Pith. sign in

REVIEW 3 major objections 6 minor 43 references

CLERF: Contrastive LEaRning for Full Range Head Pose Estimation

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read CLERF shows that contrastive learning with synthetic GAN images gives head pose estimation a true full-range capability, matching state-of-the-art on standard benchmarks and beating full-range baselines by more than 10 degrees on heavily…

desk verdict The contrastive-learning recipe is interesting, but the 'true full range' claim doesn't survive the paper's own augmentation geometry. read the letter →

arxiv 2412.02066 v1 pith:4SNCAYQU submitted 2024-12-03 cs.CV

Ting-Ruen Wei , Haowei Liu , Huei-Chung Hu , Xuyang Wu , Yi Fang , Hsin-Tai Wu This is my paper · ORCID
classification cs.CV
keywords headposeestimationcontrastivelearningfullrange3D-awareGANgeometricaugmentationrotationmatrixCircleLosssyntheticdata
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 argues that contrastive learning, previously thought impractical for head pose estimation because matching head poses are vanishingly rare, becomes possible when a 3D-aware generative model supplies synthetic images. The authors propose CLERF, which generates an anchor image from a GAN with the same yaw and pitch as a real image, rotates it to match roll, and uses geometric transformations that preserve geodesic distance to build valid training triplets. They claim CLERF is the first true full-range head pose model, handling upside-down and back-of-head poses, and that it matches state-of-the-art accuracy on standard AFLW2000 and BIWI benchmarks while beating all baselines on slightly rotated and flipped test images and full-range baselines by more than 10 degrees mean absolute error on heavily rotated images. The practical point is that existing models are brittle to small test-time rotations, and a representation learned by contrasting nearby versus distant head orientations fixes that brittleness without sacrificing standard accuracy.

What carries the argument

The carrying mechanism is the guaranteed anchor-positive pair, assembled from a 3D-aware GAN (PanoHead) and the rotation decomposition of [15]. The paper solves for the (yaw, pitch, roll) triad of the anchor's rotation matrix, asks PanoHead to generate an image with the same yaw and pitch, then applies the roll rotation to the synthetic image so its orientation matches the anchor exactly. The geometric transformations used for augmentation are rotations in the image plane and flips across a line; Theorem 3.1 proves both preserve the geodesic distance $d(A,B)=\cos^{-1}((\operatorname{tr}(AB^T)-1)/2)$, which guarantees an augmented triplet remains a valid anchor-positive-negative set. Contrastive training uses Circle loss with geodesic-similarity thresholds to admit nearby poses as additional positives and hard negatives by embedding distance.

What would settle it

Generate a set of PanoHead images at known left-right and up-down angles, estimate their actual 3D head orientation with an independent pose estimator whose ground truth is known, and compare requested versus estimated orientation; if the mean pose error is large or correlated with pose, CLERF's positives are invalid. A companion ablation that trains CLERF with the GAN positives replaced by random-image positives would show how much of the contrastive gain depends on the same-pose guarantee.

Watch

Extended reading notes

Core claim

CLERF's central discovery is that head pose can be learned as a contrastive geometry problem rather than a direct regression. For every real anchor image, the method generates a PanoHead synthetic image at the same yaw and pitch, then applies a roll rotation so the synthetic positive's rotation matrix exactly equals the anchor's. Since same-pose pairs are guaranteed, the sparsity obstacle disappears; geometric rotations and flips preserve the geodesic distance between any two rotation matrices, so augmented images remain valid triplets. Training a Swin Transformer with Circle loss separates neighboring head orientations from distant ones, and a frozen representation feeds an MLP that outputs a 6D rotation matrix projected to SO(3) by Gram-Schmidt. The claimed result is on-par accuracy on AFLW2000 and BIWI, superior robustness on slightly rotated and flipped versions, and a more than 10 degree mean MAE advantage over existing full-yaw-range models on heavily rotated and flipped test images.

Load-bearing premise

The load-bearing premise is that the generative model actually makes a head facing exactly the requested left-right and up-down direction; the paper relies on this to build matching pairs but never measures how accurate the generated head is.

