Pith. sign in

REVIEW 4 major objections 5 minor 59 references

Semi-supervised Single-view 3D Reconstruction via Multi Shape Prior Fusion Strategy and Self-Attention

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read The paper claims a semi-supervised network can reconstruct 3D point clouds from a single image using as little as 1% labeled data, beating supervised baselines on ShapeNet and Pix3D.

desk verdict Solid, incremental semi-supervised point-cloud reconstruction paper with a category-mean prior that is never fully specified; the reported gains are consistent but small and possibly confounded by prior provenance. read the letter →

arxiv 2411.15420 v1 pith:5KTZPVCD submitted 2024-11-23 cs.CV

Wei Zhoua , Xinzhe Shia , Yunfeng Shea , Kunlong Liua , Yongqin Zhanga This is my paper · ORCID
classification cs.CV
keywords single-view3Dreconstructionsemi-supervisedlearningpointcloudshapepriorfusionself-attentionpseudo-labelingChamferdistanceteacher-student
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 sets out to show that single-view 3D reconstruction does not need large labeled datasets. It proposes a semi-supervised framework that trains a point-cloud deformation network on a few labeled image-shape pairs plus many unlabeled images, and claims this is the first such method to operate directly on point clouds. On ShapeNet with only 1%, 10%, and 20% labeled data, the method reports lower Chamfer distances than supervised baselines, reaching 5.91 at 20%. The same recipe transfers to the real-world Pix3D dataset at 10% labeling. The practical stake is that 3D annotation, which is expensive, can be largely replaced by unlabeled images.

What carries the argument

The load-bearing object is the multi shape prior fusion point cloud $P_o$, a category-level mean shape obtained by encoding training point clouds with a pretrained 3D autoencoder, running K-means on the latent codes to get $K$ shape prototypes $\hat P_k$, and fusing them with inverse-Chamfer-distance weights: $P_o = \sum_i P_i \cdot \frac{1/d_{CD}(P_i,P_{avg})}{\sum_j 1/d_{CD}(P_j,P_{avg})}$. This fused cloud replaces the sphere as the deformable starting shape; image features then displace each point through an attribute flow encoder, a PointTransformer-based shape matching deformer with adaptive instance normalization, and a self-attention decoder. The semi-supervised loop uses an EMA-updated teacher to produce pseudo-labels for unlabeled images and a student trained with image-level and feature-level perturbations.

What would settle it

Compute the fusion prior twice: once from the labeled subset only and once from the entire training pool including unlabeled shapes, then retrain at 1%, 10%, and 20% labels. If the gap over supervised baselines shrinks or vanishes when the prior excludes unlabeled geometry, the reported gains come from prior leakage rather than from the semi-supervised loop. A second check: hold out one ShapeNet category from the prior and test on it; if reconstruction collapses, the method depends on seeing the category in the prior rather than on the image.

Watch

Extended reading notes

Core claim

The central discovery is that replacing the standard spherical initial point cloud with a category mean shape, computed by clustering point clouds in a pretrained autoencoder latent space and weighting cluster centers by inverse Chamfer distance, gives a deformation-based reconstruction network a much better starting geometry. Combined with a teacher-student semi-supervised loop that generates pseudo-labels for unlabeled images and enforces consistency between weakly, strongly, and feature-perturbed views, the model learns from unlabeled data and outperforms fully supervised methods trained on the same small labeled subsets. The paper reports average L1 Chamfer distances of 7.59, 6.26, and 5.91 on ShapeNet at 1%, 10%, and 20% labeled data, and 6.53 on Pix3D at 10%.

Load-bearing premise

The method only works if the right category-level fusion point cloud is available for every test image, and the paper does not specify how that category is obtained, whether the prior is built only from the labeled split, or how the method behaves for categories never seen in the prior.

Editorial extensions

If this is right

  • At 20% ShapeNet labels, SSMP reaches an average Chamfer distance of 5.91, lower than the strongest supervised baseline 3DAttriFlow at 6.16, with the largest gains in complex categories such as sofa and vessel.
  • Even at 1% labeled data the method reports 7.59, below Pixel2Point (9.86) and 3DAttriFlow (7.75) at the same label budget.
  • The multi shape prior fusion strategy alone improves over a single spherical prior by roughly 0.12 Chamfer distance on ShapeNet and 0.18 on Pix3D.
  • Swapping the MLP decoder for a self-attention decoder lowers Chamfer distance for Pixel2Point, 3DAttriFlow, and the proposed method on both datasets.
  • The semi-supervised design beats generic SSL transplants (MeanTeacher 6.04, MixMatch 6.23, FixMatch 6.10) under the same backbone and data split.

