Pith. sign in

REVIEW 4 major objections 4 minor 30 references

PIXIE claims a single untextured CAD model is enough to estimate the 6D pose of an unseen object from one RGB image, despite texture changes and assembly defects.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

A training-free geometry-only pipeline matches RGB images against rendered depth/normal maps to estimate 6D poses of unseen, textureless, and slightly defective objects from one image.

T0 review reviewed 2026-08-01 challenge →

load-bearing objection A clean, useful zero-shot geometry-only pose pipeline with a valuable defect dataset, but the central claim leans on an untested cross-modality matcher transfer that needs an ablation. the 4 major comments →

arxiv 2607.16015 v2 pith:QFPM33X4 submitted 2026-07-17 cs.CV cs.RO

PIXIE: A Zero-Shot texture-invariant 6D pose estimation framework for unseen objects with assembly defects

classification cs.CV cs.RO MSC 68T4568U10
keywords zero-shot 6D pose estimationtexture invariancegeometry-only pose estimationcross-modality feature matchinguntextured CAD modelassembly defectsdepth and normal renderingPnP pose estimation
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The reading

PIXIE is a zero-shot system that estimates the 6D pose — position plus orientation — of an object from a single RGB photo, using nothing more than an untextured CAD model of the object. Instead of learning what the object looks like, it renders depth and surface-normal maps from a set of reference viewpoints, converts them to color images, and matches those synthetic views to the real photo with a pre-trained cross-modality feature matcher. Matched keypoints are back-projected to known 3D model coordinates, so a PnP solver with RANSAC can compute the pose, and an iterative step re-matches from the nearest reference view to refine it. The authors report strong results on textureless industrial benchmarks (43.7% Average Recall on T-LESS, 26.7% on ITODD, 35.1% on TUD-L) and 78.4% on a new dataset with texture changes, occlusions, and assembly defects, and argue this is the first training-free, geometry-only zero-shot pipeline that is robust to both texture changes and geometric deviations. If the claim holds, factories could estimate poses for frequently changing parts without any per-object data pipeline.

Core claim

PIXIE's central claim is that 6D pose of an unseen object can be recovered from one RGB image and an untextured 3D model, with no object-specific training and no reliance on texture. The key move is geometric rendering plus cross-modality matching: depth and normal maps rendered from viewpoints on a Fibonacci sphere are colormapped into 8-bit images, and a pretrained matcher (MINIMA with the LightGlue backbone) finds correspondences between these synthetic geometric views and the real photo. Because the rendering pipeline is fully known, every matched pixel back-projects to a 3D object coordinate, giving 2D–3D correspondences for PnP+RANSAC; an iterative view-selection loop then picks the cl

What carries the argument

The load-bearing mechanism is the pairing of colormapped geometric renderings with a pretrained cross-modality feature matcher. Depth and normal maps rendered from reference viewpoints are normalized, mapped to a fixed colormap, and treated as ordinary images, which lets a matcher trained on diverse modalities find correspondences between synthetic geometry and an RGB query. The controlled rendering geometry is what makes the pipeline work: each pixel in a reference view has a known 3D object coordinate, so a 2D match can be lifted to a 2D–3D correspondence for PnP. Iterative view selection — picking the reference whose viewing axis is closest to the current pose estimate and re-matching — i

Load-bearing premise

The whole approach depends on a generic image-matching network being able to pair real photos with colormapped depth/normal renders of industrial parts it has never seen, and the paper does not isolate how much of the result rests on that one network.