Editorial extensions

If this is right

  • If CLERF's central claim holds, head pose estimation no longer needs a training set that densely samples every orientation: synthetic positives plus geometric augmentation can cover the full range.
  • Slight rotations or flips of test images, which sharply degrade non-full-range models, become a non-issue for a contrastively trained full-range representation.
  • The geodesic-preservation result means any rotation or flip augmentation can be applied jointly to anchor-positive-negative triplets without breaking the contrastive objective.
  • A frozen contrastive representation plus a small MLP head can match or beat end-to-end supervised models, so pose-specific head design can be decoupled from representation learning.

Reading between the lines

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

  • Editorial inference: if PanoHead's pose control is as accurate as the method requires, the same anchored-positive recipe should transfer to other sparse-pose tasks such as gaze or body-pose estimation, where same-pose pairs are also rare.
  • Editorial inference: the paper leaves the GAN pose error unmeasured; a direct measurement would distinguish true pose-geometry learning from learning GAN-specific artifacts.
  • Editorial inference: because the gains on SA and FA sets come from geometric augmentations, a natural extension is to sample rotations uniformly over SO(3) and verify that test-time robustness scales with spherical coverage.
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

3 major / 6 minor

Summary. The paper proposes CLERF, a contrastive learning framework for head pose estimation. It uses PanoHead, a 3D-aware GAN, to generate synthetic images with the same yaw and pitch as real training images, then applies in-plane rotation to match roll, creating anchor-positive pairs. Geometric transformations (in-plane rotations and flips) are used to augment the data and, the authors claim, to achieve full-range coverage. A Swin Transformer encoder is trained with Circle Loss and a downstream MLP regresses the rotation matrix. Experiments on AFLW2000 and BIWI and their slightly/full-augmented variants are reported, along with ablations.

Significance. If the full-range claim were supported, the paper would make a notable contribution: it is the first to use 3D-aware GANs for contrastive positive mining in HPE, and the empirical gains on slightly augmented and front-hemisphere full-augmented tests are meaningful. The ablation showing consistent gains from contrastive learning and geometric augmentations is a strength, as is the explicit proof of Theorem 3.1. However, the central claim of 'true full range' is not backed by the experimental protocol, as detailed in the major comments.

major comments (3)
  1. [Section 3.2, Eqs. (1)-(2), Figure 2] The claim that the geometric transformations enable 'full range' coverage is not supported by the paper's own analysis. Both the rotation in Eq. (1) and the flip in Eq. (2) preserve the sign of the third-row/third-column z-component of the rotation matrix. Starting from 300W-LP, which contains only front-facing heads, every augmented image still has a positive z-component; that is, the head remains in the front hemisphere. Figure 2(c) indeed shows the augmented distribution is a semi-sphere, not a sphere. Consequently, back-of-head poses (yaw near ±180°, z-component negative) are never in the training set nor in the FA test sets described in Section 4.1. The FA results in Table 1 therefore do not evaluate the claimed full range, and the statement in the Abstract that CLERF predicts 'any head pose' is an overstatement that is internally inconsistent with the definition of full range given in Section 1.
  2. [Section 3.1, Figure 6] The anchor-positive construction relies on the assumption that PanoHead can synthesize an image whose actual 3D head orientation matches the requested yaw and pitch. No measurement or validation of this pose accuracy is provided anywhere in the paper. If the GAN's pose control is systematically biased or noisy, the anchor-positive pairs are not genuine same-pose pairs, so the contrastive learning objective may not learn the intended pose-equivariant representation. The authors should either quantify the pose error of PanoHead outputs (e.g., by running a trained HPE model on generated images) or provide a reasoned argument for why this assumption is safe.
  3. [Section 5.1, Table 1] The claim that CLERF outperforms 'other existing full-yaw range models at the full range capability' is not established, because the FA test sets only cover the front hemisphere (as argued in major comment 1). To support this claim, the authors need an evaluation protocol that includes back-of-head poses, for example by generating test images with PanoHead across the full yaw range or by collecting/annotating real images with yaw near ±180°.
minor comments (6)
  1. [Abstract] The phrase 'currently available FR models are not in our standard full range' is confusing; consider rephrasing to make clear that existing FR models cannot handle certain orientations.
  2. [Section 3.2] 'Let H be a image geometric transformation function' should be 'Let H be an image geometric transformation function', and 'Bpresent' should be 'B present'.
  3. [Section 5.1] The phrase 'An interestingly observation' should be 'An interesting observation'.
  4. [Section 4.1] The specific angles for the SA versions (10 degrees and 85 degrees) are given without justification; adding a sentence on why these values were chosen would improve reproducibility.
  5. [Appendix A] The notation 'tr(A, B)' in the last paragraph is ambiguous; it should be 'tr(A × B^T)' or the trace of the product should be written explicitly.
  6. [Section 4.2] The hyperparameters TGD and v are not described in detail; for instance, a brief explanation of how TGD is computed from rotation matrices would help readers reproduce the triplet sampling.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the contrastive-learning derivation is self-contained and the central claims are independently supported by standard AFLW2000 and BIWI benchmarks. The 'true full range' wording overstates the semi-sphere coverage admitted in Figure 2, but that is a correctness/scope concern, not a circular derivation.

