Pith. sign in

REVIEW 4 major objections 4 minor 49 references

Corr2Distrib: Making Ambiguous Correspondences an Ally to Predict Reliable 6D Pose Distributions

T0 review · 4 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read Treating ambiguous 2D-3D matches as votes recovers the full set of valid 6D poses from a single RGB image.

desk verdict A genuinely new correspondence-based pose-distribution method with a plausible but unproven symmetry-learning assumption; worth refereeing, but the SOTA claim is softer than the abstract suggests. read the letter →

arxiv 2505.02501 v1 pith:EP52O7MQ submitted 2025-05-05 cs.CV cs.AIcs.LGcs.RO

Asma Brazi , Boris Meden , Fabrice Mayran de Chamisso , Steve Bourgeois , Vincent Lepetit This is my paper · ORCID
classification cs.CVcs.AIcs.LGcs.RO
keywords 6Dobjectposeestimationdistributionsymmetry-awarerepresentation2D-3DcorrespondenceslocalframelearningBOP-DistribRGBambiguity
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 introduces Corr2Distrib, a method that estimates a distribution of valid 6D object poses from a single RGB image instead of committing to one pose. It builds on 2D-3D correspondences, the approach family that currently dominates single-pose benchmarks, but treats ambiguous matches as evidence rather than noise: each ambiguous correspondence votes for one rotation hypothesis. The paper argues this is the first correspondence-based method for pose distributions and reports state-of-the-art results on the BOP-Distrib benchmark for both distribution quality and best single pose. A sympathetic reader would care because the method turns a known failure mode of geometric pose solvers, symmetry-induced ambiguity, into the signal that recovers the full set of valid poses.

What carries the argument

The load-bearing object is the symmetry-aware object model $\mathcal{P}$: a dense point cloud on the object surface where each 3D point $X$ carries a descriptor $\mathcal{P}_d(X)$ and a local frame $\mathcal{P}_f(X)$. This representation is learned jointly from RGB crops and the 3D model using an InfoNCE contrastive loss for descriptors and a rotation-distance loss for local frames, with no explicit symmetry annotations. At inference the descriptor field produces many-to-many 2D-3D matches; the key identity is Eq. (5), which turns one match into one rotation hypothesis by composing the inferred camera-to-local-frame rotation with the model's local-frame-to-object rotation. The rest of the pipeline filters these hypotheses by density on an equi-volumetric SO(3) grid, groups correspondences whose rotations fall in the same bin, and refines each group with PnP-RANSAC plus a descriptor-similarity and mask-agreement score.

What would settle it

Compare the learned descriptors of pairs of 3D points related by an object's known symmetry with descriptors of unrelated pairs on a held-out symmetric object. If symmetry-related points are not systematically more similar, the implicit symmetry-awareness assumption is false and the rotation hypotheses from ambiguous matches lose their validity; equivalently, run the method on an object with a known discrete symmetry and check that all symmetry-equivalent ground-truth rotations receive comparable probability.

Watch

Extended reading notes

Core claim

The central claim is that a symmetry-aware surface representation, learned without explicit symmetry labels, lets a single 2D-3D correspondence produce a rotation hypothesis, and that aggregating these hypotheses across an image recovers the full pose distribution. For each surface point the method learns a descriptor and a local frame, with the property that points related by a symmetry share descriptors and have frames linked by the symmetry transformation. At inference a pixel can match several surface points, and each match yields a camera rotation via $R^{X}_{C\leftarrow O}=R_{C\leftarrow L_X}R_{L_X\leftarrow O}$, so ambiguous matches generate a cloud of rotation hypotheses. Density filtering on a discretized SO(3) removes spurious hypotheses, geometrically consistent correspondences are grouped, and PnP-RANSAC with a descriptor-plus-mask score turns each group into a scored 6D pose. On the BOP-Distrib protocol, the paper reports that this outperforms previous distribution methods and dedicated single-pose methods on the T-LESS benchmark.

Load-bearing premise

The method assumes that the learned descriptors and local frames will automatically become symmetry-aware from synthetic training images alone, without explicit symmetry labels; if that implicit learning fails, the per-pixel rotation hypotheses that feed the entire pipeline would be wrong.

