Pith. sign in

REVIEW 4 major objections 5 minor 2 cited by

CRISP: Object Pose and Shape Estimation with Test-Time Adaptation

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 A correct-and-certify self-training loop lets a category-agnostic RGB-D pose-and-shape model adapt to new domains using only test images.

desk verdict A clever correct-and-certify system with a genuinely useful active shape decoder; the math is sound, but the self-training certificate shares its decoder with the pseudo-labels, so the paper's central empirical premise about simplex reliability carries more weight than the authors admit. read the letter →

arxiv 2412.01052 v1 pith:4TIZ5IH3 submitted 2024-12-02 cs.CV cs.RO

Jingnan Shi , Rajat Talak , Harry Zhang , David Jin , Luca Carlone This is my paper · ORCID
classification cs.CVcs.RO
keywords objectposeestimationshapeRGB-Dcategory-agnostictest-timeadaptationself-trainingsigneddistancefieldactivemodel
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

CRISP is a category-agnostic pipeline that takes one RGB-D image of a segmented object and returns both its 6D pose and its 3D shape. The paper's central claim is that the same pipeline can be adapted to a new domain at test time without any synthetic data, by repeatedly correcting its own estimates and using the corrected estimates that pass a certificate as pseudo-labels. The enabling observation is that a neural signed-distance-field decoder trained on a set of CAD models behaves well only when the latent shape code is a convex combination of training codes, so the corrector projects onto that simplex and approximates the decoder by an active shape model. That approximation turns shape correction into a constrained linear least-squares problem, solvable quickly with interior-point methods. If the claim holds, systems facing sim-to-real gaps could self-improve from unlabelled observations rather than waiting for curated synthetic data.

What carries the argument

The load-bearing object is the active shape decoder, a linear surrogate for the trained SDF decoder over the simplex $S_K$ of latent codes: $f_a(z\mid c)=c_0d_0f_d(z\mid h)+\sum_{k=1}^{K}c_kd_kf_d(z\mid h_k)$, with $c$ constrained to the probability simplex and $d_k$ the inverse bounding-box diameters. Keeping the encoder output $h=f_e(I)$ as one basis element is what makes the approximation work; dropping it degrades self-training. Because the surrogate is linear in $c$, updating the shape code given fixed pose-normalized coordinates $Z$ becomes a constrained linear least-squares problem solvable by an interior-point method, and the matrix $F(Z)$ built from decoder evaluations also gives a degeneracy check through the eigenvalues of $F(Z)^TF(Z)$.

What would settle it

Take a trained CRISP model and a held-out set of annotated test objects. If the estimates that pass the certificate do not have systematically lower Chamfer distance than estimates that fail it, the certificate is not actually certifying correctness; alternatively, if an object whose ground-truth shape corresponds to a latent code outside the training simplex is still corrected to a low-Chamfer shape, the simplex assumption is not the active constraint.

Watch

Extended reading notes

Core claim

The central discovery is a test-time repair loop for joint pose-and-shape estimation. Given a domain gap, the trained estimates may be wrong in ways that violate the depth observation. The paper argues that the shape decoder is well-posed exactly inside the convex hull (simplex) of latent codes seen during training, and that projecting corrections onto that simplex keeps the decoder in its reliable regime. It then replaces the decoder with an active shape decoder $f_a(z\mid c)=c_0d_0f_d(z\mid h)+\sum_{k=1}^{K}c_kd_kf_d(z\mid h_k)$, where $c$ lives in the simplex and $h=f_e(I)$ is kept as a basis, so that the shape update becomes $\min_{c\ge 0,\mathbf{1}^Tc=1}\|F(Z)Dc\|^2$, a constrained linear least-squares problem. The corrector is solved by block coordinate descent with an interior-point method, and a certificate checks that corrected pose-inverted depth points lie near the zero level set of the corrected SDF. Only estimates passing the certificate become pseudo-labels for self-training, and the paper demonstrates on YCBV, SPE3R, and NOCS that this procedure improves a synthetically trained model across the sim-to-real gap and also improves shape and pose for unseen test objects.

Load-bearing premise

The whole repair loop rests on the empirical claim that the trained shape decoder produces plausible shapes exactly when the latent code is a convex combination of the training codes, and implausible shapes outside that simplex; if that boundary fails for an unseen object, the corrector's projections and every pseudo-label derived from them can be systematically wrong.