full rationale

The paper's core mechanism is not circular. Anchor-positive pairs are constructed by generating a PanoHead image with the same yaw and pitch as a real image and then rotating it to match the roll (Section 3.1, Corollary 3.1.1). This uses the label-based rotation matrix to define the positive pair; it does not fit a parameter and then rename that fit as a prediction. The downstream MLP is trained with geodesic loss on the same labels, and the representation model is evaluated on external AFLW2000 and BIWI benchmarks where it performs on par with SOTA (Table 1). The self-citation to Hu et al. [15] for Euler decomposition and flip/rotation formulas is load-bearing for the math, but the formulas are standard, parameter-free linear algebra, and the paper independently proves the key geodesic-distance invariance in Appendix A. Thus the citation is real evidence and does not raise the circularity score. The main caveat is not circularity but overclaiming: the paper itself states in Figure 2 that after rotation and flipping augmentation 'the resulting distribution does not fully match the uniform distribution (a semi-sphere versus a sphere)', because the transformations preserve the sign of the head's z-axis. Consequently, back-of-head poses with yaw near ±180° are never trained or tested, so the abstract's 'true full range ... any head pose' language exceeds what the augmentation and FA evaluation actually cover. This is an internal-consistency/overstatement issue, not a case where a result reduces to its own input by construction. The standard benchmark results, the supervised-vs-contrastive ablation (Table 2), and the augmentation ablation (Table 3) provide independent empirical content for the main claims.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central claim rests on standard rotation-matrix math, the authors' own [15] convention, PanoHead pose fidelity, and the validity of the self-constructed full-range test sets. No new physical entities are introduced.

free parameters (6)
  • TGD = 0.8
    Geodesic similarity threshold for including additional anchor-positives; chosen by hand; affects which pairs are treated as positives during contrastive training (Section 3.3, Section 4.2).
  • v = 0.1
    Margin for hard/semi-hard negative mining based on Euclidean distance between embeddings; chosen by hand (Section 3.3, Section 4.2).
  • m = 0.4
    Relaxation margin in Circle Loss; chosen by hand (Section 4.2).
  • gamma = 80
    Scaling factor in Circle Loss; chosen by hand (Section 4.2).
  • augmentation_probabilities = 0.5 rotate, 0.3 flip
    Probabilities of applying rotation and flip augmentations during training; chosen by hand (Section 4.2).
  • PanoHead_pose_range = yaw [-3.14, 3], pitch [-1.5, 0.1]
    Sampling range for generated synthetic images; a design choice that defines the pose coverage of generated positives (Section 3.1).
assumptions (4)
  • standard math Geodesic distance on SO(3) is invariant under the rotation and flip transformations defined in Eq. 1 and Eq. 2 (Theorem 3.1, Appendix A).
    Used to guarantee that augmented triplets remain valid; proof uses trace cyclicity; mathematically sound assuming Eq. 1-2 match the intended image transformations.
  • domain assumption The Euler decomposition R = Rroll * (Rpitch * Ryaw) and the extrinsic rotation formulas from Hu et al. [15] correctly describe the 300W-LP reference system and the effect of image rotation/flip on pose labels.
    The paper cites its own prior work [15] for this math and applies it to both training positives and the SA/FA test-label construction; no external validation is provided.
  • domain assumption PanoHead-generated images accurately reflect the requested yaw and pitch with negligible error.
    The anchor-positive guarantee in Section 3.1 relies on this; the paper does not quantify GAN pose accuracy.
  • domain assumption Rotated and flipped versions of frontal test images are a valid proxy for natural full-range head poses.
    The FA test sets used to support the 'true full range' claim are constructed by the authors from AFLW2000 and BIWI; real upside-down heads may differ in appearance (hair, occlusion, lighting).