Editorial extensions

If this is right

  • Robotic grasping can query the distribution directly: a grasp is only attempted if some pose in the distribution is reachable, which matters when symmetric objects admit only some feasible poses.
  • Single-pose estimation can be improved by first predicting the full distribution and then picking the best scored pose, as the paper's best-pose results exceed dedicated single-pose methods.
  • Correspondence-based pipelines do not have to force every match into one consistent pose; keeping multiple matches per pixel and grouping by rotation hypotheses keeps PnP-RANSAC tractable.
  • Density-based filtering on SO(3) gives a principled way to decide whether an object is unambiguous, discretely symmetric, or continuously symmetric from the shape of the hypothesis cloud.
  • Scoring reprojection error in descriptor space rather than pixel space reduces the penalty for correct symmetric matches.

Reading between the lines

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

  • If the implicit symmetry-awareness claim holds, explicit symmetry annotations and symmetry-group enumeration could be removed entirely for new objects, reducing dataset preparation cost.
  • A category-level variant seems feasible: the per-object implicit fields for descriptors and local frames could be conditioned on a category code, since nothing in the loss requires an object-specific geometry.
  • A natural stress test is to evaluate on objects with near-symmetries, where the paper's own failure analysis predicts the distribution will over-spread; an attention mechanism over small disambiguating elements is a plausible remedy.
  • The reported runtime is dominated by per-group PnP refinement, so using rotation-bin densities to prune groups before refinement is a cheap, testable speed-up.
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 / 4 minor

Summary. The manuscript introduces Corr2Distrib, a correspondence-based method for estimating a distribution over 6DoF poses of a known rigid object from a single RGB image crop. The authors learn per-point descriptors and local coordinate frames for the object model, together with a pixel-wise predictor of descriptors, local frames, and object masks. At inference, each pixel is matched to a set of model points by descriptor similarity; each 2D-3D correspondence yields a rotation hypothesis via the learned local-frame composition; hypotheses are density-filtered over a HealPix discretization of SO(3); correspondences whose hypotheses fall in the same rotation bin are grouped; PnP-RANSAC and a descriptor/mask agreement score convert each group into a scored 6D pose. Experiments on T-LESS with the BOP-Distrib protocol compare against LiePose and SpyroPose for pose-distribution estimation and also evaluate the best pose from the distribution against single-pose methods. The paper claims state-of-the-art performance on both tasks.

Significance. If the central mechanism works as claimed, this is a genuinely new approach to pose ambiguity: ambiguous correspondences are used to enumerate the modes of the pose distribution rather than being discarded, and the method outputs an explicit set of scored poses rather than an implicit density. The pipeline is described in sufficient detail to be reimplementable, and the components (InfoNCE descriptor learning, local-frame rotation composition, SO(3) density filtering, correspondence grouping, PnP-RANSAC, and scoring) are clearly specified. The reported improvements over LiePose and SpyroPose on T-LESS are plausible. However, the significance is conditional on two things the paper does not currently establish: (i) that the learned descriptors and local frames are actually symmetry-aware in the exact sense required by Eq. (5), and (ii) that the evaluation, which relies on a single non-public benchmark and aggregate single-run numbers, is robust. Section IV-F itself acknowledges the single-dataset limitation, which further tempers the strength of the state-of-the-art claim.