Editorial extensions

If this is right

  • A model trained on synthetic RGB-D data can be self-trained on unlabelled real test images, improving both shape and pose metrics on YCBV without any synthetic data during adaptation.
  • The corrector also helps at inference time even when no self-training is run; on SPE3R, correction lowers mean shape error and pose error for unseen satellites.
  • The category-agnostic pipeline outperforms category-level baselines in shape reconstruction on NOCS/REAL275 and achieves the best mean average precision at 3D IoU 50 and IoU 75 among compared methods.
  • Because the SDF decoder stays frozen during self-training and only the lightweight shape and PNC heads are updated, runtimes stay compatible with real-time robotic perception.

Reading between the lines

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

  • Beyond the paper, the simplex-reliability trick should transfer to any implicit decoder whose latent space is trained on a finite set of exemplars: wrap it in an active model and certify by whatever consistency check the field supplies.
  • The linear least-squares form means the corrector is fitting a convex shape prior, so the minimum eigenvalue of $F(Z)^TF(Z)$ could be exposed as an observability score for downstream planners.
  • A testable next step is replacing the simplex with a low-dimensional PCA ellipsoid, which should let the same correct-and-certify loop scale to much larger object libraries.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper presents CRISP, a category-agnostic RGB-D pipeline for simultaneous object pose and shape estimation, together with an optimization-based corrector and a self-training method (CRISP-ST) that adapts the model at test time without synthetic data. The corrector replaces the trained shape decoder with an active shape model built from training latent codes, reducing shape correction to a constrained linear least squares problem solved by an interior-point method (Alg. 2), while an alternative block coordinate descent solver (Alg. 1) is also provided. Self-training uses a correct-and-certify loop in which corrected estimates pass an observable correctness certificate before becoming pseudo-labels. Experiments are reported on YCBV, SPE3R, and NOCS/REAL275, with the paper claiming high performance on all datasets and that self-training bridges large domain gaps.

Significance. If the claims are supported, the paper would make a useful contribution: the corrector is elegant and fast (the LSQ solver is much faster than BCD), the Appendix A equivalence proof is correct, and the system shows strong shape reconstruction on NOCS/REAL275. The idea of avoiding synthetic data in self-training by combining geometric correction with a certificate is also of practical interest. However, the central empirical claims are currently undermined by evaluation-protocol and statistical issues, and the certificate's reliance on the same decoder that generates pseudo-labels raises a correctness risk for out-of-simplex shapes. The paper's strengths are its clean formulation and the machine-checkable proof in Appendix A, but the experimental evidence does not yet establish the headline claim that self-training bridges large domain gaps.

major comments (4)
  1. [Section 6.2 (SPE3R)] The paper states that CRISP-ST is self-trained for 10 epochs on the test set (Section 6.2, Setup) and is then evaluated on that same test set (Table 3). This protocol measures transductive fitting to the evaluation set, not adaptation to a previously unseen domain, so the abstract's claim that self-training is capable of bridging a large domain gap is not supported by this experiment. Please re-run the evaluation with a held-out split of the target data, or with self-training restricted to a separate unlabeled set, and report metrics on a disjoint evaluation set.
  2. [Section 5 and Algorithm 2] The certificate (Eq. 13) evaluates |fd(\hat z_i | \hat h)| using the same frozen decoder fd that produces the pseudo-labels, and Algorithm 2 computes \hat h = \hat c0 d0 h + sum_k \hat c_k d_k h_k without projecting \hat h back into the simplex S_K. Since Section 4.1 observes that fd is unreliable outside S_K, the certificate can be satisfied by a self-consistent extrapolated code with an incorrect shape. Please either project \hat h onto S_K in Algorithm 2 or provide empirical evidence that codes passing the certificate remain in S_K and that the certificate precision is correlated with shape error on a held-out set.
  3. [Appendix D, Table 8 vs Section 6.2, Table 3] On SPE3R, CRISP + Corrector alone achieves eL1_shape 0.139 and eL1_pose 0.191, while CRISP-ST achieves 0.141 and 0.224 (mean values). Thus the reported self-training gains over the corrector are negative for pose and negligible for shape on this dataset, so the contribution of the self-training loop itself is not demonstrated. Please disentangle the effect of the corrector from the effect of self-training in the main results and discuss the discrepancy.
  4. [Section 6, Tables 1-6 and Appendix D, Tables 8-13] All experimental results appear to come from a single training run, with no error bars, confidence intervals, or multiple random seeds reported. Several comparisons central to the paper are small (e.g., Table 1: CRISP-Syn-ST (LSQ) 0.037 vs CRISP-Syn-ST (BCD) 0.039 in mean eshape; Table 6: Proposed ADD-S AUC 0.42 vs No Corrector 0.30), and without variance estimates it is impossible to assess whether these differences are statistically meaningful. Please report means and standard deviations over at least three independent runs, or justify why single-run results are sufficient for the claims made.
minor comments (5)
  1. [Section 1] In the first paragraph, 'FiLM conditioning' is misspelled as 'FiLM conditiong'; please correct the typo.
  2. [Figure 4 caption] The caption says 'the handle of the mud'; this should be 'the handle of the mug'.
  3. [Section 5, Certification] The phrase 'observably correctness certificate' should be 'observable correctness certificate' for grammatical consistency.
  4. [References] References [13] and [14] are the same paper (Shap-E) and should be merged to avoid duplicate citation.
  5. [Section 4.2, Eq. (9)] The sentence 'The positive constants dk prove useful in normalizing the signed distance field' is vague; please provide the definition (inverse bounding-box diameter) in the main text rather than only in Appendix D.