What would settle it

Run the pipeline on a fixed set of query images but change the colormap used to encode the depth maps (e.g., grayscale instead of the default colormap) while keeping geometry identical; if pose accuracy changes substantially, the method is not using geometry alone and the texture-invariance claim fails.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • A factory can estimate the pose of a new part from its nominal CAD model alone, without collecting or annotating training images.
  • Parts that differ from the model due to assembly defects, wear, or manufacturing tolerances still yield usable poses, so inspection can compare against the intended design rather than a per-instance scan.
  • Because the pipeline never touches texture, surface finish, paint color, or lighting have no effect on the estimate, eliminating a common source of domain shift.
  • The approach is competitive on textureless benchmarks without the refinement stage that most trained systems rely on, and the authors show both depth and normal maps contribute complementary signal.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The paper's own failure cases — feature-poor or symmetric objects and heavy occlusion — suggest a hard limit: any geometry-only method needs distinct geometric landmarks, so fusing appearance cues when texture exists could extend the approach without losing its zero-shot property.
  • If the cross-modality matcher proves to be the bottleneck, the architecture could improve simply by swapping in a matcher trained on more industrial geometry, since the rest of the pipeline is matcher-agnostic; this is a testable prediction.
  • The custom dataset with measured geometric deviations (APS and IoU3D) offers a reusable protocol for quantifying how much model-to-real deviation zero-shot pose estimators can tolerate; similar defect-inclusive benchmarks could become standard for industrial evaluation.
  • Since all reference views are precomputed, the pipeline could be adapted to video by using temporal priors to skip the coarse matching stage, which would address the 3.6 s per-image runtime.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The paper proposes PIXIE, a zero-shot 6D pose estimation pipeline that uses only an untextured CAD model and a single RGB query image. It renders synthetic depth and normal maps from offline-sampled reference viewpoints, colormaps these geometric maps, matches them to the query with a pretrained cross-modality matcher (MINIMA/LightGlue), lifts the 2D matches to 2D-3D correspondences via known rendering geometry, and solves for pose with PnP+RANSAC. A single iteration of nearest-view re-matching is used as refinement. The authors evaluate on BOP subsets (LM-O, T-LESS, TUD-L, ITODD) and introduce a custom LEGO-based dataset with texture changes, assembly defects, and occlusion. They report 43.7% AR on T-LESS, 26.7% on ITODD, 35.1% on TUD-L, and 78.4% AR on the custom dataset, and claim state-of-the-art zero-shot performance on textureless objects without object-specific training.

Significance. If validated, the core idea is significant: it would show that a generic pretrained cross-modality matcher can transfer to the unusual domain of colormapped depth/normal renderings, enabling six-degree-of-freedom pose estimation for unseen industrial objects from untextured CAD alone. This would be a practical advance for settings where textured models and per-object training data are unavailable. The paper also ships a useful public dataset with paired target/defective models, and the single-modality ablations are helpful. However, the strongest empirical and conceptual claims rest on two load-bearing assumptions that are not adequately tested: the transferability of MINIMA to the synthetic geometric-reference domain, and the provenance/completeness of the public-benchmark numbers. The absence of error bars, the small baseline set, and the overstatement of 'state-of-the-art' claims further limit confidence.

major comments (4)
  1. [Section III-D and IV-A] The pipeline's central claim depends on the pretrained cross-modality matcher MINIMA/LightGlue establishing reliable 2D matches between real RGB queries and colormapped depth/normal renderings of unseen objects. The paper states that 'Our framework is agnostic to the specific cross-modality matcher' (Section III-D), but no ablation tests this. There is no comparison with another matcher (e.g., MatchAnything [22]), no diagnostic of correspondence precision/recall in the rendered-reference domain, and no sensitivity analysis with respect to matching confidence thresholds or the number of in-plane rotations. Since every PnP estimate in Eq. (4) depends on these matches, the zero-shot transfer claim is currently unsupported.
  2. [Table I and Section IV-B] The provenance of the BOP results is ambiguous and must be clarified. The table caption says 'benchmark numbers and runtimes obtained from Pos3R [14]', while the text in IV-B credits PIXIE with 'achieved 43.7% AR' on T-LESS. If the PIXIE rows are copied from Pos3R rather than produced by the authors, the state-of-the-art claim relies entirely on a third party's evaluation and needs independent verification. If the PIXIE rows are the authors' own runs, the caption is misleading and the table row citing '[27]' for Pos3R is incorrect (reference [27] is Pose3RCNN, not Pos3R). This issue must be resolved before the empirical claims can be assessed.
  3. [Abstract, Conclusion, Table I] The 'state-of-the-art results on textureless objects' claim is contradicted by the paper's own Table I. On TUD-L, PIXIE scores 35.1% AR, while Pos3R scores 43.2% and FoundPose 46.7%, both training-free methods. The claim is only defensible if restricted to specific datasets (e.g., T-LESS and ITODD) and to the small set of compared methods. Additionally, the benchmark comparison includes only six methods, all with numbers borrowed from a single source, and no error bars or significance measures are reported. Please qualify the claims accordingly and, if possible, add confidence intervals or per-object breakdowns.
  4. [Section IV-C, Table III] The custom dataset is the only place where the method is tested under the claimed conditions of simultaneous texture variation and geometric defects, but it consists exclusively of LEGO brick assemblies. This limits the generality of the 'real-world applicability' conclusion. Moreover, the comparison on this dataset omits Pos3R, which is the strongest zero-shot baseline in Table I and whose exclusion weakens the 'significantly outperformed' claim. Please either add Pos3R to the custom-dataset comparison or justify its omission.
minor comments (4)
  1. [Table I / References] There is a citation inconsistency: the table uses '[27]' for Pos3R, but reference [27] is 'Pose3RCNN'. The text in Section IV-A says benchmark numbers are from Pos3R [14]. Please correct.
  2. [Equation (5)] The notation for APS reuses D_i for the target-model depth map and D_d for the defective-model depth map, while D_i was already used for the reference depth map in Eq. (3). This is confusing and should be renamed.
  3. [Section IV-A] The number of in-plane rotations (8 at 45-degree increments) is fixed with no discussion of its effect on runtime or accuracy. A brief sensitivity note or a reference to a supplementary figure would help.
  4. [Section III-F] The termination criterion says 'converges across iterations' but only a single iteration is used in all experiments. Please state explicitly whether the reported results use exactly one iteration and why further iterations were not needed.

