Hyperparameter-Free Losses for Model-Based Monocular Reconstruction

Eduard Ramon; Guillermo Ruiz; Thomas Batard; Xavier Gir\'o-i-Nieto

REVIEW 3 major objections 3 minor 35 references

Hyperparameter-Free Losses for Model-Based Monocular Reconstruction

T0 review · 3 major / 3 minor · reviewed 2026-08-14 · 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 that single-image 3D face reconstruction can be trained with hyperparameter-free geometric losses that match tuned multiterm baselines in accuracy while cutting total training time roughly tenfold.

desk verdict A solid, honest loss-design paper whose MRL random-projection regularization is the real contribution; the 'hyperparameter-free' claim is overstated and the single-run evidence is the main weakness. read the letter →

arxiv 1908.09001 v1 pith:FIETUHZV submitted 2019-08-16 cs.CV cs.AI

Eduard Ramon , Guillermo Ruiz , Thomas Batard , Xavier Giró-i-Nieto This is my paper · ORCID

classification cs.CVcs.AI

keywords hyperparameter-freelosses3Dmorphablemodelmonocularreconstructionfacereprojectionerrorimplicitregularizationrandomvirtualprojectionscameraposeestimation

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish that the labor-intensive step of tuning loss weights can be removed from single-image 3D face reconstruction when a morphable model is used. Its proposed losses fuse shape error and camera-pose error into one geometric term, so a network can be trained once with a fixed learning rate and still match the accuracy of baselines whose weights were selected by Bayesian optimization. The multiview reprojection loss (MRL) adds random virtual camera views as an implicit regularizer, which prevents the flattened-shape failure mode that a pure reprojection loss produces. If the claim is right, training such models becomes faster, simpler, and more reproducible, and the bottleneck shifts from loss engineering to the quality of the underlying 3D registration.

What carries the argument

The load-bearing object is a 3D Morphable Model (3DMM), a PCA-based linear space of face geometry in which a shape is $\hat{x} = m + \Phi_{\mathrm{id}} \hat{\alpha}_{\mathrm{id}}$. The identity that carries the argument is the fusion of shape, rotation, and translation errors into one geometric term: GAL compares $[R(q)|t]x_H$ with $[R(\hat{q})|\hat{t}]\hat{x}_H$; SRL replaces the 3D comparison with the projected difference $||P(q,t)(x_H) - P(\hat{q},\hat{t})(\hat{x}_H)||_1$; MRL averages the SRL error over $V$ random views, with the predicted shape distorted by the isometric transform $D$ that encodes the relative pose between predicted and true cameras. The random projections act as an implicit regularizer on the shape parameters, so no explicit parameter-norm term or weighting hyperparameter is needed.

What would settle it

Retrain the five losses on the same images using a 3DMM built by an independent registration method (different template topology or landmark initialization), then compare their shape errors on the two public face datasets; if the ranking of GAL and MRL against the tuned baselines changes, the reported parity was an effect of the particular PCA basis rather than of the loss formulation.

Watch

Extended reading notes

Core claim

The central claim is that a single, hyperparameter-free loss term can carry the full training signal for a 3DMM-based reconstruction network. The paper demonstrates three such losses: GAL aligns the predicted and ground-truth shapes in 3D space using the predicted camera pose and measures the $\ell^1$ distance; SRL projects both shapes into the image plane and measures reprojection error; MRL repeats the SRL computation over several random camera views, passing the predicted shape through the relative pose between predicted and ground-truth cameras. The experiments on a large internal dataset and two public face datasets show that GAL and MRL reach accuracy comparable to tuned multiterm baselines, with MRL giving the lowest reprojection errors and stable shapes. The conclusion is that geometry itself can replace hand-chosen weights, reducing optimization complexity and total training time while maintaining accuracy, robustness, and generalization.

Load-bearing premise

The load-bearing premise is that the deformable registration of the reference template to the internal 3D scans produces a face-shape space that faithfully covers the test subjects; if that registration is biased, all losses being compared inherit the same distorted shape space and the reported accuracy rankings may not transfer to other faces.

Editorial extensions

If this is right

A 3D face reconstruction model can be trained in a single run with a fixed learning rate, with no loss-weight search, saving roughly an order of magnitude in total training time.
MRL reduces reprojection error to a few pixels while keeping shape error near 3 mm, which is what alignment-critical applications such as augmented reality need.
The shape-accuracy gap between hyperparameter-free and tuned baselines shrinks on previously unseen face datasets, suggesting the geometric losses generalize at least as well as tuned ones.
Two random views are enough for MRL's implicit regularization; more views change shape error by less than a tenth of a millimeter and only add linear compute.