Circularity Check

0 steps flagged · score 0.0 of 10

No load-bearing circularity: corrector and certificate are grounded in observed depth and validated against external ground truth.

full rationale

CRISP's derivation chain is not circular. The corrector (Eq. 5/8) minimizes an SDF-consistency objective over PNC and latent codes using the observed depth point cloud as input; it is initialized from network estimates but is not defined in terms of the quantities it is later used to predict. The active shape decoder (Eq. 9) is an explicit linear surrogate of the frozen SDF decoder, introduced as an approximation and tested against ground-truth Chamfer and ADD-S metrics; its use in Algorithm 2 is an algebraic reduction to constrained linear least squares, not an assumption of the conclusion. The certificate (Eq. 13) thresholds the same SDF residual that the corrector minimizes, but this is an observable geometric consistency check against input depth rather than a self-defined quality label, and Figure 6a shows that the check selects instances with lower ground-truth Chamfer distances, an external validation. The self-training loss (Eq. 14) trains the encoder and PNC head toward corrected pseudo-labels, and all reported gains are measured against ground-truth models (Tables 1-3), so improvements are not forced by construction. The self-citations [36,38] are acknowledged as inspiration for the correct-and-certify idea and are not used to establish correctness or to import a uniqueness theorem. The SPE3R evaluation protocol self-trains on the test set and evaluates on the same set, which is a transductive test-time adaptation setup and a limitation for generalization claims, but it is an evaluation-protocol issue, not a circular derivation.

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

Most free parameters are training and certificate hyperparameters chosen by hand; none are derived from first principles. The central assumptions are the empirical simplex reliability of the shape decoder, the existence of a good latent code for unseen objects, and the validity of the depth-consistency certificate as a pseudo-label filter.

free parameters (6)
  • Certifier threshold epsilon = 0.01 (YCBV), 0.02 (SPE3R)
    Used in the observable correctness certificate (13) to decide which corrected estimates become pseudo-labels; chosen per dataset in Appendix C.2.
  • Certifier quantile p = 0.98 (YCBV), 0.97 (SPE3R)
    Chosen per dataset for the certificate (13), controlling how many depth points must satisfy the SDF residual bound.
  • Loss hyperparameters alpha, beta, gamma1, gamma2, gamma3 = alpha=5e3 or 50, beta=0.1 or 1, gamma1=3e3, gamma2=2e2, gamma3=50
    Set by hand in Appendix C for each dataset; they weight the PNC loss and the SDF loss terms during supervised training.
  • Soft-L1 threshold zeta = 0.1
    Switching threshold in the PNC loss (20), fixed across all datasets.
  • Normalization constants dk = Inverse bounding-box diameter per training shape
    Used in (9) and the diagonal matrix D; Table 13 shows including this normalization improves self-training performance.
  • Corrector step sizes and iteration counts = Z step 1e-3, 50 iters; h step 1e-2, 25 iters; buffer 50 frames
    Manual choices for the BCD and LSQ solvers in Appendix C.2; these affect how much correction is applied and how long self-training takes.
assumptions (6)
  • standard math Arun's least-squares rigid alignment solves (4) for pose given PNC.
    Used in Section 3.2 to recover R and t from the predicted pose-normalized coordinates.
  • standard math Interior-point methods (CVXPY) solve the constrained linear least squares problem in Algorithm 2.
    Used in Section 4.2 for the active shape decoder update.
  • ad hoc to paper The trained shape decoder is reliable only when the latent shape code lies in the simplex of training codes, and unreliable outside it.
    Empirical observation in Section 4.1 that motivates both the projection step in BCD and the construction of the active shape decoder.
  • domain assumption For any test object there exists a latent shape code generating the observed signed distance field (well-trained decoder).
    Remark 2 assumes this to argue the corrector objective reaches zero at ground truth; it may fail for genuinely unseen object classes.
  • domain assumption The observable correctness certificate (13) with chosen epsilon and p is a valid proxy for correct pose and shape.
    Self-training relies on this certificate to filter pseudo-labels; it is a heuristic depth-consistency check with no formal guarantee.
  • domain assumption Ground-truth object data association is available at test time.
    Stated as a limitation in Section 7; tracking errors are not modeled.