major comments (4)
  1. [III-A1, Eqs. (1)-(3)] The correctness of the method depends on the learned descriptors and local frames being symmetry-aware: for a symmetry mapping X to X', the descriptor of X' must be similar enough to the image descriptor to pass the threshold of Eq. (4), and the local frames must satisfy Pf(X') = Pf(X) * s^{-1} so that Eq. (5) produces the equivalent rotation. The text argues this emerges from visual ambiguities even though the losses do not explicitly encode symmetries. The InfoNCE loss in Eq. (1) has a single positive per pixel, and every other model point, including symmetry-equivalent points, enters the denominator of Eq. (2) as a negative, so the loss explicitly penalizes the required descriptor invariance. The claim in Section III-A1 that the match/non-match imbalance overrides this is not a derivation. Similarly, the local-frame loss in Eq. (3) supervises only the composed rotation with the annotated pose and contains no term tying phi_LF(X') to phi_LF(X) after the symmetry. This premise is load-bearing because if it fails, the rotation hypotheses in Eq. (5) are not all valid poses and the subsequent density filtering, grouping, and PnP-RANSAC stages inherit wrong inputs. I request a direct measurement of symmetry-awareness, for example the fraction of symmetric matches that pass Eq. (4), the descriptor similarity between symmetric model points, or the local-frame equivariance error under known symmetries, or an ablation that trains with explicit symmetry labels.
  2. [Eq. (4)] As written, the threshold in Eq. (4) cannot select high-similarity matches. Eq. (2) defines simdesc as a log-probability, hence non-positive. With delta = tau_desc * simdesc(I,x,Xhat) and tau_desc = 0.65, taking the maximal log-similarity to be, say, -2 gives delta = -1.3. The condition simdesc(I,x,X) < delta then selects descriptors with values below -1.3, i.e., the least similar descriptors, and in fact all model points with similarity at most -2 would satisfy it. This would make the match set S(I,x) trivially large. The intended behavior is presumably a threshold that keeps descriptors above a fraction of the maximum similarity. Please clarify whether the matching score in Eq. (4) is the unnormalized inner product rather than the log-probability, whether the inequality should be reversed, or whether delta should be defined differently. This is not merely notational: Eq. (4) is the entry point for all subsequent stages.
  3. [IV-C and IV-D, Tables I-II] The state-of-the-art claim rests on BOP-Distrib [27], an unpublished benchmark whose baseline numbers are taken from that paper, and the tables report only aggregate means over the 30 T-LESS objects. There are no per-object results, no standard deviations, no confidence intervals, and no indication of the number of runs. Because the RANK column in Table I compares three methods on a few aggregate numbers, it is unclear whether the reported differences are statistically meaningful. To support the central claim, the paper should provide object-wise results and error bars or a statistical test, and should either rerun the baselines under the same training data and crop protocol or include enough BOP-Distrib protocol details for the reader to verify comparability. Section IV-F acknowledges the single-dataset limitation, which further qualifies the claim of state-of-the-art performance as stated in the abstract and introduction.
  4. [IV-E and III-A] The ablation study in Table III covers only the hyperparameters k, tau_corr, tau_dens, and tau_score. It does not test the components that are presented as the core contribution, namely the learned symmetry-aware descriptor field and the learned local-frame field. Without an ablation that, for example, replaces the learned local frames with fixed geometric frames, or replaces the multi-match selection of Eq. (4) with nearest-neighbor-only matching, it is difficult to attribute the reported performance to the proposed mechanism. A causal test of the central design choice is needed to substantiate the claim that ambiguous correspondences are being exploited rather than merely tolerated.
minor comments (4)
  1. [Eq. (1) and Eq. (2)] The signature of simdesc is inconsistent: Eq. (1) calls simdesc(psi_d(I,x), h(x)) while Eq. (2) defines simdesc(I,x,X). Please unify the notation so that the arguments match the definitions.
  2. [Eq. (3)] The sentence after Eq. (3) says 'with M_i the object mask in image C', but C denotes the camera frame earlier in the same section. It should presumably be image I.
  3. [References [27] and [23]] BOP-Distrib [27] is described only as 'under review' with an arXiv preprint number but no version or URL, and GPose [23] is an unpublished submission. Since Tables I and II depend on numbers from [27], please provide an archival reference or a supplement with the protocol details.
  4. [Table III] The table marks the best precision-recall trade-off in red, which is not distinguishable in grayscale. Please use bold or a symbol to indicate the best values.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the pose-distribution pipeline is a supervised prediction pipeline, not a fitted input renamed as a prediction.