Reading between the lines

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

Not tested in the paper, but the same geometry-fusion trick should transfer to other categories with linear shape models, such as bodies or hands, where pose errors are larger and removing the weighting hyperparameters might matter even more.
MRL's random-view consistency could be reused as a self-supervised fine-tuning signal on unlabelled images, by projecting the predicted shape into random views and minimizing reprojection consistency without any 3D label.
If the central claim holds, run-to-run variability of hyperparameter-free training should be visibly lower than the variability across Bayesian-optimized baselines, because a whole source of tuning variance is removed; comparing random-seed spreads would be a direct test.

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.

Desk Editor's Note

read the letter

Bottom line: this is a competent and mostly honest paper, and the MRL idea is genuinely worth knowing about. The central claim — that you can drop loss-weight tuning without losing accuracy — is plausible but not as firmly established as the paper suggests, because the headline numbers come from single training runs with no variance.

What's new: GAL collapses the usual shape/pose multiterm loss into one geometric alignment term; SRL is a reprojection variant of Kendall-style geometric loss; and MRL is the real contribution. It projects the predicted shape into V random virtual camera views and uses the relative pose between predicted and ground-truth cameras as an isometric transform, so the loss implicitly regularizes the shape without an explicit norm penalty. That is a simple, transferable trick. The math in Eqs. 5-10 checks out, and the paper is clear about SRL's failure mode (flattened shapes), which is honest.

What's done well: fixed architecture and dataset across all losses, Bayesian optimization for baselines (20 runs) vs single run for proposed, and the public release of the loss implementations and the MICC/FaceWarehouse annotations is real reproducibility value. The generalization tables (Table 3) show GAL and MRL tracking the multiterm baselines, which supports the 'comparable' reading.

Soft spots. The biggest is Table 2: one number per metric, no error bars, no seeds. The differences are small (e.g., 2.3 vs 3.0 mm), but with one run we don't know whether that's method or noise. The paper's practical pitch is exactly 'train once and trust,' so this missing variance matters. Second, 'hyperparameter-free' is partly a marketing label: V (number of random views) is chosen empirically in Figure 6 and the L1 norm is chosen by experiments, so the loss is not literally free of tuning. That's a minor wording problem, not a fatal one. Third, the main training data is internal and not shared, so the large-dataset numbers are not independently reproducible. The authors do release the public evaluation data, which mitigates but doesn't eliminate the issue. Fourth, the 3DMM is built via Non-Rigid ICP on their scans; a biased registration would affect all losses equally, so it doesn't invalidate the comparisons, but it is a reason to be cautious about absolute numbers.

The stress-test note is on point. I don't think the conclusion is false, but it is under-evidenced as written. A referee should ask for multi-seed runs with mean/std, a sentence acknowledging V as a hyperparameter, and ideally a release of a subsample of the training data.

For whom: anyone working on 3DMM fitting or single-view face reconstruction. The MRL trick alone is worth a citation.

Recommendation: send to peer review. It deserves a serious referee; with the missing variance addressed it would be an accept.

Referee Report

3 major / 3 minor

Summary. The paper addresses single-view 3D face reconstruction with 3D morphable models. It proposes three losses that avoid per-term weighting hyperparameters by fusing shape, rotation, and translation errors into a single geometric expression: GAL aligns predicted and ground-truth shapes in camera coordinates, SRL minimizes reprojection error, and MRL applies SRL over multiple random virtual views with an isometric transform of the predicted shape to provide implicit regularization. The authors evaluate these losses with a fixed VGG-16 encoder on a private dataset of 6,528 structured-light face scans, reporting shape, pose, and reprojection errors, and they test generalization on MICC and FaceWarehouse. They conclude that GAL and MRL are competitive with Bayesian-tuned multiterm baselines while requiring a single training run, thereby considerably reducing optimization time and complexity.