how reviews work

0 comments
Cite this review

Pith. "Pith review of CRISP: Object Pose and Shape Estimation with Test-Time Adaptation." pith.science (2026). https://pith.science/paper/4TIZ5IH3

@misc{pith2026241201052,
  author       = {Pith},
  title        = {Pith review of: CRISP: Object Pose and Shape Estimation with Test-Time Adaptation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4TIZ5IH3}},
  note         = {Machine review of arXiv:2412.01052}
}
read the original abstract

We consider the problem of estimating object pose and shape from an RGB-D image. Our first contribution is to introduce CRISP, a category-agnostic object pose and shape estimation pipeline. The pipeline implements an encoder-decoder model for shape estimation. It uses FiLM-conditioning for implicit shape reconstruction and a DPT-based network for estimating pose-normalized points for pose estimation. As a second contribution, we propose an optimization-based pose and shape corrector that can correct estimation errors caused by a domain gap. Observing that the shape decoder is well behaved in the convex hull of known shapes, we approximate the shape decoder with an active shape model, and show that this reduces the shape correction problem to a constrained linear least squares problem, which can be solved efficiently by an interior point algorithm. Third, we introduce a self-training pipeline to perform self-supervised domain adaptation of CRISP. The self-training is based on a correct-and-certify approach, which leverages the corrector to generate pseudo-labels at test time, and uses them to self-train CRISP. We demonstrate CRISP (and the self-training) on YCBV, SPE3R, and NOCS datasets. CRISP shows high performance on all the datasets. Moreover, our self-training is capable of bridging a large domain gap. Finally, CRISP also shows an ability to generalize to unseen objects. Code and pre-trained models will be available on https://web.mit.edu/sparklab/research/crisp_object_pose_shape/.

Figures

Figures reproduced from arXiv: 2412.01052 by the authors.

Figure 1
Figure 1. We introduce CRISP, a category-agnostic object pose and shape estimation pipeline, and a test-time adaptive self-training method CRISP-ST to bridge domain gaps. Top: Qualitative ex￾amples of CRISP on the YCBV dataset [45]. Bottom: Qualitative examples of CRISP on the SPE3R dataset [27]. object 6D pose and shape estimation [39, 47], 3D bounding box estimation [2, 23], and more recently category-agnostic object pose a… view at source ↗
Figure 2
Figure 2. Overview of our contributions. Given an segmented RGB image [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Visualization of the mesh extracted from SDF produced by the shape decoder [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (10 more)
Figure 4
Figure 4. Figure 4: The minimum eigenvalue of the matrix F (Z) TF (Z) as a function of keyframes N. Each keyframe captures the mug from a different viewing angle. F (Z) is computed using the estimated PNC Z, aggregated over all keyframes till N. of the matrix F(Z) TF(Z) to determine shape…
Figure 6
Figure 6. Figure 6: (a) Cumulative distribution function of the Chamfer [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 5
Figure 5. Figure 5: Qualitative examples of CRISP on the YCBV dataset. Top: projection of transformed reconstructed mesh with our es￾timation. Bottom: reconstructed mesh. See Appendix for more examples. improve on mean eshape by 17% and 12% and eshape (AUC) (5 cm) by 22% and 11% respectiv…
Figure 7
Figure 7. Figure 7: Qualitative example of self-training improving shape [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: Diagram of our architecture for CRISP’s shape head and shape decoder [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: Diagram of our architecture for CRISP’s PNC head. YCBV Dataset For the YCBV dataset, we train two su￾pervised models: CRISP-Syn and CRISP-Real. For CRISP-Syn, we train using synthetic rendered images of the YCBV objects using BlenderProc [8]. We generate a total of 420…
Figure 10
Figure 10. Figure 10: Two sample images from our YCBV synthetic dataset. [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]
Figure 11
Figure 11. Figure 11: Cumulative distribution function of the Chamfer dis [PITH_FULL_IMAGE:figures/full_fig_p013_11.png]
Figure 12
Figure 12. Figure 12: Cumulative distribution function of the ADD-S scores [PITH_FULL_IMAGE:figures/full_fig_p013_12.png]
Figure 13
Figure 13. Figure 13: (a) Purple: Z predicted by the PNC head after self￾training if we adopt normalization for Z (equivalent to NOCS [42]), Black: ground truth CAD model. (b) Blue: ground truth trans￾formed CAD model. Black: depth point cloud. Red: CAD model transformed using estimated po…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Finding NeMO: A Geometry-Aware Representation of Template Views for Few-Shot Perception