Circularity Check

0 steps flagged

No significant circularity: PIXIE composes an external pretrained matcher, standard PnP geometry, and fixed CAD renderings without fitting parameters to the reported poses.

full rationale

The paper's derivation chain is: render depth/normal maps from an untextured CAD model (Sec. III-C), match the RGB query to colormapped renders with the external pretrained matcher MINIMA/LightGlue (Sec. III-D), back-project matched pixels to 3D using known rendering matrices (Eqs. 1-4), solve PnP+RANSAC (Sec. III-E), and refine by selecting the nearest precomputed reference view (Sec. III-F). None of these steps fits a free parameter to the target pose or to the benchmark numbers. MINIMA/LightGlue is an independently trained matcher (Refs. 20/26) whose weights are not adapted to any object in T-LESS, ITODD, TUD-L, or the custom dataset; relying on it is an input assumption, not a circular prediction. The method's own design choices—colormap encoding, 121 reference views, 8 in-plane rotations, RANSAC thresholds—are manual constants, not regression fits, and the only ablations (depth-only vs normal-only) vary the pipeline's own inputs rather than fitting to outputs. The self-citations in Sec. II [6],[7] are ordinary related-work citations and are not used to justify the zero-shot or geometry-only claim. There is no self-citation chain, no imported uniqueness theorem, and no equation where the predicted pose is defined in terms of the evaluation target. The lack of an ablation isolating MINIMA's transfer to colormapped depth/normal renders is a genuine evidence gap about external validity, but it is not circularity: the reported AR values are external benchmark outcomes produced with a fixed, pretrained matcher. The custom dataset is self-created but not circular, since ground truth comes from marker-based calibration and comparisons use official baseline implementations.

Axiom & Free-Parameter Ledger

4 free parameters · 6 axioms · 0 invented entities

The central claim rests on standard camera geometry, the zero-shot generalization of an external pretrained matcher, the informativeness of geometry-only renderings, detector quality, and the accuracy of custom ground truth. No free physical constants are introduced; the manually chosen thresholds and view counts are listed above and lack sensitivity analysis. No new physical entities are proposed; the new dataset and APS metric are evaluation tools rather than theoretical entities.