Significance. If the claims hold, the practical contribution is substantial: model-based monocular reconstruction can be trained without loss-weight search, and MRL introduces an implicit regularizer that needs no additional annotations. The paper's design choices are mostly sound: the geometric derivations are explicit, all losses share the same architecture and training data, and the authors commit to releasing implementations and external annotations. The implicit-random-projection idea is a useful addition to the loss-engineering toolbox. However, the evidence for the central 'comparable accuracy' claim is incomplete because the comparisons are single-run point estimates, and the term 'hyperparameter-free' is overstated for MRL.

major comments (3)

[Section 4.4, Table 2; Section 5] Table 2 reports one point estimate per loss: the proposed losses are trained once, while the multiterm baselines are selected as the best of 20 Bayesian-optimization runs. The differences that support the concluding claim of 'comparable accuracy' are small (e.g., shape 3D error 2.3 mm for Coarse vs. 2.8 mm for GAL and 3.0 mm for MRL; camera rotation 3.0 degrees for Coarse vs. 3.1 degrees for GAL and 4.3 degrees for MRL), and no seed-to-seed variance, error bars, or statistical tests are reported. Because the central practical claim is that a user can train once and trust the result, the paper needs to show that these gaps are not explained by initialization or data-order stochasticity; at minimum, repeated-seed means and standard deviations should be reported for all losses.
[Section 3.4 and Section 4.6, Eq. (9)] The 'hyperparameter-free' label is not literal. MRL depends on the number of views V, and V=2 is selected empirically in Figure 6; likewise, the l1 norm in Section 3.2 is chosen 'from our experiments.' The paper should either restrict the claim to 'free of weighting hyperparameters' or report the cost of selecting V in the time comparison of Table 2, since the headline time saving of a single training run omits this selection step.
[Section 4.1 and Section 4.7] The entire evaluation rests on the quality of the Non-Rigid ICP registration used to construct the 3DMM and the ground-truth shapes, but the paper reports no validation of registration accuracy or of how much test geometry is captured by the first 100 principal components. If the reference template registration is biased for some subjects, all losses share the same corrupted shape space, and the absolute errors in Table 2 and the generalization results in Table 3 may not transfer. I ask for a quantitative registration/correspondence check and an explained-variance or reconstruction-error statement for the PCA model.

minor comments (3)

[Section 4.5 and Section 5] The text contains the typo 'posses' for 'poses' in the robustness discussion and the conclusions; please correct it.
[Section 4.6, Figure 6] Figure 6 would benefit from explicit numeric values or error bars; the text claims that shape-error variations are below a tenth of a millimeter, but the figure alone does not support that precisional claim.
[Section 4.2] The calibration matrix K is used in Eq. (6)-(8) but never specified in the experiments; please state the image resolution and K used for the internal dataset and for the external datasets.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the losses are constructed from geometry and validated on held-out and external data, not fitted to the target metrics.

full rationale

The paper's derivation chain is self-contained. The three proposed losses (GAL, SRL, MRL) are defined directly as functions of predicted and ground-truth 3D geometry, camera pose, and reprojection, with no fitted parameter later renamed as a prediction. The 3DMM is built from the training split by non-rigid ICP and PCA; evaluation uses a separate test split and external datasets (MICC and FaceWarehouse), so the comparison is not circular. The multiterm baselines are optimized with Bayesian hyperparameter search, while the proposed losses are trained once with a fixed learning rate; the reported accuracy differences are empirical outcomes, not consequences of the loss definitions. The only tuning choices, such as the l1 norm (Section 3.2) and the number of random views V = 2 (Section 4.6), are selected on the validation set and are standard model selection, not definitional circularity. There are no load-bearing self-citations: the paper does not invoke prior work by the same authors to justify a uniqueness claim or to smuggle in an ansatz. The label 'hyperparameter-free' is somewhat overstated because a learning rate and V remain, but this is a precision-of-language issue, not a circular derivation. The central claim that the proposed losses achieve comparable accuracy with less tuning is supported by held-out and external experiments rather than by construction.

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

The proposed losses introduce no new parameters beyond V and the L1 norm choice; the core of the method is a geometric reformulation of existing error terms. The load-bearing external input is the 3DMM and the known-camera model, both standard in the field.

free parameters (2)

V (number of random views in MRL) = 2
In Section 4.6 the authors vary V and conclude V=2 is sufficient; V is a hyperparameter even though the loss is called hyperparameter-free.
Distance norm in GAL, SRL, and MRL = L1
The authors state 'From our experiments, we find L1 norm to behave the best in terms of stability and accuracy' (Section 3.2), so the norm is an empirical choice rather than a theoretical necessity.