    cs.CV 2026-02 conditional novelty 7.0 of 10

    A single network, given a few RGB views of an unseen object, builds a NeMO point-cloud representation that supports few-shot detection, segmentation, surface reconstruction, and 6DoF pose estimation without retraining.

  2. Box Pose and Shape Estimation and Domain Adaptation for Large-Scale Warehouse Automation

    cs.RO 2025-07 conditional novelty 5.0 of 10

    BOSS uses certificate-checked pseudo-labels to self-train a stereo keypoint network, improving box pose and shape estimates on real warehouse data without manual labels.

Reference graph

Works this paper leans on

48 extracted references · 43 canonical work pages · cited by 2 Pith papers

  1. [1]

    Arun, T.S

    K.S. Arun, T.S. Huang, and S.D. Blostein. Least-squares fit- ting of two 3-D point sets.IEEE Trans. Pattern Anal. Machine Intell., 9(5):698–700, 1987. 3

    work page 1987

  2. [2]

    Omni3d: A large benchmark and model for 3d object detection in the wild

    Garrick Brazil, Abhinav Kumar, Julian Straub, Nikhila Ravi, Justin Johnson, and Georgia Gkioxari. Omni3d: A large benchmark and model for 3d object detection in the wild. In IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), pages 13154–13164, 2023. 1

    work page 2023

  3. [3]

    pi-gan: Periodic implicit generative adversarial networks for 3d-aware image synthesis

    Eric R Chan, Marco Monteiro, Petr Kellnhofer, Jiajun Wu, and Gordon Wetzstein. pi-gan: Periodic implicit generative adversarial networks for 3d-aware image synthesis. In IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), pages 5799–5809, 2021. 3

    work page 2021

  4. [4]

    Satel- lite pose estimation with deep landmark regression and non- linear pose refinement

    Bo Chen, Jiewei Cao, Alvaro Parra, and Tat-Jun Chin. Satel- lite pose estimation with deep landmark regression and non- linear pose refinement. In 2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW), pages 2816–2824. IEEE, 2019. 1

    work page 2019

  5. [5]

    Learning canonical shape space for category-level 6d object pose and size estimation

    Dengsheng Chen, Jun Li, Zheng Wang, and Kai Xu. Learning canonical shape space for category-level 6d object pose and size estimation. In IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), pages 11973–11982, 2020. 8, 14

    work page 2020

  6. [6]

    Sgpa: Structure-guided prior adap- tation for category-level 6d object pose estimation

    Kai Chen and Qi Dou. Sgpa: Structure-guided prior adap- tation for category-level 6d object pose estimation. In Intl. Conf. on Computer Vision (ICCV), pages 2773–2782, 2021. 8, 14

    work page 2021

  7. [7]

    Unsupervised template-assisted point cloud shape correspondence network

    Jiacheng Deng, Jiahao Lu, and Tianzhu Zhang. Unsupervised template-assisted point cloud shape correspondence network. In IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), pages 5250–5259, 2024. 2

    work page 2024

  8. [8]

    Blenderproc: Reducing the reality gap with photorealistic rendering

    Maximilian Denninger, Martin Sundermeyer, Dominik Winkelbauer, Dmitry Olefir, Tomas Hodan, Youssef Zidan, Mohamad Elbadrawy, Markus Knauer, Harinandan Katam, and Ahsan Lodhi. Blenderproc: Reducing the reality gap with photorealistic rendering. In 16th Robotics: Science and Systems, RSS 2020, Workshops, 2020. 6, 12

    work page 2020

Show all 48 references
  1. [9]