full rationale

The derivation chain is self-contained: Eqs. 1-2 train descriptor fields with an InfoNCE loss, Eq. 3 trains local frames against ground-truth rotations, Eqs. 4-5 convert per-correspondence predictions into rotation hypotheses, Eq. 6 filters by SO(3) density, and Eqs. 7-9 score pose hypotheses with the same learned descriptors and predicted mask. No stage uses the BOP-Distrib precision/recall metrics as a training input or as a fitted parameter renamed as a prediction. The statement in Section III-A1 that object symmetries 'will naturally be respected due to the visual ambiguities they introduce' is an empirical assumption about what the contrastive loss learns; it may be unsupported, but it is not a definitional identity that equates input with output. The evaluation relies on BOP-Distrib [27], which is under review, and only on T-LESS, as acknowledged in Section IV-F; these are external-validity concerns, not circularity. Citations such as SurfEmb [15] and BOP Challenge [24] provide training and evaluation conventions, but the central pose-distribution output is not enforced by those citations. No quotable equation-level reduction from output back to input was found, so the paper does not exhibit significant circularity.

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

The central claim rests on the emergent symmetry-awareness of learned descriptors and local frames, on the heuristic density-based filtering, and on the validity of the unpublished BOP-Distrib benchmark. No new physical entities are introduced; the symmetry-aware representation is a learned parameter field, not a postulated particle or force.

free parameters (4)
  • tau_corr (descriptor matching threshold ratio) = 0.65
    Controls the 2D-3D correspondence threshold delta = tau_corr * max similarity; ablated in Table III and selected to optimize test-set precision-recall.
  • tau_dens (rotation hypothesis density threshold) = 10
    Filters rotation hypotheses by SO(3) bin density; ablated in Table III and chosen as a precision-recall balance.
  • tau_score (pose scoring threshold) = 0.9
    Filters final poses; has the largest effect in Table III and is selected on the test set to maximize the precision-recall trade-off.
  • Grid level k for SO(3) discretization (G_k) = 4
    Equi-volumetric partition of SO(3) used for density filtering and grouping; ablated in Table III.
assumptions (4)
  • domain assumption Visual ambiguities force the InfoNCE-trained descriptors and local frames to become symmetry-aware without explicit symmetry supervision.
    The paper relies on this emergent property in Section III-A1/2 to generate correct rotation hypotheses from symmetric matches; it is asserted, not proven.
  • domain assumption Per-pixel local frame fields, combined with object-surface local frames, yield a valid camera rotation for each correspondence.
    Equation 5 assumes psi_LF and phi_LF are consistent so that R_C<-O = R_C<-LX * R_LX<-O; if local frames are misaligned, single-correspondence rotations are wrong.
  • ad hoc to paper The density of rotation hypotheses in HealPix bins identifies valid pose regions.
    Section III-B2; the density threshold tau_dens is a heuristic with no theoretical guarantee that dense bins correspond to valid poses.
  • domain assumption BOP-Distrib ground-truth distributions correctly model the set of valid poses under occlusion and symmetry.
    Evaluation relies on this unpublished benchmark; if its ground truths are flawed, the quantitative claims do not hold.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Corr2Distrib: Making Ambiguous Correspondences an Ally to Predict Reliable 6D Pose Distributions." pith.science (2026). https://pith.science/paper/EP52O7MQ

@misc{pith2026250502501,
  author       = {Pith},
  title        = {Pith review of: Corr2Distrib: Making Ambiguous Correspondences an Ally to Predict Reliable 6D Pose Distributions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EP52O7MQ}},
  note         = {Machine review of arXiv:2505.02501}
}
read the original abstract

We introduce Corr2Distrib, the first correspondence-based method which estimates a 6D camera pose distribution from an RGB image, explaining the observations. Indeed, symmetries and occlusions introduce visual ambiguities, leading to multiple valid poses. While a few recent methods tackle this problem, they do not rely on local correspondences which, according to the BOP Challenge, are currently the most effective way to estimate a single 6DoF pose solution. Using correspondences to estimate a pose distribution is not straightforward, since ambiguous correspondences induced by visual ambiguities drastically decrease the performance of PnP. With Corr2Distrib, we turn these ambiguities into an advantage to recover all valid poses. Corr2Distrib first learns a symmetry-aware representation for each 3D point on the object's surface, characterized by a descriptor and a local frame. This representation enables the generation of 3DoF rotation hypotheses from single 2D-3D correspondences. Next, we refine these hypotheses into a 6DoF pose distribution using PnP and pose scoring. Our experimental evaluations on complex non-synthetic scenes show that Corr2Distrib outperforms state-of-the-art solutions for both pose distribution estimation and single pose estimation from an RGB image, demonstrating the potential of correspondences-based approaches.