free parameters (4)
  • RANSAC reprojection error threshold = 5 pixels
    Manually chosen PnP inlier threshold used for all experiments; no sensitivity analysis reported (Section IV-A).
  • Minimum inlier ratio = 0.5
    Manual threshold for accepting PnP pose; no sensitivity analysis (Section IV-A).
  • Number of fine reference views = 115 (plus 6 coarse)
    Fibonacci-lattice sampling density chosen without ablation; affects matching coverage and runtime (Section III-C/IV-A).
  • In-plane rotation augmentation = 8 rotations at 45-degree increments
    Heuristic to handle in-plane rotation; no ablation reported (Section III-D/IV-A).
axioms (6)
  • domain assumption Pinhole camera model with known query intrinsics K; rendering intrinsics K_start need only be close to K.
    Standard projection model for PnP; K may vary across BOP images while references use first calibration, so validity of pose recovery relies on K accuracy (Section III-A, IV-A).
  • domain assumption Pretrained MINIMA/LightGlue cross-modality matcher transfers to synthetic depth/normal vs real RGB matching for unseen industrial objects.
    Central to correspondence generation; trained on generic data [20], no fine-tuning on target objects or modality pair; if transfer fails, PnP has no inputs (Section III-D, IV-A).
  • domain assumption Geometric-only renderings contain enough discriminative features for pose estimation; texture/color are non-informative.
    Design premise; acknowledged as limitation for symmetric/low-feature objects and low-resolution occluded scenes (Section IV-D).
  • domain assumption CNOS detector provides usable segmentation for unseen objects on BOP; bounding boxes for custom dataset are sufficient.
    Masks used to filter keypoints for all methods; failure of detector would hurt all compared methods, but results depend on it (Section IV-A).
  • standard math PnP+RANSAC with fixed thresholds yields correct pose given enough correct 2D-3D correspondences.
    Standard robust geometry, assumed throughout Section III-E.
  • domain assumption Ground-truth poses in custom dataset (marker-based, 1.6 px avg reprojection error) are accurate enough for ADD/APS evaluation.
    Used for all custom dataset metrics; no uncertainty propagation (Section IV-C).

reviewed 2026-08-01 · how reviews work

0 comments
Cite this review

Pith. "Pith review of PIXIE: A Zero-Shot texture-invariant 6D pose estimation framework for unseen objects with assembly defects." pith.science (2026). https://pith.science/paper/QFPM33X4

@misc{pith2026260716015,
  author       = {Pith},
  title        = {Pith review of: PIXIE: A Zero-Shot texture-invariant 6D pose estimation framework for unseen objects with assembly defects},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QFPM33X4}},
  note         = {Machine review of arXiv:2607.16015}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

6D pose estimation remains a key challenge in robotics and computer vision, particularly in industrial environments. The deployment of currently available data-driven methods is often limited by resource-intensive data pipelines, reliance on textured 3D models, and sensitivity to geometric deviations caused by damages or assembly defects. We present PIXIE, a zero-shot framework that estimates the 6D pose of an object from an RGB image using only an untextured 3D model. Synthetic depth and normal maps are rendered from sampled reference viewpoints and matched to the query image via a pretrained cross-modality feature matcher. Matched keypoints are back-projected to obtain 2D--3D correspondences for PnP-based pose estimation. Relying exclusively on geometry makes the method inherently robust to lighting and texture variation, while correspondence filtering handles geometric deviations between the model and physical object. We evaluate on widely-used public benchmarks, reporting state-of-the-art results on texture-less objects without object-specific training, and introduce a novel dataset with assembly defects, texture variations, and occlusion to demonstrate real-world applicability.

Figures

Figures reproduced from arXiv: 2607.16015 by Alejandro Maga\~na, Daniel Werdehausen, Gautham Mohan, Leon Jungemeyer, Matthias Karl.

Figure 1
Figure 1. Figure 1: Inference pipeline overview. Optional object localization extracts the region of interest from the query image. Pre [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Examples of geometric reference views generated [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Cross-modality feature matching establishes corre [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Qualitative results of our method on the T-LESS [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Examples from our custom dataset: texture variations [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

30 extracted references · 5 linked inside Pith

  1. [1]

    Epnp: An accurate o (n) solution to the pnp problem,

    V . Lepetit, F. Moreno-Noguer, and P. Fua, “Epnp: An accurate o (n) solution to the pnp problem,” inInternational Conference on Computer Vision, pp. 1–8, IEEE, 2009

  2. [2]

    The moped framework: Object recognition and pose estimation for manipulation,

    A. Collet, M. Martinez, and S. S. Srinivasa, “The moped framework: Object recognition and pose estimation for manipulation,”The inter- national journal of robotics research, vol. 30, no. 10, pp. 1284–1306, 2011

  3. [3]

    Posecnn: A convolutional neural network for 6d object pose estimation in cluttered scenes,

    Y . Xiang, T. Schmidt, V . Narayanan, and D. Fox, “Posecnn: A convolutional neural network for 6d object pose estimation in cluttered scenes,” inRobotics: Science and Systems, 2018

  4. [4]