    CVXPY: A Python- embedded modeling language for convex optimization

    Steven Diamond and Stephen Boyd. CVXPY: A Python- embedded modeling language for convex optimization. Jour- nal of Machine Learning Research, 17(83):1–5, 2016. 5

    2016

  2. [10]

    Industrial robotics

    Martin Hägele, Klas Nilsson, J Norberto Pires, and Rainer Bischoff. Industrial robotics. Springer handbook of robotics, pages 1385–1422, 2016. 1

    2016

  3. [11]

    BOP challenge 2020 on 6D object localization.Eu- ropean Conference on Computer Vision Workshops (ECCVW),

    Tomáš Hodaˇn, Martin Sundermeyer, Bertram Drost, Yann Labbé, Eric Brachmann, Frank Michel, Carsten Rother, and Jiˇrí Matas. BOP challenge 2020 on 6D object localization.Eu- ropean Conference on Computer Vision Workshops (ECCVW),

    2020

  4. [12]

    Shapo: Im- plicit representations for multi-object shape, appearance, and pose optimization

    Muhammad Zubair Irshad, Sergey Zakharov, Rares Ambrus, Thomas Kollar, Zsolt Kira, and Adrien Gaidon. Shapo: Im- plicit representations for multi-object shape, appearance, and pose optimization. In European Conf. on Computer Vision (ECCV), pages 275–292. Springer, 2022. 2

    2022

  5. [14]

    Shap-e: Generating conditional 3d implicit functions

    Heewoo Jun and Alex Nichol. Shap-e: Generating conditional 3d implicit functions. arXiv preprint arXiv:2305.02463, 2023. 2

    2023 arXiv

  6. [15]

    Labbe, J

    Y . Labbe, J. Carpentier, M. Aubry, and J. Sivic. CosyPose: Consistent multi-view multi-object 6D pose estimation. In European Conf. on Computer Vision (ECCV), 2020. 1, 2, 6, 7

    2020

  7. [16]

    Megapose: 6d pose estimation of novel objects via render & compare

    Yann Labbé, Lucas Manuelli, Arsalan Mousavian, Stephen Tyree, Stan Birchfield, Jonathan Tremblay, Justin Carpentier, Mathieu Aubry, Dieter Fox, and Josef Sivic. Megapose: 6d pose estimation of novel objects via render & compare. 2022. 1, 2

    2022

  8. [17]

    Category-level metric scale object shape and pose estimation

    Taeyeop Lee, Byeong-Uk Lee, Myungchul Kim, and In So Kweon. Category-level metric scale object shape and pose estimation. IEEE Robotics and Automation Letters , 6(4): 8575–8582, 2021. 8, 14

    2021

  9. [18]

    DeepIM: Deep Iterative Matching for 6D Pose Estimation

    Yi Li, Gu Wang, Xiangyang Ji, Yu Xiang, and Dieter Fox. DeepIM: Deep Iterative Matching for 6D Pose Estimation. In European Conf. on Computer Vision (ECCV), pages 683–698,

  10. [19]

    Dualposenet: Category-level 6d object pose and size estimation using dual pose network with refined learning of pose consistency

    Jiehong Lin, Zewei Wei, Zhihao Li, Songcen Xu, Kui Jia, and Yuanqing Li. Dualposenet: Category-level 6d object pose and size estimation using dual pose network with refined learning of pose consistency. InIntl. Conf. on Computer Vision (ICCV), pages 3560–3569, 2021. 8, 14

    2021

  11. [20]

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

    Minghua Liu, Chao Xu, Haian Jin, Linghao Chen, Mukund Varma T, Zexiang Xu, and Hao Su. One-2-3-45: Any single image to 3d mesh in 45 seconds without per-shape optimiza- tion. Advances in Neural Information Processing Systems (NIPS), 36, 2024. 1, 2

    2024

  12. [21]

    Xingyu Liu, Ruida Zhang, Chenyangguang Zhang, Bowen Fu, Jiwen Tang, Xiquan Liang, Jingyi Tang, Xiaotian Cheng, Yukang Zhang, Gu Wang, and Xiangyang Ji. Gdrnpp. https : / / github . com / shanice - l / gdrnpp_bop2022, 2022. 6, 7

    2022

  13. [22]