Figures

Figures reproduced from arXiv: 2505.02501 by the authors.

Figure 1
Figure 1. Pose distribution comparison of our method with SpyroPose [12] and LiePose [11] on the T-LESS dataset [13]. We illustrate three cases of the object 1 from top to bottom: no symmetry, six-fold rotational symmetry, and continuous rotational symmetry. Rotations are shown in 2D (Mollweide projection), with color indicating tilt direction and marker size representing pose probability [14]. The circles represent ground tr… view at source ↗
Figure 2
Figure 2. Overview of Corr2Distrib. We first learn offline a Discrete Symmetry-Aware object model (Section III-A), consisting in a descriptor and a local frame per object surface point. At inference, given an image crop, we estimate for each pixel a descriptor and a local frame orientation relative to the camera. Then, we establish many-to-many 2D-3D correspondences between image pixels and symmetry-aware model points. From e… view at source ↗
Figure 3
Figure 3. Visualization of the transformations used in local frames learning. The different frames are: camera frame, object frame, and local frames (LX). For a given 2D-3D correspondence (x, X), ψLF returns RC←LX and ϕLF returns RLX←O. The combination of these rotations provides the desired rotation RC←O, which is supervised by R gt C←O. To reach this objective, we train ψd and ϕd simultaneously using the contrastive loss Ld… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Evolution of the rotation distribution from coarse to fine. We start by estimating an initial set of rotation hypotheses from single 2D-3D correspondences (Section III-B). This set is then filtered by discarding rotations according to observa￾tion conditions and SO(3) …
Figure 5
Figure 5. Figure 5: Failure cases. The figure illustrates two key sources of error: too small disambiguating element (first row), and texture inconsistencies between 3D object model and observed object (second row). D. Single Pose Evaluation Evaluated methods. For a deeper comprehensive a…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 46 canonical work pages

  1. [27]

    Bop-distrib: Revisiting 6d pose estimation benchmark for better evaluation under visual ambiguities,

    under review, “Bop-distrib: Revisiting 6d pose estimation benchmark for better evaluation under visual ambiguities,” in arXiv Preprint, 2024

    work page 2024

  2. [1]

    Densefusion: 6d object pose estimation by iterative dense fusion,

    C. Wang, D. Xu, Y . Zhu, R. Mart ´ın-Mart´ın, C. Lu, L. Fei-Fei, and S. Savarese, “Densefusion: 6d object pose estimation by iterative dense fusion,” in CVPR, 2019

    work page 2019

  3. [2]

    Posecnn: A convolu- tional neural network for 6d object pose estimation in cluttered scenes,

    Y . Xiang, T. Schmidt, V . Narayanan, and D. Fox, “Posecnn: A convolu- tional neural network for 6d object pose estimation in cluttered scenes,” RSS, 2017. 8

    work page 2017

  4. [3]

    Robotic task success evaluation under multi-modal non-parametric object pose uncertainty,

    L. Naik, T. M. Iversen, A. Kramberger, and N. Kr ¨uger, “Robotic task success evaluation under multi-modal non-parametric object pose uncertainty,” arXiv preprint arXiv:2403.10874, 2024

    arXiv 2024

  5. [4]

    Computer vision systems for industrial inspection and assembly,

    G. J. Agin, “Computer vision systems for industrial inspection and assembly,” Computer, 1980

    work page 1980

  6. [5]

    Vision-guided robotic assembly using uncalibrated vision,

    B. Zhang, J. Wang, G. Rossano, and C. Martinez, “Vision-guided robotic assembly using uncalibrated vision,” in ICMA. IEEE, 2011

    work page 2011

  7. [6]

    Stereo vision based automation for a bin-picking solution,

    J.-K. Oh, S. Lee, and C.-H. Lee, “Stereo vision based automation for a bin-picking solution,” IJCAS, 2012

    work page 2012

  8. [7]

    Depth image–based deep learning of grasp planning for textureless planar-faced objects in vision-guided robotic bin-picking,