assumptions (4)

domain assumption Faces are well approximated by a 100-dimensional PCA morphable model built from internal scans.
Invoked in Section 3.1, Equation 2, where x_hat = m + Phi_id alpha_hat; the method and all baselines depend on this shape space.
domain assumption The camera is a pinhole model with known intrinsics K.
Used in Equation 8 to define projections; SRL and MRL require known K and cannot handle unknown calibration.
domain assumption Non-Rigid ICP registration of the template to scans produces dense correspondences suitable for PCA.
Described in Section 4.1; the validity of the morphable model depends on this registration, and errors would bias all evaluated losses.
standard math The isometric transform D in Equation 10 correctly maps predicted geometry into the ground-truth camera frame for random projections.
D is a rigid-body transform composition, a standard manipulation of camera extrinsics; the assumption is that pose estimates and ground-truth poses are in the same world coordinate system.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Hyperparameter-Free Losses for Model-Based Monocular Reconstruction." pith.science (2026). https://pith.science/paper/FIETUHZV

@misc{pith2026190809001,
  author       = {Pith},
  title        = {Pith review of: Hyperparameter-Free Losses for Model-Based Monocular Reconstruction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FIETUHZV}},
  note         = {Machine review of arXiv:1908.09001}
}

read the original abstract

This work proposes novel hyperparameter-free losses for single view 3D reconstruction with morphable models (3DMM). We dispense with the hyperparameters used in other works by exploiting geometry, so that the shape of the object and the camera pose are jointly optimized in a sole term expression. This simplification reduces the optimization time and its complexity. Moreover, we propose a novel implicit regularization technique based on random virtual projections that does not require additional 2D or 3D annotations. Our experiments suggest that minimizing a shape reprojection error together with the proposed implicit regularization is especially suitable for applications that require precise alignment between geometry and image spaces, such as augmented reality. We evaluate our losses on a large scale dataset with 3D ground truth and publish our implementations to facilitate reproducibility and public benchmarking in this field.

Figures

Figures reproduced from arXiv: 1908.09001 by the authors.

**Figure 2.** Schemes of the presented hyperparameter-free losses. From top to bottom: Transformations applied to the ground [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** Effect of training with SRL. While the reprojec [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Figures from the paper (4 more)

**Figure 4.** Figure 4: Camera angles distributions. Average Split # subjects # images views/subject Train 4543 20349 4.4 Validation 675 2976 4.4 Test 1310 6347 4.8 [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]

**Figure 5.** Figure 5: Shape 3D errors (top) and reprojection errors (bot [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

**Figure 6.** Figure 6: Effect of the number of views on the models [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

**Figure 7.** Figure 7: Qualitative evaluation of the shape 3D errors on [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

35 extracted references · 29 canonical work pages

[1]

A. D. Bagdanov, A. Del Bimbo, and I. Masi. The ﬂorence 2d/3d hybrid face dataset. In Proceedings of the 2011 joint ACM workshop on Human gesture and behavior understand- ing, pages 79–80. ACM, 2011. 2, 7, 8

work page 2011
[2]

Blanz and T

V . Blanz and T. Vetter. A morphable model for the synthesis of 3d faces. In Proceedings of the 26th annual conference on Computer graphics and interactive techniques , pages 187–
[3]

M. M. Bronstein, J. Bruna, Y . LeCun, A. Szlam, and P. Van- dergheynst. Geometric deep learning: going beyond eu- clidean data. IEEE Signal Processing Magazine, 34(4):18– 42, 2017. 1, 2

2017
[4]

C. Cao, Y . Weng, S. Zhou, Y . Tong, and K. Zhou. Faceware- house: A 3d facial expression database for visual computing. IEEE Transactions on Visualization and Computer Graphics, 20(3):413–425, 2014. 2, 7, 8

work page 2014
[5]

Carreira, P

J. Carreira, P. Agrawal, K. Fragkiadaki, and J. Malik. Human pose estimation with iterative error feedback. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 4733–4742, 2016. 3

work page 2016
[6]

A. X. Chang, T. Funkhouser, L. Guibas, P. Hanrahan, Q. Huang, Z. Li, S. Savarese, M. Savva, S. Song, H. Su, J. Xiao, L. Yi, and F. Yu. ShapeNet: An Information-Rich 3D Model Repository. Technical Report arXiv:1512.03012 [cs.GR], Stanford University — Princeton University — Toyota Technological Institute at Chicago, 2015. 2

arXiv 2015
[7]

