paper additionally runs 4Deform as a baseline on the same point clouds, so the deformation module is not justified by an unverified uniqueness claim. The admitted dependence on Hunyuan3D quality in Sec. 4.4 is a correctness/generality limitation, not a circularity: it says the output is conditional on the external backbone, not that the output was derived from itself. No fitted parameter is renamed as a prediction, no equation reduces to its own input, and no central claim is forced by a self-citation chain.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claim rests on the accuracy of external components (Hunyuan3D, Diff3F, functional maps) and on the assumption that regularized smooth velocity fields are an adequate model of physical plausibility. The loss weights and correspondence threshold are undisclosed free parameters that are likely tuned on the evaluation data. No new physical entities are introduced.

free parameters (2)
  • Loss weights λv, λo, λn, λs, λm, λd
    The relative weights of the six loss terms in Eq. (15) are not reported. They are presumably chosen by hand or by tuning on the evaluation sequences, which is not disclosed.
  • Correspondence distance threshold δd
    Used in Eq. (1) to filter mismatched vertices; the numerical value is not given.
assumptions (4)
  • domain assumption Hunyuan3D produces sufficiently accurate and consistent meshes from input images
    The entire pipeline depends on the quality of the 3D generation backbone (Sec. 3.1). The authors note in Sec. 4.4 that significant artifacts in the keyframe meshes degrade results.
  • domain assumption Diff3F features are semantically consistent across the two generated meshes
    Registration in Sec. 3.2 relies on cosine similarity of per-vertex diffusion features; if the features are not consistent, correspondences are wrong.
  • domain assumption A smooth regularity-constrained velocity field yields physically plausible deformations
    The physical plausibility of the trajectory is imposed via losses (smoothness, normal, stretching, distortion) rather than derived from mechanics or validated by external physical simulation.
  • domain assumption Surface area standard deviation is an adequate proxy for physical plausibility
    The only quantitative physical-plausibility metric reported is SAσ from 4Deform.

how reviews work

0 comments
Cite this review

Pith. "Pith review of TwoSquared: 4D Generation from 2D Image Pairs." pith.science (2026). https://pith.science/paper/UTN4RNXQ

@misc{pith2026250412825,
  author       = {Pith},
  title        = {Pith review of: TwoSquared: 4D Generation from 2D Image Pairs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UTN4RNXQ}},
  note         = {Machine review of arXiv:2504.12825}
}
read the original abstract

Despite the astonishing progress in generative AI, 4D dynamic object generation remains an open challenge. With limited high-quality training data and heavy computing requirements, the combination of hallucinating unseen geometry together with unseen movement poses great challenges to generative models. In this work, we propose TwoSquared as a method to obtain a 4D physically plausible sequence starting from only two 2D RGB images corresponding to the beginning and end of the action. Instead of directly solving the 4D generation problem, TwoSquared decomposes the problem into two steps: 1) an image-to-3D module generation based on the existing generative model trained on high-quality 3D assets, and 2) a physically inspired deformation module to predict intermediate movements. To this end, our method does not require templates or object-class-specific prior knowledge and can take in-the-wild images as input. In our experiments, we demonstrate that TwoSquared is capable of producing texture-consistent and geometry-consistent 4D sequences only given 2D images.

Figures

Figures reproduced from arXiv: 2504.12825 by the authors.