Reading between the lines

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

  • Inference beyond the paper: if the fusion prior is computed from the full unlabeled training pool rather than only the labeled subset, part of the reported gain could come from geometry information that is functionally equivalent to extra labels; the paper does not specify which split builds the prior.
  • Inference beyond the paper: the method assumes a category label is available at test time to select the matching fusion cloud; a natural extension is to predict the category from the image itself or to learn a per-instance prior, which would extend the method to unseen categories.
  • Inference beyond the paper: the self-attention decoder's gains appear partly independent of the semi-supervised setup, since the ablation shows it also improves supervised baselines, so the two contributions could be combined separately in future systems.
  • Inference beyond the paper: a direct stress test is to shrink the labeled fraction to 1% and measure how the fusion prior degrades, since the prior's fidelity likely depends on the amount and diversity of data used to compute it.
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 SSMP, a semi-supervised learning framework for single-view 3D point cloud reconstruction. The method combines a teacher-student pseudo-labeling pipeline with EMA updates and two architectural additions: a "multi shape prior fusion strategy" that initializes the deformation network from a category-level average point cloud, and a self-attention decoder. The fusion prior is obtained by KMeans clustering point clouds in a pretrained 3D autoencoder's latent space and Chamfer-distance-weighted averaging of the decoded cluster centers. Experiments on ShapeNet at 1%, 10%, and 20% labeled data and on Pix3D at 10% labeled data report L1 Chamfer distances below the supervised baselines considered. The authors claim this is the first semi-supervised point cloud reconstruction method from a single image, and they release code at the stated GitHub URL.

Significance. If the reported gains are robust and leak-free, the paper makes a useful empirical contribution: it shows that a standard SSL loop, combined with a category-level shape prior, can improve single-view point cloud reconstruction under small label budgets. The paper includes several strengths: code is released, comparisons are made against three image-domain SSL methods adapted to reconstruction (Mean-Teacher, MixMatch, FixMatch), and ablations are provided for the fusion strategy, the fusion weighting, and the decoder type. The main result is plausible and the experimental design is mostly standard. However, the central claim currently rests on single-run comparisons with no variance information, and the construction of the shape prior leaves open a potentially severe data-leakage confound and an unspecified test-time category requirement. The significance is therefore conditional on those issues being resolved.

major comments (4)
  1. [Section 3.1, Eqs. (1)-(3); Section 4.1] The fusion prior Po is category-specific, but the paper never states how the category is supplied at test time. Section 4.1 says only that the training set was randomly divided into supervised and unlabeled data; the test-time image is described as a "single image" in the contribution claim. If the method requires the object category to select Po, this must be stated explicitly, and the baselines should either receive the same category information or the evaluation should be category-conditional. As written, the comparison in Table 3 is not a like-for-like single-image setting.
  2. [Section 3.1, Eqs. (1)-(3); Section 4.1] The paper does not specify the data used to pretrain the 3D autoencoder or to compute the KMeans clusters and fusion point cloud. If the autoencoder and clustering are run on the full ShapeNet set of 36,169 models, the prior will contain geometry from the test split, giving SSMP an advantage over the sphere-initialized baselines. The authors should define the exact split used for the prior computation, recompute the prior using only the training (labeled plus unlabeled) split, and show whether the reported 5.91 versus 6.16 result in Table 2 survives this protocol.
  3. [Section 4.2, Tables 1-3] All quantitative results are reported as single point estimates with no error bars, no seeds, and no significance tests. The margins over the closest baseline are small at every ShapeNet ratio (7.59 versus 7.75 at 1%, 6.26 versus 6.39 at 10%, and 5.91 versus 6.16 at 20%), and the labeled/unlabeled split itself is random. Without multiple seeds or a significance test, the claim that SSMP "consistently outperforms" the baselines at each data proportion is not yet established.
  4. [Section 4.3, Tables 4-7] The ablation tables do not unambiguously isolate the contributions of the fusion prior, the SSL strategy, and the self-attention decoder. Table 4 reports 6.27 for "Ours w/o" and 6.04 for Mean-Teacher, while Table 5 reports 6.03 for "Single Shape Priors Strategy" and Table 7 reports 5.98 for "Ours MLP Decoder". It is unclear which rows include the semi-supervised stage, which rows use the fusion prior, and why an SSL-ablated model can be worse than the single-shape-prior model. Each configuration should be defined as a complete method, for example "SSMP without SSL" or "SSMP with sphere prior and self-attention decoder", so the ablation supports the attribution claims in the text.
minor comments (5)
  1. [Section 3.1, Eq. (2)] The sentence beginning "In [43], the average shape is derived..." cites reference [43] (AlphaMatch), which is not a voxel shape prior method; this appears to be an incorrect citation and should be replaced with the intended shape-prior reference such as [49] or [50].
  2. [Section 3.1, Eq. (3)] In Eq. (3), the quantities Pi, Pavg, and the index ranges are not clearly defined; in particular, Pavg is never defined in the text, and the summation indices differ from those in Eq. (1). Please define all terms and make the normalization explicit.
  3. [Section 4.2, Table 1 and surrounding text] The text states that in the miscellaneous category there is "a slight increase in error", but the table shows Ours (13.44) is worse than both Pixel2point (12.20) and Pix2Vox (10.21), while only better than 3DAttriFlow (13.60). This discrepancy should be acknowledged more precisely.
  4. [Section 4.2] The paper uses inconsistent capitalization and naming for the baselines, including "3DAttriFlow" in Table 2 but "3D attributeflow" in Table 1, and "AltasNet" in the text where reference [55] is "AtlasNet".
  5. [Section 4.2] The word "significantly" is used in the Pix3D discussion (e.g., "achieved an optimal performance ... significantly outperforming") without any statistical test; the wording should be softened to "reports a lower Chamfer distance" unless significance testing is added.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reconstruction results are externally benchmarked, and the fusion prior is dataset-statistic initialization rather than a self-defined input-output loop.