C. B. Choy, D. Xu, J. Gwak, K. Chen, and S. Savarese. 3d- r2n2: A uniﬁed approach for single and multi-view 3d ob- ject reconstruction. In European Conference on Computer Vision, pages 628–644. Springer, 2016. 1, 2

work page 2016
[8]

J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei- Fei. Imagenet: A large-scale hierarchical image database. In Computer Vision and Pattern Recognition, 2009. CVPR

work page 2009

Show all 35 references

[9]

Y . Feng, F. Wu, X. Shao, Y . Wang, and X. Zhou. Joint 3d face reconstruction and dense alignment with position map regression network. In Proceedings of the European Con- ference on Computer Vision (ECCV), pages 534–551, 2018. 5

2018
[10]

Goodfellow, Y

I. Goodfellow, Y . Bengio, A. Courville, and Y . Bengio.Deep learning, volume 1. MIT press Cambridge, 2016. 1, 2

2016
[11]

Y . Guo, J. Zhang, J. Cai, B. Jiang, and J. Zheng. Cnn-based real-time dense face reconstruction with inverse-rendered photo-realistic face images. IEEE transactions on pattern analysis and machine intelligence, 2018. 5

2018
[12]

H ¨ane, S

C. H ¨ane, S. Tulsiani, and J. Malik. Hierarchical surface pre- diction for 3d object reconstruction. In 2017 International Conference on 3D Vision (3DV) , pages 412–420. IEEE,

2017
[13]

B. K. Horn. Shape from shading: A method for obtaining the shape of a smooth opaque object from one view. 1970. 1

1970
[14]

A. S. Jackson, A. Bulat, V . Argyriou, and G. Tzimiropoulos. Large pose 3d face reconstruction from a single image via di- rect volumetric cnn regression. In Computer Vision (ICCV), 2017 IEEE International Conference on , pages 1031–1039. IEEE, 2017. 2

2017
[15]

Kanazawa, M

A. Kanazawa, M. J. Black, D. W. Jacobs, and J. Malik. End- to-end recovery of human shape and pose. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 7122–7131, 2018. 3

2018
[16]

Kanazawa, S

A. Kanazawa, S. Tulsiani, A. A. Efros, and J. Malik. Learn- ing category-speciﬁc mesh reconstruction from image col- lections. In Proceedings of the European Conference on Computer Vision (ECCV), pages 371–386, 2018. 1, 2

2018
[17]

A. Kar, C. H ¨ane, and J. Malik. Learning a multi-view stereo machine. In Advances in Neural Information Processing Sys- tems, pages 364–375, 2017. 2

2017
[18]

H. Kato, Y . Ushiku, and T. Harada. Neural 3d mesh renderer. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 3907–3916, 2018. 3

2018
[19]

Kendall, R

A. Kendall, R. Cipolla, et al. Geometric loss functions for camera pose regression with deep learning. In Proc. CVPR, volume 3, page 8, 2017. 3, 4, 6

2017
[20]

Kendall, M

A. Kendall, M. Grimes, and R. Cipolla. Posenet: A convolu- tional network for real-time 6-dof camera relocalization. In Proceedings of the IEEE international conference on com- puter vision, pages 2938–2946, 2015. 6

2015
[21]

D. P. Kingma and J. Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014. 6

2014 arXiv
[22]

Krizhevsky, I

A. Krizhevsky, I. Sutskever, and G. E. Hinton. Imagenet classiﬁcation with deep convolutional neural networks. In Advances in neural information processing systems , pages 1097–1105, 2012. 2

2012
[23]

Richardson, M

E. Richardson, M. Sela, and R. Kimmel. 3d face reconstruc- tion by learning from synthetic data. In 3D Vision (3DV), 2016 Fourth International Conference on , pages 460–469. IEEE, 2016. 2, 3, 5, 6

2016
[24]

Richardson, M

E. Richardson, M. Sela, R. Or-El, and R. Kimmel. Learning detailed face reconstruction from a single image. In 2017 IEEE Conference on Computer Vision and Pattern Recogni- tion (CVPR), pages 5553–5562. IEEE, 2017. 2, 3, 5, 6, 7, 8

2017
[25]

J. Roth, Y . Tong, and X. Liu. Unconstrained 3d face recon- struction. Trans. Graph, 33(4):43, 2014. 2

2014
[26]

Simonyan and A

K. Simonyan and A. Zisserman. Very deep convolutional networks for large-scale image recognition. In International Conference on Learning Representations (ICLR), 2015. 6

2015
[27]

Tewari, M

A. Tewari, M. Zollh ¨ofer, P. Garrido, F. Bernard, H. Kim, P. P´erez, and C. Theobalt. Self-supervised multi-level face model learning for monocular reconstruction at over 250 hz. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 2549–2559...