Figure 1
Figure 1. TwoSquared takes a pair of 2D images representing the initial and final states of an object as input and generates texture [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Pipeline of TwoSquared: TwoSquared processes two input images through an image-to-3D generation block, producing two 3D meshes. We then extract per-vertex features and compute a cosine similarity map, which is refined using a functional map module and a close loop check module to obtain point-to-point correspondences and a confidence map. These registered points are then fed into our shape deformation module, where … view at source ↗
Figure 3
Figure 3. Comparison with other methods: TwoSquared gen￾erates texture-consistent, physically plausible 4D sequences, and it is more robust then 4Deform [38] to correspondences noise. In contrast, other methods show artifacts in the intermediate shapes. we perform a quantitative ablation for the proposed losses on 4D-DRESS. Tab. 2 shows the error metrics on different settings. The overlapping loss Lo enforces global alignment… view at source ↗
Figures from the paper (4 more)
Figure 5
Figure 5. Figure 5: 4D reconstruction from web images. Our method can take a pair of images as input and generate the temporal consistent 4D deformation between these two objects. 4D generation from different types of images. We demonstrate that our method accommodates a wide range of inp…
Figure 4
Figure 4. Figure 4: Visualization of ablations: We show the qualitative results of our ablation study. while the visual results coincide with the error number reported in Tab. 2, the stretching loss and over￾lapping loss help the shape to remain physically plausible. While the normal loss…
Figure 6
Figure 6. Figure 6: Sketch images example: Our method can take simple, low-quality images such as sketch images as input and generate high-quality meshes. I0 I1 DiffMorpher [59] + Hunyuan3D [46] DreamMover [40] + Hunyuan3D [46] Ours front Ours side Ours back ◦ ◦ [PITH_FULL_IMAGE:figures/…
Figure 7
Figure 7. Figure 7: Daily images example: Our method can directly take images that are taken in daily life and generate continuous 4D se￾quences for different objects, animals, and humans. In the figure, we demonstrate the different angles of the 4D mesh sequences to show that our method …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

65 extracted references · 30 canonical work pages

  1. [38]

    4deform: Neural surface deformation for robust shape interpolation

    L Sang, Z Canfes, D Cao, R Marin, F Bernard, and D Cre- mers. 4deform: Neural surface deformation for robust shape interpolation. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2025. 3, 5, 6, 1

    work page 2025

  2. [1]

    MIT Press, 2019

    Virtual Reality. MIT Press, 2019. ISBN: 9780262354684. 1

    work page 2019

  3. [2]

    Videophy: Evaluating physical commonsense for video generation

    Hritik Bansal, Zongyu Lin, Tianyi Xie, Zeshun Zong, Michal Yarom, Yonatan Bitton, Chenfanfu Jiang, Yizhou Sun, Kai- Wei Chang, and Aditya Grover. Videophy: Evaluating physical commonsense for video generation. arXiv preprint arXiv:2406.03520, 2024. 2

    arXiv 2024

  4. [3]

    JAX: composable transformations of Python+NumPy programs, 2018

    James Bradbury, Roy Frostig, Peter Hawkins, Matthew James Johnson, Chris Leary, Dougal Maclau- rin, George Necula, Adam Paszke, Jake VanderPlas, Skye Wanderman-Milne, and Qiao Zhang. JAX: composable transformations of Python+NumPy programs, 2018. 5

    work page 2018

  5. [4]

    Optimal transportation of particles, fluids and currents

    Yann Brenier. Optimal transportation of particles, fluids and currents. In Variational methods for evolving objects, pages 59–86. Mathematical Society of Japan, 2015. 4

    work page 2015

  6. [5]

    Motion2vecsets: 4d latent vector set diffusion for non-rigid shape reconstruction and tracking, 2024

    Wei Cao, Chang Luo, Biao Zhang, Matthias Nießner, and Jiapeng Tang. Motion2vecsets: 4d latent vector set diffusion for non-rigid shape reconstruction and tracking, 2024. 3

    work page 2024

  7. [6]

    A physical coherence benchmark for evalu- ating video generation models via optical flow-guided frame prediction

    Yongfan Chen, Xiuwen Zhu, Tianyu Li, Hao Chen, and Chunhua Shen. A physical coherence benchmark for evalu- ating video generation models via optical flow-guided frame prediction. arXiv preprint arXiv:2502.05503, 2025. 2

    arXiv 2025

  8. [7]

    Nonisotropic gaussian diffusion for realistic 3d human motion prediction

    C Curreli, D Muhle, A Saroha, Z Ye, R Marin, and D Cre- mers. Nonisotropic gaussian diffusion for realistic 3d human motion prediction. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, 2025. 2

    work page 2025

Show all 65 references
  1. [8]

    Objaverse: A universe of annotated 3d objects

    Matt Deitke, Dustin Schwenk, Jordi Salvador, Luca Weihs, Oscar Michel, Eli VanderBilt, Ludwig Schmidt, Kiana Ehsani, Aniruddha Kembhavi, and Ali Farhadi. Objaverse: A universe of annotated 3d objects. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Rec...