full rationale

The paper's contributions are empirical and are evaluated against external supervised baselines (Pixel2Point, 3DAttriFlow, AtlasNet, Pix2Vox, OccNet) and standard SSL extensions (MeanTeacher, MixMatch, FixMatch) on ShapeNet and Pix3D, so the headline gain is not manufactured from its own fitted values. The multi shape prior fusion strategy (Eqs. 1-3) is an initialization computed by clustering pretrained latent codes of point clouds and Chamfer-distance weighting; it is not defined in terms of the predicted output and does not force the reported Chamfer distances by construction. The teacher-student pseudo-label/EMA loop is ordinary self-training with consistency losses, not a derivation whose conclusion equals its premise. SSP3D is cited as inspiration, but no load-bearing uniqueness theorem or circular ansatz is imported from the authors' prior work. Possible concerns about whether the shape prior was computed on held-out geometry or whether a category label is needed at test time are dataset-hygiene and applicability questions, not circular reductions, and the text does not exhibit the specific reduction needed to substantiate circularity.

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

The paper introduces no new physical or mathematical entities. The "fusion point cloud" and "feature perturbation" are algorithmic constructs, not postulated entities requiring independent evidence.

free parameters (4)
  • K (number of KMeans clusters in shape prior)
    Number of clusters for shape prototypes in Eq. (1); never reported, so the granularity of the category prior is unspecified.
  • xi (orthogonality loss weight) = 100
    Set by hand in Section 3.3, Eq. (16); balances L_Orth against reconstruction loss.
  • lambda and nu (unsupervised loss weights) = 0.5
    Set to 0.5 in Section 3.3, Eq. (17); governs pseudo-label and consistency losses.
  • EMA momentum alpha = 0.9996
    Set in Section 4.1 for teacher updating; cosine scheduled toward 1 in Section 3.2.
assumptions (4)
  • domain assumption A pre-trained 3D autoencoder's latent space is structured enough that KMeans clustering yields meaningful shape prototypes.
    Stated in Section 3.1: "the latent representations learned by the auto-encoder can capture the main features and structural information of point cloud data." No analysis or benchmark validates this.
  • domain assumption The category-specific fusion point cloud is available at test time.
    The method feeds a category mean shape as initial input (Fig. 3, Section 3.1), but the paper does not explain how the category is obtained for a novel image.
  • domain assumption Chamfer-distance-weighted averaging of decoded cluster centers produces a better initial point cloud than a sphere.
    Supported only by the ablation in Table 5, which shows a small improvement (6.03 to 5.91 on ShapeNet) without error bars.
  • domain assumption SSL consistency regularization and pseudo-labels transfer from 2D image classification to 3D point cloud regression.
    The paper adapts MeanTeacher/MixMatch/FixMatch-style losses (Section 2.2, Eq. 17) without a theoretical or empirical study of when these assumptions hold for point cloud outputs.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Semi-supervised Single-view 3D Reconstruction via Multi Shape Prior Fusion Strategy and Self-Attention." pith.science (2026). https://pith.science/paper/5KTZPVCD