    P. Jiang, Y . Ishihara, N. Sugiyama, J. Oaki, S. Tokura, A. Sugahara, and A. Ogawa, “Depth image–based deep learning of grasp planning for textureless planar-faced objects in vision-guided robotic bin-picking,” Sensors, 2020

    work page 2020

Show all 49 references
  1. [8]

    A machine learning approach for collaborative robot smart manufacturing inspection for quality control systems,

    T. Brito, J. Queiroz, L. Piardi, L. A. Fernandes, J. Lima, and P. Leit˜ao, “A machine learning approach for collaborative robot smart manufacturing inspection for quality control systems,” Procedia Manufacturing, vol. 51, 2020

    2020

  2. [9]

    Self-learning processes in smart factories: Deep reinforcement learning for process control of robot brine injection,

    R. E. Andersen, S. Madsen, A. B. Barlo, S. B. Johansen, M. Nør, R. S. Andersen, and S. Bøgh, “Self-learning processes in smart factories: Deep reinforcement learning for process control of robot brine injection,” Procedia Manufacturing, 2019

    2019

  3. [10]

    Explaining the ambiguity of object detection and 6d pose from visual data,

    F. Manhardt, D. M. Arroyo, C. Rupprecht, B. Busam, T. Birdal, N. Navab, and F. Tombari, “Explaining the ambiguity of object detection and 6d pose from visual data,” in ICCV, 2019

    2019

  4. [11]

    Confronting ambiguity in 6d object pose estimation via score-based diffusion on se(3),

    T.-C. Hsiao, H.-W. Chen, H.-K. Yang, and C.-Y . Lee, “Confronting ambiguity in 6d object pose estimation via score-based diffusion on se(3),” in CVPR, 2024

    2024

  5. [12]

    Spyropose: Se(3) pyramids for object pose distribution estimation,

    R. L. Haugaard, F. Hagelskjær, and T. M. Iversen, “Spyropose: Se(3) pyramids for object pose distribution estimation,” in ICCV, 2023

    2023

  6. [13]

    T-less: An rgb-d dataset for 6d pose estimation of texture-less objects,

    T. Hodan, P. Haluza, ˇS. Obdr ˇz´alek, J. Matas, M. Lourakis, and X. Zab- ulis, “T-less: An rgb-d dataset for 6d pose estimation of texture-less objects,” in W ACV. IEEE, 2017

    2017

  7. [14]

    Implicit-pdf: Non-parametric representation of probability distributions on the rotation manifold,

    K. A. Murphy, C. Esteves, V . Jampani, S. Ramalingam, and A. Makadia, “Implicit-pdf: Non-parametric representation of probability distributions on the rotation manifold,” in ICML. PMLR, 2021

    2021

  8. [15]

    Surfemb: Dense and continuous correspondence distributions for object pose estimation with learnt surface embeddings,

    R. L. Haugaard and A. G. Buch, “Surfemb: Dense and continuous correspondence distributions for object pose estimation with learnt surface embeddings,” in CVPR, 2022

    2022

  9. [16]

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

    Y . Labb ´e, L. Manuelli, A. Mousavian, S. Tyree, S. Birchfield, J. Trem- blay, J. Carpentier, M. Aubry, D. Fox, and J. Sivic, “Megapose: 6d pose estimation of novel objects via render & compare,” CoRL, 2022

    2022

  10. [17]

    Augmented autoencoders: Implicit 3d orientation learning for 6d object detection,

    M. Sundermeyer, Z.-C. Marton, M. Durner, and R. Triebel, “Augmented autoencoders: Implicit 3d orientation learning for 6d object detection,” IJCV, 2020

    2020

  11. [18]

    Zebrapose: Coarse to fine surface encoding for 6dof object pose estimation,

    Y . Su, M. Saleh, T. Fetzer, J. Rambach, N. Navab, B. Busam, D. Stricker, and F. Tombari, “Zebrapose: Coarse to fine surface encoding for 6dof object pose estimation,” in CVPR, 2022

    2022

  12. [19]

    Dpod: 6d pose object detector and refiner,

    S. Zakharov, I. Shugurov, and S. Ilic, “Dpod: 6d pose object detector and refiner,” in ICCV, 2019

    2019

  13. [20]