    2023

  2. [9]

    Objaverse-xl: A universe of 10m+ 3d objects

    Matt Deitke, Ruoshi Liu, Matthew Wallingford, Huong Ngo, Oscar Michel, Aditya Kusupati, Alan Fan, Christian Laforte, Vikram V oleti, Samir Yitzhak Gadre, et al. Objaverse-xl: A universe of 10m+ 3d objects. Advances in Neural Informa- tion Processing Systems, 36, 2024. 2

    2024

  3. [10]

    Niladri Shekhar Dutt, Sanjeev Muralikrishnan, and Niloy J. Mitra. Diffusion 3d features (diff3f): Decorating untextured shapes with distilled semantic features. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 4494–4504, 2024. 3, 5, 1

    2024

  4. [11]

    Cf vanloan, matrix computations

    Gene H Golub et al. Cf vanloan, matrix computations. The Johns Hopkins, 113(10):23–36, 1996. 5

    1996

  5. [12]

    Nerfdiff: Single-image view synthesis with nerf-guided dis- tillation from 3d-aware diffusion

    Jiatao Gu, Alex Trevithick, Kai-En Lin, Joshua M Susskind, Christian Theobalt, Lingjie Liu, and Ravi Ramamoorthi. Nerfdiff: Single-image view synthesis with nerf-guided dis- tillation from 3d-aware diffusion. In International Confer- ence on Machine Learning , pages 11808–11826. PMLR,

  6. [13]

    Robust statistics

    FR Hampel, EM Ronchetti, PJ Rousseeuw, and W A Stahel. Robust statistics. mathematics, 1986. 5

    1986

  7. [14]

    Lora: Low-rank adaptation of large language models

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen- Zhu, Yuanzhi Li, Shean Wang, Lu Wang, Weizhu Chen, et al. Lora: Low-rank adaptation of large language models. ICLR, 1(2):3, 2022. 5

    2022

  8. [15]

    Animate anyone: Consistent and controllable image-to-video synthesis for character animation

    Li Hu, Xin Gao, Peng Zhang, Ke Sun, Bang Zhang, and Liefeng Bo. Animate anyone: Consistent and controllable image-to-video synthesis for character animation. arXiv preprint arXiv:2311.17117, 2023. 1

    2023 arXiv

  9. [16]

    Continuum Mechanics in Curvilinear Coor- dinates, pages 599–624

    Fridtjov Irgens. Continuum Mechanics in Curvilinear Coor- dinates, pages 599–624. Springer Berlin Heidelberg, Berlin, Heidelberg, 2008. 4

    2008

  10. [17]

    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 Trans. Graph., 42(4):139–1,

  11. [18]

    Dessie: Disentanglement for artic- ulated 3d horse shape and pose estimation from images

    Ci Li, Yi Yang, Zehang Weng, Elin Hernlund, Silvia Zuffi, and Hedvig Kjellstr ¨om. Dessie: Disentanglement for artic- ulated 3d horse shape and pose estimation from images. In Asian Conference on Computer Vision, 2024. 2

    2024

  12. [19]

    Dreammesh4d: Video-to-4d generation with sparse-controlled gaussian- mesh hybrid representation

    Zhiqi Li, Yiming Chen, and Peidong Liu. Dreammesh4d: Video-to-4d generation with sparse-controlled gaussian- mesh hybrid representation. Advances in Neural Information Processing Systems, 37:21377–21400, 2025. 2

    2025

  13. [20]

    Plataniotis, Yao Zhao, and Yunchao Wei

    Hanwen Liang, Yuyang Yin, Dejia Xu, Hanxue Liang, Zhangyang Wang, Konstantinos N. Plataniotis, Yao Zhao, and Yunchao Wei. Diffusion4d: Fast spatial-temporal con- sistent 4d generation via video diffusion models, 2024. 2