    Deep object pose estimation for semantic robotic grasping of household objects,

    J. Tremblay, T. To, B. Sundaralingam, Y . Xiang, D. Fox, and S. Birch- field, “Deep object pose estimation for semantic robotic grasping of household objects,” inConference on Robot Learning, pp. 306–316, PMLR, 2018

  5. [5]

    Dpod: 6d pose object detector and refiner,

    S. Zakharov, I. Shugurov, and S. Ilic, “Dpod: 6d pose object detector and refiner,” inProceedings of the IEEE/CVF International Conference on Computer Vision, pp. 1941–1950, 2019

  6. [6]

    Posenetwork: Pipeline for the automated generation of synthetic training data and cnn for object detection, segmentation, and orientation estimation,

    A. Maga ˜na, H. Wu, P. Bauer, and G. Reinhart, “Posenetwork: Pipeline for the automated generation of synthetic training data and cnn for object detection, segmentation, and orientation estimation,” in2020 25th IEEE International conference on emerging technologies and factory automation (ETFA), pp. 587–594, IEEE, 2020. (a) Different states of the same 3D...

  7. [7]

    Fast adaptable 6d object pose estimation for autonomous robotics,

    T. Nguyen-Cong, J. Zhu, C. Glasenapp, and M. Karl, “Fast adaptable 6d object pose estimation for autonomous robotics,” in2021 IEEE 17th International Conference on Automation Science and Engineering (CASE), IEEE, 2021

  8. [8]

    Onepose: One-shot object pose estimation without cad models,

    J. Sun, Z. Wang, S. Zhang, X. He, H. Zhao, G. Zhang, and X. Zhou, “Onepose: One-shot object pose estimation without cad models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 6825–6834, 2022

  9. [9]

    Onepose++: Keypoint-free one-shot object pose estimation without cad mod- els,

    X. He, J. Sun, Y . Wang, D. Huang, H. Bao, and X. Zhou, “Onepose++: Keypoint-free one-shot object pose estimation without cad mod- els,”Advances in Neural Information Processing Systems, vol. 35, pp. 35103–35115, 2022

  10. [10]

    Gen6d: Generalizable model-free 6-dof object pose estimation from rgb images,

    Y . Liu, Y . Wen, S. Peng, C. Lin, X. Long, T. Komura, and W. Wang, “Gen6d: Generalizable model-free 6-dof object pose estimation from rgb images,” inEuropean Conference on Computer Vision, pp. 298– 315, Springer, 2022

  11. [11]

    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,”arXiv preprint arXiv:2212.06870, 2022

  12. [12]

    Gigapose: Fast and robust novel object pose estimation via one correspondence,

    V . N. Nguyen, T. Groueix, M. Salzmann, and V . Lepetit, “Gigapose: Fast and robust novel object pose estimation via one correspondence,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 9903–9913, 2024

  13. [13]

    Genflow: Generalizable recurrent flow for 6d pose refinement of novel objects,

    S. Moon, H. Son, D. Hur, and S. Kim, “Genflow: Generalizable recurrent flow for 6d pose refinement of novel objects,” inProceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 10039–10049, 2024

  14. [14]

    Pos3r: 6d pose estimation for unseen objects made easy,

    W. Deng, D. Campbell, C. Sun, J. Zhang, S. Kanitkar, M. E. Shaffer, and S. Gould, “Pos3r: 6d pose estimation for unseen objects made easy,” inProceedings of the Computer Vision and Pattern Recognition Conference, pp. 16818–16828, 2025

  15. [15]

    Bop: Benchmark for 6d object pose estimation,

    T. Hodan, F. Michel, E. Brachmann, W. Kehl, A. GlentBuch, D. Kraft, B. Drost, J. Vidal, S. Ihrke, X. Zabulis,et al., “Bop: Benchmark for 6d object pose estimation,” inProceedings of the European conference on computer vision (ECCV), pp. 19–34, 2018

  16. [16]

    Zs6d: Zero-shot 6d object pose estimation using vision transformers,

    P. Ausserlechner, D. Haberger, S. Thalhammer, J.-B. Weibel, and M. Vincze, “Zs6d: Zero-shot 6d object pose estimation using vision transformers,” in2024 IEEE International Conference on Robotics and Automation (ICRA), pp. 463–469, IEEE, 2024

  17. [17]