    Gdr-net: Geometry- guided direct regression network for monocular 6d object pose estima- tion,

    G. Wang, F. Manhardt, F. Tombari, and X. Ji, “Gdr-net: Geometry- guided direct regression network for monocular 6d object pose estima- tion,” in CVPR, 2021

    2021

  14. [21]

    Pix2pose: Pixel-wise coordinate regression of objects for 6d pose estimation,

    K. Park, T. Patten, and M. Vincze, “Pix2pose: Pixel-wise coordinate regression of objects for 6d pose estimation,” in ICCV, 2019

    2019

  15. [22]

    Yolo-6d-pose: Enhanc- ing yolo for single-stage monocular multi-object 6d pose estimation,

    D. Maji, S. Nagori, M. Mathew, and D. Poddar, “Yolo-6d-pose: Enhanc- ing yolo for single-stage monocular multi-object 6d pose estimation,” in 3DV. IEEE, 2024

    2024

  16. [23]

    GPose2023, a submission to the BOP Challenge 2023,

    R. Zhang, Z. Huang, G. Wang, X. Liu, C. Zhang, and X. Ji, “GPose2023, a submission to the BOP Challenge 2023,” 2023, unpublished, http: //bop.felk.cvut.cz/method info/410/

    2023

  17. [24]

    Bop challenge 2020 on 6d object localization,

    T. Hoda ˇn, M. Sundermeyer, B. Drost, Y . Labb ´e, E. Brachmann, F. Michel, C. Rother, and J. Matas, “Bop challenge 2020 on 6d object localization,” in ECCV Worksh. Springer, 2020

    2020

  18. [25]

    Complete solution classification for the perspective-three-point problem,

    X.-S. Gao, X.-R. Hou, J. Tang, and H.-F. Cheng, “Complete solution classification for the perspective-three-point problem,” IEEE TPAMI, 2003

    2003

  19. [26]

    Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography,

    M. A. Fischler and R. C. Bolles, “Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography,” Commun. ACM, 1981

    1981

  20. [28]

    Bop challenge 2023 on detection segmentation and pose estimation of seen and unseen rigid objects,

    T. Hodan, M. Sundermeyer, Y . Labbe, V . N. Nguyen, G. Wang, E. Brach- mann, B. Drost, V . Lepetit, C. Rother, and J. Matas, “Bop challenge 2023 on detection segmentation and pose estimation of seen and unseen rigid objects,” in CVPR, 2024

    2023

  21. [29]

    Bb8: A scalable, accurate, robust to partial occlusion method for predicting the 3d poses of challenging objects without using depth,

    M. Rad and V . Lepetit, “Bb8: A scalable, accurate, robust to partial occlusion method for predicting the 3d poses of challenging objects without using depth,” in ICCV, 2017

    2017

  22. [30]

    6d-diff: A keypoint diffusion framework for 6d object pose estimation,

    L. Xu, H. Qu, Y . Cai, and J. Liu, “6d-diff: A keypoint diffusion framework for 6d object pose estimation,” in CVPR, 2024

    2024

  23. [31]

    Dense object nets: Learn- ing dense visual object descriptors by and for robotic manipulation,

    P. R. Florence, L. Manuelli, and R. Tedrake, “Dense object nets: Learn- ing dense visual object descriptors by and for robotic manipulation,” CoRL, 2018

    2018

  24. [32]

    Posenet: A convolutional network for real-time 6-dof camera relocalization,

    A. Kendall, M. Grimes, and R. Cipolla, “Posenet: A convolutional network for real-time 6-dof camera relocalization,” in ICCV, 2015

    2015

  25. [33]

    Real-time seamless single shot 6d object pose prediction,

    B. Tekin, S. N. Sinha, and P. Fua, “Real-time seamless single shot 6d object pose prediction,” in CVPR, 2018

    2018

  26. [34]

    Cosypose: Consistent multi-view multi-object 6d pose estimation,

    Y . Labb ´e, J. Carpentier, M. Aubry, and J. Sivic, “Cosypose: Consistent multi-view multi-object 6d pose estimation,” in ECCV, 2020

    2020

  27. [35]