    2024

  14. [21]

    Phys4dgen: A physics-driven framework for con- trollable and efficient 4d content generation from a single image

    Jiajing Lin, Zhenzhong Wang, Shu Jiang, Yongjie Hou, and Min Jiang. Phys4dgen: A physics-driven framework for con- trollable and efficient 4d content generation from a single image. arXiv preprint arXiv:2411.16800, 2024. 3

    2024

  15. [22]

    One-2-3-45: Any single image to 3d mesh in 45 seconds without per-shape optimiza- tion

    Minghua Liu, Chao Xu, Haian Jin, Linghao Chen, Mukund Varma T, Zexiang Xu, and Hao Su. One-2-3-45: Any single image to 3d mesh in 45 seconds without per-shape optimiza- tion. Advances in Neural Information Processing Systems , 36, 2024. 2

    2024

  16. [23]

    Zero-1-to- 3: Zero-shot one image to 3d object

    Ruoshi Liu, Rundi Wu, Basile Van Hoorick, Pavel Tok- makov, Sergey Zakharov, and Carl V ondrick. Zero-1-to- 3: Zero-shot one image to 3d object. In Proceedings of the IEEE/CVF international conference on computer vision, pages 9298–9309, 2023. 2

    2023

  17. [24]

    Syncdreamer: Gen- erating multiview-consistent images from a single-view im- age

    Yuan Liu, Cheng Lin, Zijiao Zeng, Xiaoxiao Long, Lingjie Liu, Taku Komura, and Wenping Wang. Syncdreamer: Gen- erating multiview-consistent images from a single-view im- age. arXiv preprint arXiv:2309.03453, 2023

    2023 arXiv

  18. [25]

    Wonder3d: Sin- gle image to 3d using cross-domain diffusion

    Xiaoxiao Long, Yuan-Chen Guo, Cheng Lin, Yuan Liu, Zhiyang Dou, Lingjie Liu, Yuexin Ma, Song-Hai Zhang, 9 Marc Habermann, Christian Theobalt, et al. Wonder3d: Sin- gle image to 3d using cross-domain diffusion. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and ...

    2024

  19. [26]

    Clay: A controllable large-scale generative model for creat- ing high-quality 3d assets

    Zhang Longwen, Wang Ziyu, Zhang Qixuan, Qiu Qiwei, Pang Anqi, Jiang Haoran, Yang Wei, Xu Lan, and Yu Jingyi. Clay: A controllable large-scale generative model for creat- ing high-quality 3d assets. arXiv preprint arXiv:2406.13897,

    arXiv

  20. [27]

    Smooth non-rigid shape matching via effective dirichlet energy optimization

    Robin Magnet, Jing Ren, Olga Sorkine-Hornung, and Maks Ovsjanikov. Smooth non-rigid shape matching via effective dirichlet energy optimization. In 2022 International Confer- ence on 3D Vision (3DV), pages 495–504. IEEE, 2022. 4

    2022

  21. [28]

    Realfusion: 360deg reconstruction of any object from a single image

    Luke Melas-Kyriazi, Iro Laina, Christian Rupprecht, and Andrea Vedaldi. Realfusion: 360deg reconstruction of any object from a single image. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 8446–8455, 2023. 2

    2023

  22. [29]

    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. 2

    2021

  23. [30]

    Gen- eralized connectivity constraints for spatio-temporal 3d re- construction

    Martin Ralf Oswald, Jan St ¨uhmer, and Daniel Cremers. Gen- eralized connectivity constraints for spatio-temporal 3d re- construction. In Computer Vision–ECCV 2014: 13th Eu- ropean Conference, Zurich, Switzerland, September 6-12, 2014, Proceedings, Part IV 13, pages 32–46. Spr...