@misc{pith2026241115420,
  author       = {Pith},
  title        = {Pith review of: Semi-supervised Single-view 3D Reconstruction via Multi Shape Prior Fusion Strategy and Self-Attention},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5KTZPVCD}},
  note         = {Machine review of arXiv:2411.15420}
}
read the original abstract

In the domain of single-view 3D reconstruction, traditional techniques have frequently relied on expensive and time-intensive 3D annotation data. Facing the challenge of annotation acquisition, semi-supervised learning strategies offer an innovative approach to reduce the dependence on labeled data. Despite these developments, the utilization of this learning paradigm in 3D reconstruction tasks remains relatively constrained. In this research, we created an innovative semi-supervised framework for 3D reconstruction that distinctively uniquely introduces a multi shape prior fusion strategy, intending to guide the creation of more realistic object structures. Additionally, to improve the quality of shape generation, we integrated a self-attention module into the traditional decoder. In benchmark tests on the ShapeNet dataset, our method substantially outperformed existing supervised learning methods at diverse labeled ratios of 1\%, 10\%, and 20\%. Moreover, it showcased excellent performance on the real-world Pix3D dataset. Through comprehensive experiments on ShapeNet, our framework demonstrated a 3.3\% performance improvement over the baseline. Moreover, stringent ablation studies further confirmed the notable effectiveness of our approach. Our code has been released on https://github.com/NWUzhouwei/SSMP

Figures

Figures reproduced from arXiv: 2411.15420 by the authors.

Figure 1
Figure 1. Comparison of reconstruction results using the initial spherical and [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The differences between semi-supervised learning, supervised learning and unsupervised learning. (a) Illustration of supervised single-view 3D reconstruction, which requires a large amount of labeled data pairs.(b) Illustration of semi-supervised single-view 3D reconstruction, our proposed SSMP (Semi-Supervised Multi Shape Prior Fusion Reconstruction) model can predict the 3D shape of unlabeled images after training… view at source ↗
Figure 3
Figure 3. SSMP consists of two stages. The warm-up stage: we train the 3D reconstruction network using the available supervised data and the fusion shape [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: (a) SSP3D baseline. (b) Our proposed feature perturbations method [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Examples of single-view 3D reconstruction on the Pix3D dataset using only 10% labeled data. [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Examples of single-view 3D reconstruction on the ShapeNet dataset using only 20% labeled data. [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 56 canonical work pages

  1. [43]

    Alphamatch: Improving consistency for semi-supervised learning with alpha-divergence

    Gong, C, Wang, D, Liu, Q. Alphamatch: Improving consistency for semi-supervised learning with alpha-divergence. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2021, p. 13683–13692

    work page 2021

  2. [1]

    Topological 3d reconstruction of multiple anatomical structures from volumetric medical data

    Gerbaud, S, Cavalier, A, Horna, S, Zrour, R, Naudin, M, Guillevin, C, et al. Topological 3d reconstruction of multiple anatomical structures from volumetric medical data. Computers & Graphics 2024;121:103947. URL: https://www.sciencedirect. com/science/article/pii/S0097849324000827. doi: https: //doi.org/10.1016/j.cag.2024.103947

    arXiv 2024

  3. [2]

    3d face recon- struction from a single image based on hybrid-level contextual infor- mation with weak supervision

    Liu, Y , Ran, T, Yuan, L, Lv, K, Zheng, G. 3d face recon- struction from a single image based on hybrid-level contextual infor- mation with weak supervision. Computers & Graphics 2024;118:80–

    work page 2024

  4. [3]

    Reconfusion: 3d reconstruction with di ffusion priors

    Wu, R, Mildenhall, B, Henzler, P, Park, K, Gao, R, Watson, D, et al. Reconfusion: 3d reconstruction with di ffusion priors. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2024, p. 21551–21561

    work page 2024

  5. [4]

    Tri- plane meets gaussian splatting: Fast and generalizable single-view 3d re- construction with transformers