    Ssd-6d: Making rgb-based 3d detection and 6d pose estimation great again,

    W. Kehl, F. Manhardt, F. Tombari, S. Ilic, and N. Navab, “Ssd-6d: Making rgb-based 3d detection and 6d pose estimation great again,” in ICCV, 2017

    2017

  28. [36]

    Osop: A multi-stage one shot object pose estimation framework,

    I. Shugurov, F. Li, B. Busam, and S. Ilic, “Osop: A multi-stage one shot object pose estimation framework,” in CVPR, 2022

    2022

  29. [37]

    Templates for 3d object pose estimation revisited: Generalization to new objects and robustness to occlusions,

    V . N. Nguyen, Y . Hu, Y . Xiao, M. Salzmann, and V . Lepetit, “Templates for 3d object pose estimation revisited: Generalization to new objects and robustness to occlusions,” in CVPR, 2022

    2022

  30. [38]

    Hyperposepdf - hyper- networks predicting the probability distribution on so(3),

    T. H ¨ofer, B. Kiefer, M. Messmer, and A. Zell, “Hyperposepdf - hyper- networks predicting the probability distribution on so(3),” in W ACV, January 2023

    2023

  31. [39]

    Sc6d: Symmetry-agnostic and correspondence-free 6d object pose estimation,

    D. Cai, J. Heikkil ¨a, and E. Rahtu, “Sc6d: Symmetry-agnostic and correspondence-free 6d object pose estimation,” in 3DV. IEEE, 2022

    2022

  32. [40]

    Ki-pode: Keypoint- based implicit pose distribution estimation of rigid objects,

    T. M. Iversen, R. L. Haugaard, and A. G. Buch, “Ki-pode: Keypoint- based implicit pose distribution estimation of rigid objects,” in BMVC. BMV A Press, 2022

    2022

  33. [41]

    3d-rcnn: Instance-level 3d object reconstruction via render-and-compare,

    A. Kundu, Y . Li, and J. M. Rehg, “3d-rcnn: Instance-level 3d object reconstruction via render-and-compare,” in CVPR, 2018

    2018

  34. [42]

    Generating uniform incremental grids on so (3) using the hopf fibration,

    A. Yershova, S. M. LaValle, and J. C. Mitchell, “Generating uniform incremental grids on so (3) using the hopf fibration,” in W AFR. Springer, 2010

    2010

  35. [43]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in CVPR, 2016

    2016

  36. [44]

    Imagenet: A large-scale hierarchical image database,

    J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei, “Imagenet: A large-scale hierarchical image database,” in CVPR. Ieee, 2009

    2009

  37. [45]

    Implicit neural representations with periodic activation functions,

    V . Sitzmann, J. Martel, A. Bergman, D. Lindell, and G. Wetzstein, “Implicit neural representations with periodic activation functions,” NeurIPS, vol. 33, 2020

    2020

  38. [46]

    On the continuity of rotation representations in neural networks,

    Y . Zhou, C. Barnes, J. Lu, J. Yang, and H. Li, “On the continuity of rotation representations in neural networks,” in CVPR, 2019

    2019

  39. [47]

    HccePose2024, a submission to the BOP Challenge 2024,

    H. L. Yulin Wang, Mengting Hu and C. Luo, “HccePose2024, a submission to the BOP Challenge 2024,” 2024, unpublished, https: //bop.felk.cvut.cz/method info/636/

    2024

  40. [48]

    GDet2023, a submission to the BOP Challenge 2023,

    R. Zhang, Z. Huang, G. Wang, X. Liu, C. Zhang, and X. Ji, “GDet2023, a submission to the BOP Challenge 2023,” 2023, unpublished, http: //bop.felk.cvut.cz/method info/411/

    2023

  41. [49]

    GDRNPPDet, a submission to the BOP Challenge 2022,

    X. Liu, R. Zhang, C. Zhang, B. Fu, J. Tang, X. Liang, J. Tang, X. Cheng, Y . Zhang, G. Wang, and X. Ji, “GDRNPPDet, a submission to the BOP Challenge 2022,” 2022, unpublished, http://bop.felk.cvut.cz/ method info/281/

    2022

Pith tools

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