    2014

  24. [31]

    Functional maps: a flexible representation of maps between shapes

    Maks Ovsjanikov, Mirela Ben-Chen, Justin Solomon, Adrian Butscher, and Leonidas Guibas. Functional maps: a flexible representation of maps between shapes. ACM Transactions on Graphics (TOG), 31(4):30:1–30:11, 2012. 4

    2012

  25. [32]

    Sdxl: Improving latent diffusion mod- els for high-resolution image synthesis

    Dustin Podell, Zion English, Kyle Lacey, Andreas Blattmann, Tim Dockhorn, Jonas M ¨uller, Joe Penna, and Robin Rombach. Sdxl: Improving latent diffusion mod- els for high-resolution image synthesis. arXiv preprint arXiv:2307.01952, 2023. 2

    2023 arXiv

  26. [33]

    Dreamfusion: Text-to-3d using 2d diffusion

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

    2022 arXiv

  27. [34]

    D-nerf: Neural radiance fields for dynamic scenes

    Albert Pumarola, Enric Corona, Gerard Pons-Moll, and Francesc Moreno-Noguer. D-nerf: Neural radiance fields for dynamic scenes. In Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition , pages 10318–10327, 2021. 1

    2021

  28. [35]

    Magic123: One image to high-quality 3d object generation using both 2d and 3d diffusion priors.arXiv preprint arXiv:2306.17843,

    Guocheng Qian, Jinjie Mai, Abdullah Hamdi, Jian Ren, Aliaksandr Siarohin, Bing Li, Hsin-Ying Lee, Ivan Sko- rokhodov, Peter Wonka, Sergey Tulyakov, et al. Magic123: One image to high-quality 3d object generation using both 2d and 3d diffusion priors.arXiv preprint arXiv:2306.17843,

    arXiv

  29. [36]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022. 2, 5

    2022

  30. [37]

    Implicit neural surface deformation with explicit velocity fields

    Lu Sang, Zehranaz Canfes, Dongliang Cao, Florian Bernard, and Daniel Cremers. Implicit neural surface deformation with explicit velocity fields. In ICLR, 2025. 4

    2025

  31. [39]

    Augmented Real- ity: Principles and Practice

    Dieter Schmalstieg and Tobias Hollerer. Augmented Real- ity: Principles and Practice . Addison-Wesley Professional,

  32. [40]

    Dreammover: Leveraging the prior of diffusion models for image interpolation with large motion

    Liao Shen, Tianqi Liu, Huiqiang Sun, Xinyi Ye, Baopu Li, Jianming Zhang, and Zhiguo Cao. Dreammover: Leveraging the prior of diffusion models for image interpolation with large motion. arXiv preprint arXiv:2409.09605, 2024. 5, 6, 7, 8, 1, 2

    2024 arXiv

  33. [41]

    Anything- 3d: Towards single-view anything reconstruction in the wild

    Qiuhong Shen, Xingyi Yang, and Xinchao Wang. Anything- 3d: Towards single-view anything reconstruction in the wild. arXiv preprint arXiv:2304.10261, 2023. 2

    2023 arXiv

  34. [42]

    Zero123++: a single image to consistent multi-view dif- fusion base model

    Ruoxi Shi, Hansheng Chen, Zhuoyang Zhang, Minghua Liu, Chao Xu, Xinyue Wei, Linghao Chen, Chong Zeng, and Hao Su. Zero123++: a single image to consistent multi-view dif- fusion base model. arXiv preprint arXiv:2310.15110, 2023. 2

    arXiv

  35. [43]

    Dimensionx: Create any 3d and 4d scenes from a single image with controllable video diffusion

    Wenqiang Sun, Shuo Chen, Fangfu Liu, Zilong Chen, Yueqi Duan, Jun Zhang, and Yikai Wang. Dimensionx: Create any 3d and 4d scenes from a single image with controllable video diffusion. arXiv preprint arXiv:2411.04928, 2024. 2, 3

    2024 arXiv

  36. [44]

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

    Stanislaw Szymanowicz, Chrisitian Rupprecht, and Andrea Vedaldi. Splatter image: Ultra-fast single-view 3d recon- struction. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 10208– 10217, 2024. 2