    Zou, ZX, Yu, Z, Guo, YC, Li, Y , Liang, D, Cao, YP, et al. Tri- plane meets gaussian splatting: Fast and generalizable single-view 3d re- construction with transformers. In: Proceedings of the IEEE /CVF Con- ference on Computer Vision and Pattern Recognition. 2024, p. 10324– 10335

    work page 2024

  6. [5]

    Gamba: Marry gaussian splatting with mamba for single view 3d reconstruction

    Shen, Q, Yi, X, Wu, Z, Zhou, P, Zhang, H, Yan, S, et al. Gamba: Marry gaussian splatting with mamba for single view 3d reconstruction. arXiv preprint arXiv:240318795 2024

    work page 2024

  7. [6]

    3dfires: Few image 3d reconstruc- tion for scenes with hidden surfaces

    Jin, L, Kulkarni, N, Fouhey, DF. 3dfires: Few image 3d reconstruc- tion for scenes with hidden surfaces. In: Proceedings of the IEEE /CVF Conference on Computer Vision and Pattern Recognition. 2024, p. 9742– 9751

    work page 2024

  8. [7]

    Splatter image: Ultra-fast single-view 3d reconstruction

    Szymanowicz, S, Rupprecht, C, Vedaldi, A. Splatter image: Ultra-fast single-view 3d reconstruction. In: Proceedings of the IEEE /CVF Con- ference on Computer Vision and Pattern Recognition. 2024, p. 10208– 10217

    work page 2024

Show all 59 references
  1. [8]

    Pc2: Projection- conditioned point cloud diffusion for single-image 3d reconstruction

    Melas-Kyriazi, L, Rupprecht, C, Vedaldi, A. Pc2: Projection- conditioned point cloud diffusion for single-image 3d reconstruction. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pat- tern Recognition. 2023, p. 12923–12932

    2023

  2. [9]

    3d soc- net: Deep 3d reconstruction network based on self-organizing clustering mapping

    Gan, YS, Chen, W, Yau, WC, Zou, Z, Liong, ST, Wang, SY . 3d soc- net: Deep 3d reconstruction network based on self-organizing clustering mapping. Expert Systems with Applications 2023;213:119209

    2023

  3. [10]

    3d shape recon- struction from 2d images with disentangled attribute flow

    Wen, X, Zhou, J, Liu, YS, Su, H, Dong, Z, Han, Z. 3d shape recon- struction from 2d images with disentangled attribute flow. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2022, p. 3803–3813

    2022

  4. [11]

    3d-r2n2: A unified approach for single and multi-view 3d object reconstruction

    Choy, CB, Xu, D, Gwak, J, Chen, K, Savarese, S. 3d-r2n2: A unified approach for single and multi-view 3d object reconstruction. In: Com- puter Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part VIII 14. Springer; 2016...

    2016

  5. [12]

    Cignet: Category-and- intrinsic-geometry guided network for 3d coarse-to-fine reconstruction

    Gao, J, Kong, D, Wang, S, Li, J, Yin, B. Cignet: Category-and- intrinsic-geometry guided network for 3d coarse-to-fine reconstruction. Neurocomputing 2023;554:126607

    2023

  6. [13]

    Pix2vox: Context-aware 3d reconstruction from single and multi-view images

    Xie, H, Yao, H, Sun, X, Zhou, S, Zhang, S. Pix2vox: Context-aware 3d reconstruction from single and multi-view images. In: Proceedings of the IEEE/CVF international conference on computer vision. 2019, p. 2690–2698

    2019

  7. [14]

    Pix2vox ++: Multi-scale context-aware 3d object reconstruction from single and multiple images

    Xie, H, Yao, H, Zhang, S, Zhou, S, Sun, W. Pix2vox ++: Multi-scale context-aware 3d object reconstruction from single and multiple images. International Journal of Computer Vision 2020;128(12):2919–2935. Preprint Submitted for review/ Computers & Graphics (2024) 13

    2024

  8. [15]

    Legoformer: Trans- formers for block-by-block multi-view 3d reconstruction

    Yagubbayli, F, Wang, Y , Tonioni, A, Tombari, F. Legoformer: Trans- formers for block-by-block multi-view 3d reconstruction. arXiv preprint arXiv:210612102 2021

    2021

  9. [16]