how reviews work

0 comments
Cite this review

Pith. "Pith review of CLERF: Contrastive LEaRning for Full Range Head Pose Estimation." pith.science (2026). https://pith.science/paper/4SNCAYQU

@misc{pith2026241202066,
  author       = {Pith},
  title        = {Pith review of: CLERF: Contrastive LEaRning for Full Range Head Pose Estimation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4SNCAYQU}},
  note         = {Machine review of arXiv:2412.02066}
}
read the original abstract

We introduce a novel framework for representation learning in head pose estimation (HPE). Previously such a scheme was difficult due to head pose data sparsity, making triplet sampling infeasible. Recent progress in 3D generative adversarial networks (3D-aware GAN) has opened the door for easily sampling triplets (anchor, positive, negative). We perform contrastive learning on extensively augmented data including geometric transformations and demonstrate that contrastive learning allows networks to learn genuine features that contribute to accurate HPE. On the other hand, we observe that existing HPE works struggle to predict head poses as accurately when test image rotation matrices are slightly out of the training dataset distribution. Experiments show that our methodology performs on par with state-of-the-art models on standard test datasets and outperforms them when images are slightly rotated/ flipped or full range head pose. To the best of our knowledge, we are the first to deliver a true full range HPE model capable of accurately predicting any head pose including upside-down pose. Furthermore, we compared with other existing full-yaw range models and demonstrated superior results.

Figures

Figures reproduced from arXiv: 2412.02066 by the authors.

Figure 1
Figure 1. Proposed method for contrastive learning in full range head pose representation, with a [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Visualization of 300W-LP dataset [43] after randomized rotation and flipping augmentations. A point on the sphere ∈ R 3 is formed by projecting a rotation matrix to the 3D sphere, which multiplies unit vectors in (a) x, (b) y, and (c) z coordinate axes. The top row shows random rotation matrices (in green) along with those in the 300W-LP dataset (in blue), and the bottom row shows the augmented 300W-LP dataset (in r… view at source ↗
Figure 3
Figure 3. Change of model performance across different choices in (a) batch size, (b) epochs, and (c) [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Head pose test predictions of (d) CLERF and (a) [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: 3D TSNE plot of the embedding vectors of a video across (a) [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: The first row shows the rotation matrices of 300W-LP dataset images. The second row [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 35 canonical work pages

  1. [15]

    Mathematical Foundation and Corrections for Full Range Head Pose Estimation

    Huei-Chung Hu, Xuyang Wu, Yuan Wang, Yi Fang, and Hsin-Tai Wu. Mathematical foundation and corrections for full range head pose estimation. arXiv preprint arXiv:2403.18104, 2024

    work page Pith review arXiv 2024

  2. [1]

    Panohead: Geometry-aware 3d full-head synthesis in 360deg

    Sizhe An, Hongyi Xu, Yichun Shi, Guoxian Song, Umit Y Ogras, and Linjie Luo. Panohead: Geometry-aware 3d full-head synthesis in 360deg. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 20950–20959, 2023

    work page 2023

  3. [2]