    2024

  37. [45]

    Make-it-3d: High-fidelity 3d creation from a single image with diffusion prior

    Junshu Tang, Tengfei Wang, Bo Zhang, Ting Zhang, Ran Yi, Lizhuang Ma, and Dong Chen. Make-it-3d: High-fidelity 3d creation from a single image with diffusion prior. In Pro- ceedings of the IEEE/CVF international conference on com- puter vision, pages 22819–22829, 2023. 2

    2023

  38. [46]

    Hunyuan3d 2.0: Scaling diffu- sion models for high resolution textured 3d assets generation,

    Tencent Hunyuan3D Team. Hunyuan3d 2.0: Scaling diffu- sion models for high resolution textured 3d assets generation,

  39. [47]

    Rfnet-4d++: Joint object reconstruction and flow estimation from 4d point clouds with cross-attention spatio-temporal features

    Tuan-Anh Vu, Duc Thanh Nguyen, Binh-Son Hua, Quang- Hieu Pham, and Sai-Kit Yeung. Rfnet-4d++: Joint object reconstruction and flow estimation from 4d point clouds with cross-attention spatio-temporal features. 2024. 3

    2024

  40. [48]

    4d-dress: A 4d dataset of real-world human clothing with se- mantic annotations

    Wenbo Wang, Hsuan-I Ho, Chen Guo, Boxiang Rong, Ar- tur Grigorev, Jie Song, Juan Jose Zarate, and Otmar Hilliges. 4d-dress: A 4d dataset of real-world human clothing with se- mantic annotations. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (...

    2024

  41. [49]

    Behind the scenes: Density fields for single view reconstruction

    Felix Wimbauer, Nan Yang, Christian Rupprecht, and Daniel Cremers. Behind the scenes: Density fields for single view reconstruction. arXiv preprint arXiv:2301.07668, 2023. 1

    2023 arXiv

  42. [50]

    Unique3d: High-quality and efficient 3d mesh generation from a single image

    Kailu Wu, Fangfu Liu, Zhihan Cai, Runjie Yan, Hanyang Wang, Yating Hu, Yueqi Duan, and Kaisheng Ma. Unique3d: High-quality and efficient 3d mesh generation from a single image. arXiv preprint arXiv:2405.20343, 2024. 2 10

    2024 arXiv

  43. [51]

    Neurallift-360: Lifting an in-the-wild 2d photo to a 3d object with 360deg views

    Dejia Xu, Yifan Jiang, Peihao Wang, Zhiwen Fan, Yi Wang, and Zhangyang Wang. Neurallift-360: Lifting an in-the-wild 2d photo to a 3d object with 360deg views. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4479–4489, 2023. 2

    2023

  44. [52]

    Instantmesh: Efficient 3d mesh generation from a single image with sparse-view large reconstruction models

    Jiale Xu, Weihao Cheng, Yiming Gao, Xintao Wang, Shenghua Gao, and Ying Shan. Instantmesh: Efficient 3d mesh generation from a single image with sparse-view large reconstruction models. arXiv preprint arXiv:2404.07191 ,

    arXiv

  45. [53]

    Grm: Large gaussian reconstruction model for ef- ficient 3d reconstruction and generation

    Yinghao Xu, Zifan Shi, Wang Yifan, Hansheng Chen, Ceyuan Yang, Sida Peng, Yujun Shen, and Gordon Wet- zstein. Grm: Large gaussian reconstruction model for ef- ficient 3d reconstruction and generation. In European Con- ference on Computer Vision, pages 1–20. Springer, 2024. 2

    2024

  46. [54]

    Human 3diffusion: Realistic avatar creation via explicit 3d consistent diffusion models

    Yuxuan Xue, Xianghui Xie, Riccardo Marin, and Gerard Pons-Moll. Human 3diffusion: Realistic avatar creation via explicit 3d consistent diffusion models. Arxiv, 2024. 2

    2024

  47. [55]