    Foundpose: Unseen object pose estimation with foundation features,

    E. P. ¨Ornek, Y . Labb´e, B. Tekin, L. Ma, C. Keskin, C. Forster, and T. Hodan, “Foundpose: Unseen object pose estimation with foundation features,” inEuropean Conference on Computer Vision, pp. 163–182, Springer, 2024

  18. [18]

    Su- perglue: Learning feature matching with graph neural networks,

    P.-E. Sarlin, D. DeTone, T. Malisiewicz, and A. Rabinovich, “Su- perglue: Learning feature matching with graph neural networks,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 4938–4947, 2020

  19. [19]

    Loftr: Detector- free local feature matching with transformers,

    J. Sun, Z. Shen, Y . Wang, H. Bao, and X. Zhou, “Loftr: Detector- free local feature matching with transformers,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 8922–8931, 2021

  20. [20]

    Minima: Modality invariant image matching,

    J. Ren, X. Jiang, Z. Li, D. Liang, X. Zhou, and X. Bai, “Minima: Modality invariant image matching,” inProceedings of the Computer Vision and Pattern Recognition Conference, pp. 23059–23068, 2025

  21. [21]

    Matching 2d images in 3d: Metric relative pose from metric correspondences,

    A. B.-L. Li, V . Prisacariu, and V . Larsson, “Matching 2d images in 3d: Metric relative pose from metric correspondences,”arXiv preprint arXiv:2404.06337, 2024

  22. [22]

    Matchanything: Universal cross-modality image matching with large- scale pre-training,

    X. He, H. Yu, S. Peng, D. Tan, Z. Shen, H. Bao, and X. Zhou, “Matchanything: Universal cross-modality image matching with large- scale pre-training,”arXiv preprint arXiv:2501.07556, 2025

  23. [23]

    Megadepth: Learning single-view depth pre- diction from internet photos,

    Z. Li and N. Snavely, “Megadepth: Learning single-view depth pre- diction from internet photos,” inProceedings of the IEEE conference on computer vision and pattern recognition, pp. 2041–2050, 2018

  24. [24]

    Measurement of areas on a sphere using fibonacci and latitude–longitude lattices,

    ´A. Gonz ´alez, “Measurement of areas on a sphere using fibonacci and latitude–longitude lattices,”Mathematical geosciences, vol. 42, no. 1, pp. 49–64, 2010

  25. [25]

    Cnos: A strong baseline for cad-based novel object segmentation,

    V . N. Nguyen, T. Groueix, G. Ponimatkin, V . Lepetit, and T. Hodan, “Cnos: A strong baseline for cad-based novel object segmentation,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 2134–2140, 2023

  26. [26]

    Lightglue: Local feature matching at light speed,

    P. Lindenberger, P.-E. Sarlin, and M. Pollefeys, “Lightglue: Local feature matching at light speed,” inProceedings of the IEEE/CVF international conference on computer vision, pp. 17627–17638, 2023

  27. [27]

    Pose3rcnn: Monocular 3d object pose estimation via differentiable rendering and point cloud matching,

    A. Kundu, Y . Li, and J. M. Rehg, “Pose3rcnn: Monocular 3d object pose estimation via differentiable rendering and point cloud matching,” arXiv preprint arXiv:2209.12638, 2022

  28. [28]

    Bop: Benchmark for 6d object pose estimation

    T. Hodan, F. Michel, E. Brachmann, W. Kehl, A. G. Buch, D. Kraft, B. Drost, J. Vidal, S. Ihrke, X. Zabulis,et al., “Bop: Benchmark for 6d object pose estimation.”https://bop.felk.cvut.cz, 2018

  29. [29]

    Accelerating 3d deep learning with pytorch3d,

    N. Ravi, J. Reizenstein, D. Novotny, T. Gordon, W.-Y . Lo, J. Johnson, and G. Gkioxari, “Accelerating 3d deep learning with pytorch3d,” arXiv:2007.08501, 2020

  30. [30]

    Model based training, detection and pose estimation of texture-less 3d objects in heavily cluttered scenes,

    S. Hinterstoisser, V . Lepetit, S. Ilic, S. Holzer, G. Bradski, K. Konolige, and N. Navab, “Model based training, detection and pose estimation of texture-less 3d objects in heavily cluttered scenes,” inAsian conference on computer vision, pp. 548–562, Springer, 2012

This paper was first reviewed by deepseek-v4-flash on August 1, 2026.