    3d-c2ft: Coarse-to-fine trans- former for multi-view 3d reconstruction

    Tiong, LCO, Sigmund, D, Teoh, ABJ. 3d-c2ft: Coarse-to-fine trans- former for multi-view 3d reconstruction. In: Proceedings of the Asian Conference on Computer Vision. 2022, p. 1438–1454

    2022

  10. [17]

    Pixel2mesh: Generating 3d mesh models from single rgb images

    Wang, N, Zhang, Y , Li, Z, Fu, Y , Liu, W, Jiang, YG. Pixel2mesh: Generating 3d mesh models from single rgb images. In: Proceedings of the European conference on computer vision (ECCV). 2018, p. 52–67

    2018

  11. [18]

    Std- net: Structure-preserving and topology-adaptive deformation network for single-view 3d reconstruction

    Mao, A, Dai, C, Liu, Q, Yang, J, Gao, L, He, Y , et al. Std- net: Structure-preserving and topology-adaptive deformation network for single-view 3d reconstruction. IEEE Transactions on Visualization and Computer Graphics 2021;29(3):1785–1798

    2021

  12. [19]

    Pixel2mesh ++: 3d mesh generation and refinement from multi-view images

    Wen, C, Zhang, Y , Cao, C, Li, Z, Xue, X, Fu, Y . Pixel2mesh ++: 3d mesh generation and refinement from multi-view images. IEEE Trans- actions on Pattern Analysis and Machine Intelligence 2022;45(2):2166– 2180

    2022

  13. [20]

    Single-view 3d mesh reconstruction for seen and unseen categories

    Yang, X, Lin, G, Zhou, L. Single-view 3d mesh reconstruction for seen and unseen categories. IEEE transactions on image processing 2023;32:3746–3758

    2023

  14. [21]

    T-pixel2mesh: Combining global and local transformer for 3d mesh generation from a single image

    Zhang, S, Jiang, B, He, K, Zhu, J, Tai, Y , Wang, C, et al. T-pixel2mesh: Combining global and local transformer for 3d mesh generation from a single image. In: ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE; 2024, p. 2435–2439

    2024

  15. [22]

    A point set generation network for 3d object reconstruction from a single image

    Fan, H, Su, H, Guibas, LJ. A point set generation network for 3d object reconstruction from a single image. In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2017, p. 605–613

    2017

  16. [23]

    Part-wise atlasnet for 3d point cloud reconstruc- tion from a single image

    Yu, Q, Yang, C, Wei, H. Part-wise atlasnet for 3d point cloud reconstruc- tion from a single image. Knowledge-Based Systems 2022;242:108395

    2022

  17. [24]

    Pixel2point: 3d object reconstruction from a single image using cnn and initial sphere

    Afifi, AJ, Magnusson, J, Soomro, TA, Hellwich, O. Pixel2point: 3d object reconstruction from a single image using cnn and initial sphere. IEEE Access 2020;9:110–121

    2020

  18. [25]

    Nerf: Representing scenes as neural radiance fields for view synthesis

    Mildenhall, B, Srinivasan, PP, Tancik, M, Barron, JT, Ramamoorthi, R, Ng, R. Nerf: Representing scenes as neural radiance fields for view synthesis. Communications of the ACM 2021;65(1):99–106

    2021

  19. [26]

    Lrm: Large reconstruction model for single image to 3d

    Hong, Y , Zhang, K, Gu, J, Bi, S, Zhou, Y , Liu, D, et al. Lrm: Large reconstruction model for single image to 3d. arXiv preprint arXiv:231104400 2023

    2023

  20. [27]

    Vi- sion transformer for nerf-based view synthesis from a single input image

    Lin, KE, Lin, YC, Lai, WS, Lin, TY , Shih, YC, Ramamoorthi, R. Vi- sion transformer for nerf-based view synthesis from a single input image. In: Proceedings of the IEEE /CVF Winter Conference on Applications of Computer Vision. 2023, p. 806–815

    2023

  21. [28]

    pixelnerf: Neural radiance fields from one or few images

    Yu, A, Ye, V , Tancik, M, Kanazawa, A. pixelnerf: Neural radiance fields from one or few images. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2021, p. 4578–4587

    2021

  22. [29]

    Mip-nerf: A multiscale representation for anti-aliasing neural radiance fields

    Barron, JT, Mildenhall, B, Tancik, M, Hedman, P, Martin-Brualla, R, Srinivasan, PP. Mip-nerf: A multiscale representation for anti-aliasing neural radiance fields. In: Proceedings of the IEEE /CVF international conference on computer vision. 2021, p. 5855–5864