    Nerf-ds: Neural ra- diance fields for dynamic specular objects

    Zhiwen Yan, Chen Li, and Gim Hee Lee. Nerf-ds: Neural ra- diance fields for dynamic specular objects. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 8285–8295, 2023. 1

    2023

  48. [56]

    3dshape2vecset: A 3d shape representation for neu- ral fields and generative diffusion models.ACM Transactions on Graphics (TOG), 42(4):1–16, 2023

    Biao Zhang, Jiapeng Tang, Matthias Niessner, and Peter Wonka. 3dshape2vecset: A 3d shape representation for neu- ral fields and generative diffusion models.ACM Transactions on Graphics (TOG), 42(4):1–16, 2023. 2

    2023

  49. [57]

    4diffusion: Multi-view video diffusion model for 4d generation

    Haiyu Zhang, Xinyuan Chen, Yaohui Wang, Xihui Liu, Yunhong Wang, and Yu Qiao. 4diffusion: Multi-view video diffusion model for 4d generation. arXiv preprint arXiv:2405.20674, 2024. 2

    2024 arXiv

  50. [58]

    4diffusion: Multi-view video dif- fusion model for 4d generation

    Haiyu Zhang, Xinyuan Chen, Yaohui Wang, Xihui Liu, Yun- hong Wang, and Yu Qiao. 4diffusion: Multi-view video dif- fusion model for 4d generation. Advances in Neural Infor- mation Processing Systems, 37:15272–15295, 2025. 2

    2025

  51. [59]

    Diffmorpher: Unleashing the capability of diffusion models for image morphing

    Kaiwen Zhang, Yifan Zhou, Xudong Xu, Xingang Pan, and Bo Dai. Diffmorpher: Unleashing the capability of diffusion models for image morphing. arXiv preprint arXiv:2312.07409, 2023. 5, 6, 7, 8, 1, 2

    2023 arXiv

  52. [60]

    Sifu: Side-view conditioned implicit function for real-world us- able clothed human reconstruction

    Zechuan Zhang, Zongxin Yang, and Yi Yang. Sifu: Side-view conditioned implicit function for real-world us- able clothed human reconstruction. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 9936–9947, 2024. 2

    2024

  53. [61]

    Sifu: Side-view conditioned implicit function for real-world us- able clothed human reconstruction

    Zechuan Zhang, Zongxin Yang, and Yi Yang. Sifu: Side-view conditioned implicit function for real-world us- able clothed human reconstruction. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 9936–9947, 2024. 2

    2024

  54. [62]

    Michelangelo: Conditional 3d shape generation based on shape-image-text aligned latent representation

    Zibo Zhao, Wen Liu, Xin Chen, Xianfang Zeng, Rui Wang, Pei Cheng, Bin Fu, Tao Chen, Gang Yu, and Shenghua Gao. Michelangelo: Conditional 3d shape generation based on shape-image-text aligned latent representation. Advances in Neural Information Processing Systems, 36, 2024. 2

    2024

  55. [63]

    Ar4d: Autoregressive 4d generation from monocular videos

    Hanxin Zhu, Tianyu He, Xiqian Yu, Junliang Guo, Zhibo Chen, and Jiang Bian. Ar4d: Autoregressive 4d generation from monocular videos. arXiv preprint arXiv:2501.01722 ,

    arXiv

  56. [64]

    Triplane meets gaussian splatting: Fast and generalizable single-view 3d reconstruction with transformers

    Zi-Xin Zou, Zhipeng Yu, Yuan-Chen Guo, Yangguang Li, Ding Liang, Yan-Pei Cao, and Song-Hai Zhang. Triplane meets gaussian splatting: Fast and generalizable single-view 3d reconstruction with transformers. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern...

    2024

  57. [65]

    Silvia Zuffi, Ylva Mellbin, Ci Li, Markus Hoeschle, Hedvig Kjellstr¨om, Senya Polikovsky, Elin Hernlund, and Michael J. Black. V AREN: Very accurate and realistic equine network. In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024. 2 11 TwoSquared: 4...

    2024

Pith tools

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