    Fsd: Fast self-supervised single rgb-d to categorical 3d objects

    Mayank Lunayach, Sergey Zakharov, Dian Chen, Rares Am- brus, Zsolt Kira, and Muhammad Zubair Irshad. Fsd: Fast self-supervised single rgb-d to categorical 3d objects. In 9 IEEE Intl. Conf. on Robotics and Automation (ICRA), pages 14630–14637. IEEE, 2024. 1, 2, 8, 13, 14

    2024

  14. [23]

    3d bounding box estimation using deep learn- ing and geometry

    Arsalan Mousavian, Dragomir Anguelov, John Flynn, and Jana Kosecka. 3d bounding box estimation using deep learn- ing and geometry. In IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), pages 7074–7082, 2017. 1

    2017

  15. [24]

    Stacked hour- glass networks for human pose estimation

    Alejandro Newell, Kaiyu Yang, and Jia Deng. Stacked hour- glass networks for human pose estimation. In European Conf. on Computer Vision (ECCV), pages 483–499. Springer, 2016. 2

    2016

  16. [25]

    Dinov2: Learning robust visual features without supervision

    Maxime Oquab, Timothée Darcet, Théo Moutakanni, Huy V o, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, et al. Dinov2: Learning robust visual features without supervision. arXiv preprint arXiv:2304.07193, 2023. 3, 6, 11

    2023 arXiv

  17. [26]

    J.J. Park, P. Florence, J. Straub, R. Newcombe, and S. Love- grove. DeepSDF: Learning continuous signed distance func- tions for shape representation. In IEEE Conf. on Computer Vision and Pattern Recognition (CVPR). IEEE, 2019. 2

    2019

  18. [27]

    Rapid abstraction of spacecraft 3d structure from single 2d image

    Tae Ha Park and Simone D’Amico. Rapid abstraction of spacecraft 3d structure from single 2d image. In AIAA SCITECH 2024 Forum, page 2768, 2024. 1, 6, 7

    2024

  19. [28]

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

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. Pytorch: An imperative style, high-performance deep learning library. Ad- vances in neural information processing systems, 3...

    2019

  20. [29]

    Pavlakos, X

    G. Pavlakos, X. Zhou, A. Chan, K. Derpanis, and K. Dani- ilidis. 6-dof object pose from semantic keypoints. In IEEE Intl. Conf. on Robotics and Automation (ICRA), 2017. 2

    2017

  21. [30]

    Self- supervised category-level 6d object pose estimation with deep implicit shape representation

    Wanli Peng, Jianhang Yan, Hongtao Wen, and Yi Sun. Self- supervised category-level 6d object pose estimation with deep implicit shape representation. In Nat. Conf. on Artificial Intelligence (AAAI), pages 2082–2090, 2022. 2, 8, 13, 14

    2022

  22. [31]

    FiLM: Visual Reasoning with a General Conditioning Layer

    Ethan Perez, Florian Strub, Harm de Vries, Vincent Dumoulin, and Aaron Courville. FiLM: Visual Reasoning with a General Conditioning Layer. Nat. Conf. on Artificial Intelligence (AAAI), 32(1), 2018. 2, 3, 11

    2018

  23. [32]

    Pointnet: Deep learning on point sets for 3D classification and segmentation

    Charles R Qi, Hao Su, Kaichun Mo, and Leonidas J Guibas. Pointnet: Deep learning on point sets for 3D classification and segmentation. In IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), pages 652–660, 2017. 2

    2017

  24. [33]

    Vi- sion transformers for dense prediction

    René Ranftl, Alexey Bochkovskiy, and Vladlen Koltun. Vi- sion transformers for dense prediction. In Intl. Conf. on Computer Vision (ICCV), pages 12179–12188, 2021. 2, 3, 11

    2021

  25. [34]

    The state of industrial robotics: Emerging technologies, chal- lenges, and key research directions

    Lindsay Sanneman, Christopher Fourie, Julie A Shah, et al. The state of industrial robotics: Emerging technologies, chal- lenges, and key research directions. Foundations and Trends® in Robotics, 8(3):225–306, 2021. 1

    2021

  26. [35]

    Efficient learning of label ranking by soft projections onto polyhedra

    Shai Shalev-Shwartz, Yoram Singer, Kristin P Bennett, and Emilio Parrado-Hernández. Efficient learning of label ranking by soft projections onto polyhedra. J. of Machine Learning Research, 7(7), 2006. 4

    2006

  27. [36]