    Belhumeur, David W

    Peter N. Belhumeur, David W. Jacobs, David J. Kriegman, and Neeraj Kumar. Localizing parts of faces using a consensus of exemplars. In The 24th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2011, Colorado Springs, CO, USA, 20-25 June 2011, pages 545–552. IEEE Computer Society, 2011

    work page 2011

  4. [3]

    Blanz and T

    V . Blanz and T. Vetter. Face recognition based on fitting a 3d morphable model. IEEE Transactions on Pattern Analysis and Machine Intelligence, 25(9):1063–1074, 2003

    work page 2003

  5. [4]

    A vector-based representation to enhance head pose estimation, 2020

    Zhiwen Cao, Zongcheng Chu, Dongfang Liu, and Yingjie Chen. A vector-based representation to enhance head pose estimation, 2020

    work page 2020

  6. [5]

    A simple framework for contrastive learning of visual representations

    Ting Chen, Simon Kornblith, Mohammad Norouzi, and Geoffrey Hinton. A simple framework for contrastive learning of visual representations. In International conference on machine learning, pages 1597–1607. PMLR, 2020

    2020

  7. [6]

    On the represen- tation and methodology for wide and short range head pose estimation

    Alejandro Cobo, Roberto Valle, José M Buenaposada, and Luis Baumela. On the represen- tation and methodology for wide and short range head pose estimation. Pattern Recognition, 149:110263, 2024

    work page 2024

  8. [7]

    Rankpose: Learning generalised feature with rank supervision for head pose estimation

    Donggen Dai, Wangkit Wong, and Zhuojun Chen. Rankpose: Learning generalised feature with rank supervision for head pose estimation. arXiv preprint arXiv:2005.10984, 2020

    arXiv 2005

Show all 43 references
  1. [8]

    Lwposr: Lightweight efficient fine grained head pose estimation

    Naina Dhingra. Lwposr: Lightweight efficient fine grained head pose estimation. InProceedings of the IEEE/CVF Winter Conference on applications of computer vision , pages 1495–1505, 2022

    2022

  2. [9]

    Random forests for real time 3d face analysis

    Gabriele Fanelli, Matthias Dantone, Juergen Gall, Andrea Fossati, and Luc Van Gool. Random forests for real time 3d face analysis. International journal of computer vision, 101:437–458, 2013

    2013

  3. [10]

    Towards fast, accurate and stable 3d dense face alignment

    Jianzhu Guo, Xiangyu Zhu, Yang Yang, Fan Yang, Zhen Lei, and Stan Z Li. Towards fast, accurate and stable 3d dense face alignment. In Proceedings of the European Conference on Computer Vision (ECCV), 2020

    2020

  4. [11]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016. 10

    2016

  5. [12]

    Toward robust and uncon- strained full range of rotation head pose estimation

    Thorsten Hempel, Ahmed A Abdelrahman, and Ayoub Al-Hamadi. Toward robust and uncon- strained full range of rotation head pose estimation. IEEE Transactions on Image Processing, 33:2377–2387, 2024

    2024

  6. [13]

    Temporal representation learning on monocular videos for 3d human pose estimation

    Sina Honari, Victor Constantin, Helge Rhodin, Mathieu Salzmann, and Pascal Fua. Temporal representation learning on monocular videos for 3d human pose estimation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(5):6415–6427, 2022

    2022

  7. [14]

    Quatnet: Quaternion-based head pose estimation with multiregression loss

    Heng-Wei Hsu, Tung-Yu Wu, Sheng Wan, Wing Hung Wong, and Chen-Yi Lee. Quatnet: Quaternion-based head pose estimation with multiregression loss. IEEE Transactions on Multimedia, 21(4):1035–1046, 2018

    2018

  8. [16]

    Improving head pose estimation using two-stage ensembles with top-k regression

    Bin Huang, Renwen Chen, Wang Xu, and Qinbang Zhou. Improving head pose estimation using two-stage ensembles with top-k regression. Image and Vision Computing, 93:103827, 2020

    2020

  9. [17]

    Contrastive representation learning for gaze estimation

    Swati Jindal and Roberto Manduchi. Contrastive representation learning for gaze estimation. In Annual Conference on Neural Information Processing Systems, pages 37–49. PMLR, 2023

    2023

  10. [18]

    Panoptic studio: A massively multiview system for social interaction capture

    Hanbyul Joo, Tomas Simon, Xulong Li, Hao Liu, Lei Tan, Lin Gui, Sean Banerjee, Timo- thy Scott Godisart, Bart Nabbe, Iain Matthews, Takeo Kanade, Shohei Nobuhara, and Yaser Sheikh. Panoptic studio: A massively multiview system for social interaction capture. IEEE Transactions ...