    2021

  23. [30]

    Latent-nerf for shape-guided generation of 3d shapes and textures

    Metzer, G, Richardson, E, Patashnik, O, Giryes, R, Cohen-Or, D. Latent-nerf for shape-guided generation of 3d shapes and textures. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pat- tern Recognition. 2023, p. 12663–12673

    2023

  24. [31]

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

    Poole, B, Jain, A, Barron, JT, Mildenhall, B. Dreamfusion: Text-to-3d using 2d diffusion. arXiv preprint arXiv:220914988 2022

    2022

  25. [32]

    Shap-e: Generating conditional 3d implicit functions

    Jun, H, Nichol, A. Shap-e: Generating conditional 3d implicit functions. arXiv preprint arXiv:230502463 2023

    2023

  26. [33]

    Point-e: A system for generating 3d point clouds from complex prompts

    Nichol, A, Jun, H, Dhariwal, P, Mishkin, P, Chen, M. Point-e: A system for generating 3d point clouds from complex prompts. arXiv preprint arXiv:221208751 2022

    2022

  27. [34]

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

    Liu, R, Wu, R, Van Hoorick, B, Tokmakov, P, Zakharov, S, V ondrick, C. Zero-1-to-3: Zero-shot one image to 3d object. In: Proceedings of the IEEE/CVF international conference on computer vision. 2023, p. 9298– 9309

    2023

  28. [35]

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

    Liu, M, Xu, C, Jin, H, Chen, L, Varma T, M, Xu, Z, et al. One-2-3-45: Any single image to 3d mesh in 45 seconds without per-shape optimiza- tion. Advances in Neural Information Processing Systems 2024;36

    2024

  29. [36]

    Won- der3d: Single image to 3d using cross-domain diffusion

    Long, X, Guo, YC, Lin, C, Liu, Y , Dou, Z, Liu, L, et al. Won- der3d: Single image to 3d using cross-domain diffusion. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion. 2024, p. 9970–9980

    2024

  30. [37]

    Semi-supervised learning by entropy mini- mization

    Grandvalet, Y , Bengio, Y . Semi-supervised learning by entropy mini- mization. Advances in neural information processing systems 2004;17

    2004

  31. [38]

    Pseudo-label: The simple and e fficient semi-supervised learning method for deep neural networks

    Lee, DH, et al. Pseudo-label: The simple and e fficient semi-supervised learning method for deep neural networks. In: Workshop on challenges in representation learning, ICML; vol. 3. Atlanta; 2013, p. 896

    2013

  32. [39]

    Meta pseudo labels

    Pham, H, Dai, Z, Xie, Q, Le, QV . Meta pseudo labels. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2021, p. 11557–11568

    2021

  33. [40]

    Rethinking pre-training and self-training

    Zoph, B, Ghiasi, G, Lin, TY , Cui, Y , Liu, H, Cubuk, ED, et al. Rethinking pre-training and self-training. Advances in neural information processing systems 2020;33:3833–3845

    2020

  34. [41]

    Remixmatch: Semi-supervised learning with distribution alignment and augmentation anchoring

    Berthelot, D, Carlini, N, Cubuk, ED, Kurakin, A, Sohn, K, Zhang, H, et al. Remixmatch: Semi-supervised learning with distribution alignment and augmentation anchoring. arXiv preprint arXiv:191109785 2019

    2019

  35. [42]

    Mixmatch: A holistic approach to semi-supervised learning

    Berthelot, D, Carlini, N, Goodfellow, I, Papernot, N, Oliver, A, Raf- fel, CA. Mixmatch: A holistic approach to semi-supervised learning. Advances in neural information processing systems 2019;32

    2019

  36. [44]

    Unsupervised data augmen- tation for consistency training

    Xie, Q, Dai, Z, Hovy, E, Luong, T, Le, Q. Unsupervised data augmen- tation for consistency training. Advances in neural information processing systems 2020;33:6256–6268

    2020

  37. [45]

    Virtual adversarial train- ing: a regularization method for supervised and semi-supervised learn- ing

    Miyato, T, Maeda, Si, Koyama, M, Ishii, S. Virtual adversarial train- ing: a regularization method for supervised and semi-supervised learn- ing. IEEE transactions on pattern analysis and machine intelligence 2018;41(8):1979–1993

    2018

  38. [46]