2018
[28]

Tewari, M

A. Tewari, M. Zollh ¨ofer, H. Kim, P. Garrido, F. Bernard, P. Perez, and C. Theobalt. Mofa: Model-based deep convo- lutional face autoencoder for unsupervised monocular recon- struction. In The IEEE International Conference on Com- puter Vision (ICCV), volume 2, 2017. 2, 3, 4, 6

2017
[29]

A. T. Tran, T. Hassner, I. Masi, and G. Medioni. Regressing robust and discriminative 3d morphable models with a very deep neural network. In2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , pages 1493–1502. IEEE, 2017. 3, 8

2017
[30]

N. Wang, Y . Zhang, Z. Li, Y . Fu, W. Liu, and Y .-G. Jiang. Pixel2mesh: Generating 3d mesh models from single rgb im- ages. In Proceedings of the European Conference on Com- puter Vision (ECCV), pages 52–67, 2018. 1, 2

2018
[31]

M. J. Westoby, J. Brasington, N. F. Glasser, M. J. Hambrey, and J. Reynolds. structure-from-motionphotogrammetry: A low-cost, effective tool for geoscience applications.Geomor- phology, 179:300–314, 2012. 1

2012
[32]

X. Yan, J. Yang, E. Yumer, Y . Guo, and H. Lee. Perspective transformer nets: Learning single-view 3d object reconstruc- tion without 3d supervision. In Advances in Neural Informa- tion Processing Systems, pages 1696–1704, 2016. 1, 2

2016
[33]

X. Zhu, Z. Lei, X. Liu, H. Shi, and S. Z. Li. Face alignment across large poses: A 3d solution. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recogni- tion, pages 146–155, 2016. 3, 5

2016
[194]

ACM Press/Addison-Wesley Publishing Co., 1999. 1, 3

1999
[2009]

IEEE, 2009

IEEE Conference on, pages 248–255. IEEE, 2009. 2

2009

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

A. D. Bagdanov, A. Del Bimbo, and I. Masi. The ﬂorence 2d/3d hybrid face dataset. In Proceedings of the 2011 joint ACM workshop on Human gesture and behavior understand- ing, pages 79–80. ACM, 2011. 2, 7, 8

work page 2011

[2] [2]

Blanz and T

V . Blanz and T. Vetter. A morphable model for the synthesis of 3d faces. In Proceedings of the 26th annual conference on Computer graphics and interactive techniques , pages 187–

[3] [3]

M. M. Bronstein, J. Bruna, Y . LeCun, A. Szlam, and P. Van- dergheynst. Geometric deep learning: going beyond eu- clidean data. IEEE Signal Processing Magazine, 34(4):18– 42, 2017. 1, 2

2017

[4] [4]

C. Cao, Y . Weng, S. Zhou, Y . Tong, and K. Zhou. Faceware- house: A 3d facial expression database for visual computing. IEEE Transactions on Visualization and Computer Graphics, 20(3):413–425, 2014. 2, 7, 8

work page 2014

[5] [5]

Carreira, P

J. Carreira, P. Agrawal, K. Fragkiadaki, and J. Malik. Human pose estimation with iterative error feedback. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 4733–4742, 2016. 3

work page 2016

[6] [6]

A. X. Chang, T. Funkhouser, L. Guibas, P. Hanrahan, Q. Huang, Z. Li, S. Savarese, M. Savva, S. Song, H. Su, J. Xiao, L. Yi, and F. Yu. ShapeNet: An Information-Rich 3D Model Repository. Technical Report arXiv:1512.03012 [cs.GR], Stanford University — Princeton University — Toyota Technological Institute at Chicago, 2015. 2

arXiv 2015

[7] [7]

C. B. Choy, D. Xu, J. Gwak, K. Chen, and S. Savarese. 3d- r2n2: A uniﬁed approach for single and multi-view 3d ob- ject reconstruction. In European Conference on Computer Vision, pages 628–644. Springer, 2016. 1, 2

work page 2016

[8] [8]

J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei- Fei. Imagenet: A large-scale hierarchical image database. In Computer Vision and Pattern Recognition, 2009. CVPR

work page 2009