    J. Shi, R. Talak, D. Maggio, and L. Carlone. A correct-and- certify approach to self-supervise object pose estimators via ensemble self-training. 2023. 2, 6

    2023

  28. [37]

    Implicit neural representa- tions with periodic activation functions

    Vincent Sitzmann, Julien Martel, Alexander Bergman, David Lindell, and Gordon Wetzstein. Implicit neural representa- tions with periodic activation functions. Advances in Neural Information Processing Systems (NIPS), 33:7462–7473. 3, 4, 11

  29. [38]

    Talak, L

    R. Talak, L. Peng, and L. Carlone. Certifiable 3D object pose estimation: Foundations, learning models, and self-training. IEEE Trans. Robotics, 39(4):2805–2824, 2023. 2, 6

    2023

  30. [39]

    Shape prior deformation for categorical 6d object pose and size estimation

    Meng Tian, Marcelo H Ang, and Gim Hee Lee. Shape prior deformation for categorical 6d object pose and size estimation. In European Conf. on Computer Vision (ECCV), pages 530–

  31. [40]

    Self6D: Self-supervised monocular 6D object pose estimation

    Gu Wang, Fabian Manhardt, Jianzhun Shao, Xiangyang Ji, Nassir Navab, and Federico Tombari. Self6D: Self-supervised monocular 6D object pose estimation. In European Conf. on Computer Vision (ECCV), pages 108–125, 2020. 2

    2020

  32. [41]

    GDR-Net: Geometry-Guided Direct Regression Network for Monocular 6D Object Pose Estimation

    Gu Wang, Fabian Manhardt, Federico Tombari, and Xi- angyang Ji. GDR-Net: Geometry-Guided Direct Regression Network for Monocular 6D Object Pose Estimation. In IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), pages 16611–16621, 2021. 2, 6

    2021

  33. [42]

    H. Wang, S. Sridhar, J. Huang, J. Valentin, S. Song, and L. Guibas. Normalized object coordinate space for category- level 6d object pose and size estimation. In IEEE Conf. on Computer Vision and Pattern Recognition (CVPR) , pages 2642–2651, 2019. 1, 2, 3, 6, 8, 11, 13, 14

    2019

  34. [43]

    PRNet: Self-Supervised Learning for Partial-to-Partial Registration

    Yue Wang and Justin M Solomon. PRNet: Self-Supervised Learning for Partial-to-Partial Registration. In Advances in Neural Information Processing Systems (NIPS), pages 8812– 8824, 2019. 2

    2019

  35. [44]

    Bundlesdf: Neural 6-dof tracking and 3d recon- struction of unknown objects

    Bowen Wen, Jonathan Tremblay, Valts Blukis, Stephen Tyree, Thomas Müller, Alex Evans, Dieter Fox, Jan Kautz, and Stan Birchfield. Bundlesdf: Neural 6-dof tracking and 3d recon- struction of unknown objects. In IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), pages...

    2023

  36. [45]

    PoseCNN: A convolutional neural network for 6D object pose estimation in cluttered scenes

    Yu Xiang, Tanner Schmidt, Venkatraman Narayanan, and Dieter Fox. PoseCNN: A convolutional neural network for 6D object pose estimation in cluttered scenes. In Robotics: Science and Systems (RSS), 2018. 1, 2, 6

    2018

  37. [46]

    Beyond model adaptation at test time: A survey

    Zehao Xiao and Cees GM Snoek. Beyond model adaptation at test time: A survey. arXiv preprint arXiv:2411.03687, 2024. 2

    2024 arXiv

  38. [47]

    Category-level 6d object pose estimation in the wild: A semi-supervised learning approach and a new dataset

    Yanjie Ze and Xiaolong Wang. Category-level 6d object pose estimation in the wild: A semi-supervised learning approach and a new dataset. Advances in Neural Information Process- ing Systems (NIPS), 35:27469–27483, 2022. 1, 2, 8, 14

    2022

  39. [48]

    EgoPressure: A dataset for hand pressure and pose estimation in egocentric vision

    Yiming Zhao, Taein Kwon, Paul Streli, Marc Pollefeys, and Christian Holz. EgoPressure: A dataset for hand pressure and pose estimation in egocentric vision. In ArXiv Preprint: 2409.02224, 2024. 1 10 A. Proof of Equivalence We show that the two optimization problems (5) and (7)...

    2024 arXiv

  40. [546]

    1, 2, 8, 14

    Springer, 2020. 1, 2, 8, 14

    2020

Pith tools

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