    2017

  11. [19]

    Davis E. King. Dlib-ml: A machine learning toolkit. Journal of Machine Learning Research, 10:1755–1758, 2009

    2009

  12. [20]

    Domain adaptation for head pose estimation using relative pose consistency

    Felix Kuhnke and Jörn Ostermann. Domain adaptation for head pose estimation using relative pose consistency. IEEE Transactions on Biometrics, Behavior, and Identity Science, 2023

    2023

  13. [21]

    Pose- disentangled contrastive learning for self-supervised facial representation

    Yuanyuan Liu, Wenbin Wang, Yibing Zhan, Shaoze Feng, Kejun Liu, and Zhe Chen. Pose- disentangled contrastive learning for self-supervised facial representation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 9717–9728, 2023

    2023

  14. [22]

    Swin transformer v2: Scaling up capacity and resolution

    Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, et al. Swin transformer v2: Scaling up capacity and resolution. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 12009–12019, 2022

    2022

  15. [23]

    Xm2vtsdb: The extended m2vts database

    Kieron Messer, Jiri Matas, Josef Kittler, Juergen Luettin, and Gilbert Maître. Xm2vtsdb: The extended m2vts database. 1999

    1999

  16. [24]

    Head pose estimation in computer vision: A survey

    Erik Murphy-Chutorian and Mohan Trivedi. Head pose estimation in computer vision: A survey. IEEE transactions on pattern analysis and machine intelligence, 31:607–26, 05 2009

    2009

  17. [25]

    Belongie, and Ser-Nam Lim

    Kevin Musgrave, Serge J. Belongie, and Ser-Nam Lim. Pytorch metric learning. ArXiv, abs/2008.09164, 2020

    2008 arXiv

  18. [26]

    Nataniel Ruiz, Eunji Chong, and James M. Rehg. Fine-grained head pose estimation without keypoints. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, June 2018

    2018

  19. [27]

    300 faces in-the-wild challenge: The first facial landmark localization challenge

    Christos Sagonas, Georgios Tzimiropoulos, Stefanos Zafeiriou, and Maja Pantic. 300 faces in-the-wild challenge: The first facial landmark localization challenge. In 2013 IEEE Interna- tional Conference on Computer Vision Workshops, ICCV Workshops 2013, Sydney, Australia, Decem...

    2013

  20. [28]

    Pose-based con- trastive learning for domain agnostic activity representations

    David Schneider, Saquib Sarfraz, Alina Roitberg, and Rainer Stiefelhagen. Pose-based con- trastive learning for domain agnostic activity representations. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3433–3443, 2022. 11

    2022

  21. [29]

    Self-supervised 3d hand pose estimation from monocular rgb via contrastive learning

    Adrian Spurr, Aneesh Dahiya, Xi Wang, Xucong Zhang, and Otmar Hilliges. Self-supervised 3d hand pose estimation from monocular rgb via contrastive learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 11230–11239, October 2021

    2021

  22. [30]

    Circle loss: A unified perspective of pair similarity optimization

    Yifan Sun, Changmao Cheng, Yuhan Zhang, Chi Zhang, Liang Zheng, Zhongdao Wang, and Yichen Wei. Circle loss: A unified perspective of pair similarity optimization. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 6398–6407, 2020

    2020

  23. [31]

    Efficientnet: Rethinking model scaling for convolutional neural networks

    Mingxing Tan and Quoc Le. Efficientnet: Rethinking model scaling for convolutional neural networks. In International conference on machine learning, pages 6105–6114. PMLR, 2019