    Improved regularization of convolutional neural networks with cutout

    DeVries, T, Taylor, GW. Improved regularization of convolutional neural networks with cutout. arXiv preprint arXiv:170804552 2017

    2017

  39. [47]

    Learning single-view 3d reconstruction with limited pose supervision

    Yang, G, Cui, Y , Belongie, S, Hariharan, B. Learning single-view 3d reconstruction with limited pose supervision. In: Proceedings of the European Conference on Computer Vision (ECCV). 2018, p. 86–101

    2018

  40. [48]

    Ssr: Semi- supervised soft rasterizer for single-view 2d to 3d reconstruction

    Laradji, I, Rodr ´ıguez, P, Vazquez, D, Nowrouzezahrai, D. Ssr: Semi- supervised soft rasterizer for single-view 2d to 3d reconstruction. In: Pro- ceedings of the IEEE/CVF International Conference on Computer Vision. 2021, p. 1427–1436

    2021

  41. [49]

    Semi-supervised single-view 3d reconstruction via prototype shape priors

    Xing, Z, Li, H, Wu, Z, Jiang, YG. Semi-supervised single-view 3d reconstruction via prototype shape priors. In: European Conference on Computer Vision. Springer; 2022, p. 535–551

    2022

  42. [50]

    What do single-view 3d reconstruction networks learn? In: 2019 IEEE /CVF Conference on Computer Vision and Pat- tern Recognition (CVPR)

    Tatarchenko, M, Richter, SR, Ranftl, R, Li, Z, Koltun, V , Brox, T. What do single-view 3d reconstruction networks learn? In: 2019 IEEE /CVF Conference on Computer Vision and Pat- tern Recognition (CVPR). 2019,URL: http://dx.doi.org/10.1109/ cvpr.2019.00352. doi:10.1109/cvpr.2...

    2019

  43. [51]

    Eigengan: Layer-wise eigen-learning for gans

    He, Z, Kan, M, Shan, S. Eigengan: Layer-wise eigen-learning for gans. In: Proceedings of the IEEE /CVF international conference on computer vision. 2021, p. 14408–14417

    2021

  44. [52]

    Shapenet: An information-rich 3d model repository

    Chang, AX, Funkhouser, T, Guibas, L, Hanrahan, P, Huang, Q, Li, Z, et al. Shapenet: An information-rich 3d model repository. arXiv preprint arXiv:151203012 2015

    2015

  45. [53]

    Pix3d: Dataset and methods for single-image 3d shape modeling

    Sun, X, Wu, J, Zhang, X, Zhang, Z, Zhang, C, Xue, T, et al. Pix3d: Dataset and methods for single-image 3d shape modeling. In: Proceed- ings of the IEEE conference on computer vision and pattern recognition. 2018, p. 2974–2983

    2018

  46. [54]

    Fixing weight decay regularization in adam

    Loshchilov, I, Hutter, F, et al. Fixing weight decay regularization in adam. arXiv preprint arXiv:171105101 2017;5

    2017

  47. [55]

    A papier- mˆach´e approach to learning 3d surface generation

    Groueix, T, Fisher, M, Kim, VG, Russell, BC, Aubry, M. A papier- mˆach´e approach to learning 3d surface generation. In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2018, p. 216–224

    2018

  48. [56]

    Mean teachers are better role models: Weight- averaged consistency targets improve semi-supervised deep learning re- sults

    Tarvainen, A, Valpola, H. Mean teachers are better role models: Weight- averaged consistency targets improve semi-supervised deep learning re- sults. Advances in neural information processing systems 2017;30

    2017

  49. [57]

    Fixmatch: Simplifying semi-supervised learning with consis- tency and confidence

    Sohn, K, Berthelot, D, Carlini, N, Zhang, Z, Zhang, H, Ra ffel, CA, et al. Fixmatch: Simplifying semi-supervised learning with consis- tency and confidence. Advances in neural information processing systems 2020;33:596–608

    2020

  50. [58]

    Occupancy networks: Learning 3d reconstruction in function space

    Mescheder, L, Oechsle, M, Niemeyer, M, Nowozin, S, Geiger, A. Occupancy networks: Learning 3d reconstruction in function space. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2019, p. 4460–4470

    2019

  51. [89]

    doi:https://doi.org/10.1016/j.cag

    URL: https://www.sciencedirect.com/science/article/ pii/S0097849323002881. doi:https://doi.org/10.1016/j.cag. 2023.11.007

    2023 doi

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.