    2019

  24. [32]

    Contrastive regression for domain adaptation on gaze estimation

    Yaoming Wang, Yangzhou Jiang, Jin Li, Bingbing Ni, Wenrui Dai, Chenglin Li, Hongkai Xiong, and Teng Li. Contrastive regression for domain adaptation on gaze estimation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 19376–19385, 2022

    2022

  25. [33]

    Charts on so(3) — Wikipedia, the free encyclope- dia

    Wikipedia contributors. Charts on so(3) — Wikipedia, the free encyclope- dia. https://en.wikipedia.org/w/index.php?title=Charts_on_SO(3)&oldid= 1099134153, 2022. [Online; accessed 21-May-2024]

    2022

  26. [34]

    Euler angles — Wikipedia, the free encyclopedia

    Wikipedia contributors. Euler angles — Wikipedia, the free encyclopedia. https://en. wikipedia.org/w/index.php?title=Euler_angles&oldid=1223812775, 2024. [On- line; accessed 20-May-2024]

    2024

  27. [35]

    Inter-image contrastive consistency for multi-person pose estimation

    Xixia Xu, Yingguo Gao, Xingjia Pan, Ke Yan, Xiaoyu Chen, and Qi Zou. Inter-image contrastive consistency for multi-person pose estimation. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, pages 3063–3071, 2023

    2023

  28. [36]

    Fsa-net: Learning fine-grained structure aggregation for head pose estimation from a single image

    Tsun-Yi Yang, Yi-Ting Chen, Yen-Yu Lin, and Yung-Yu Chuang. Fsa-net: Learning fine-grained structure aggregation for head pose estimation from a single image. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 1087–1096, 2019

    2019

  29. [37]

    Tokenhpe: Learning orientation tokens for efficient head pose estimation via transformers

    Cheng Zhang, Hai Liu, Yongjian Deng, Bochen Xie, and Youfu Li. Tokenhpe: Learning orientation tokens for efficient head pose estimation via transformers. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 8897–8906, 2023

    2023

  30. [38]

    Extensive facial landmark localization with coarse-to-fine convolutional network cascade

    Erjin Zhou, Haoqiang Fan, Zhimin Cao, Yuning Jiang, and Qi Yin. Extensive facial landmark localization with coarse-to-fine convolutional network cascade. In2013 IEEE International Con- ference on Computer Vision Workshops, ICCV Workshops 2013, Sydney, Australia, December 1-8, ...

    2013

  31. [39]

    Semi-supervised unconstrained head pose estimation in the wild

    Huayi Zhou, Fei Jiang, and Hongtao Lu. Semi-supervised unconstrained head pose estimation in the wild. arXiv preprint arXiv:2404.02544, 2024

    2024

  32. [40]

    Whenet: Real-time fine-grained estimation for wide range head pose, 2020

    Yijun Zhou and James Gregson. Whenet: Real-time fine-grained estimation for wide range head pose, 2020

    2020

  33. [41]

    Face detection, pose estimation, and landmark localization in the wild

    Xiangxin Zhu and Deva Ramanan. Face detection, pose estimation, and landmark localization in the wild. In 2012 IEEE Conference on Computer Vision and Pattern Recognition, Providence, RI, USA, June 16-21, 2012, pages 2879–2886. IEEE Computer Society, 2012

    2012

  34. [42]

    Xiangyu Zhu, Zhen Lei, Xiaoming Liu, Hailin Shi, and Stan Z. Li. Face alignment across large poses: A 3d solution. In 2016 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016, Las Vegas, NV , USA, June 27-30, 2016, pages 146–155. IEEE Computer Society, 2016

    2016

  35. [43]

    Face alignment in full pose range: A 3d total solution

    Xiangyu Zhu, Xiaoming Liu, Zhen Lei, and Stan Z Li. Face alignment in full pose range: A 3d total solution. IEEE transactions on pattern analysis and machine intelligence, 2017. 12 A Proof of Theorem 3.1 Proof. Suppose A and B are rotation matrices ∈ SO(3) and θ ≤ 90◦. Accordi...

    2017

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.