Pith. sign in

REVIEW 3 major objections 5 minor 42 references

XYZ-IBD: Benchmarking Robust 6D Object Pose Estimation under Real-World Industrial Complexity

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper introduces XYZ-IBD, an industrial bin-picking benchmark whose annotation pipeline claims sub-millimeter pose accuracy, and shows that state-of-the-art 6D pose estimators perform far worse on it than on household benchmarks.

desk verdict A genuinely useful industrial bin-picking benchmark, but the sub-millimeter annotation accuracy claim is not yet supported by the evidence - deserves peer review with a required revision. read the letter →

arxiv 2506.00599 v3 pith:SSPIBCT6 submitted 2025-05-31 cs.CV

Junwen Huang , Jiaqi Hu , Peter KT Yu , Slobodan Ilic , Martin Sundermeyer , Benjamin Busam This is my paper
classification cs.CV
keywords 6Dobjectposeestimationindustrialbin-pickingRGB-DbenchmarkannotationaccuracyspecularobjectssynthetictrainingdataBOPChallengemonoculardepth
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

XYZ-IBD is a new RGB-D benchmark for industrial bin-picking: 15 metallic, texture-less, mostly symmetric objects in densely stacked, highly occluded bins, captured from 50 viewpoints per scene by three cameras, with about 273,000 annotated instances across 75 real scenes and a 45,000-frame synthetic training set. The paper claims that the annotation pipeline — anti-reflection spray, multi-view depth fusion, and iterative closest point refinement — reaches sub-millimeter positional accuracy (mean 0.999 mm) and sub-degree angular accuracy (0.432 degrees), validated by running the same pipeline on simulated scenes with known ground truth. On this benchmark, state-of-the-art pose estimators degrade sharply: the best unseen-object method reaches 0.547 BOP average precision, while seen-object methods trained on the synthetic data perform worse. If the accuracy claim holds, XYZ-IBD gives the field a reliable and hard testbed that measures progress on the problems that actually occur in industrial robotic picking.

What carries the argument

The carrying mechanism is the annotation-and-validation pipeline. An anti-reflection spray suppresses specular highlights so a fused multi-view depth point cloud can be built from 50 calibrated viewpoints; four precision calibration spheres and iterative closest point alignment establish relative camera poses at about 0.248 mm RMSE. Annotators align CAD models on the fused cloud with constrained GUI increments (±1 mm, ±1 degree) followed by multi-scale ICP refinement, then propagate poses to all views. To quantify error, the paper simulates the same capture in a physics-based renderer, tunes a Gaussian depth-noise level sigma = 0.26 mm so that simulated calibration RMSE (0.248 mm) matches the real one, runs the identical annotation pipeline on synthetic scenes with known ground truth, and obtains mean positional error 0.999 mm and angular error 0.432 degrees. That simulated error is the evidence for the sub-millimeter claim.

What would settle it

Take a random subset of real XYZ-IBD scenes and re-annotate them with several independent annotators using the same pipeline, or register the annotated CAD poses against a higher-precision scan of the same bin, for example a coordinate-measuring machine or a higher-resolution laser scanner. If the average disagreement among annotators or with the high-precision scan exceeds about 1 mm, the paper's sub-millimeter annotation accuracy claim is not supported for the real data.

Watch

Extended reading notes

Core claim

The central claim is that existing 6D pose estimation benchmarks are near-saturated on household objects but not on industrial bin-picking, and that XYZ-IBD captures the missing complexity. The paper argues that its dataset combines dense stochastic stacking, repeated instances, severe occlusion, high reflectivity, and industrial-scale object diversity (54–300 mm) with millimeter-accurate annotations. It reports that the strongest generalizable method drops to 0.547 BOP AP, and that monocular depth estimation also falls short of the millimeter-level precision industrial manipulation requires. Consequently, the paper positions XYZ-IBD as reference evaluation data for industrial object pose estimation, including as an official dataset in the BOP Challenge 2025 industrial track and the TRICKY Challenge 2025 monocular depth track.

Load-bearing premise

The claim of sub-millimeter ground truth rests on the assumption that a single tuned Gaussian noise level in simulation (sigma = 0.26 mm) faithfully reproduces all real sensor, calibration, and human annotation errors, since human annotator error is not independently measured on the real dataset.

Editorial extensions

If this is right

  • Seen-object pose estimators trained on XYZ-IBD's synthetic split perform worse than generalizable unseen-object methods, so synthetic-to-real transfer for metallic, symmetric parts remains an open problem.
  • The best reported pose score (0.547 BOP AP) is far below household-benchmark levels, meaning the dataset provides measurable headroom for future industrial pose estimation research.
  • Monocular depth estimation on this data falls short of the millimeter-level accuracy industrial manipulation requires, with Depth Anything V2 reporting an absolute relative error of 3.46 percent and RMSE of 41.8 mm.
  • Because the dataset is included as an official evaluation set in BOP Challenge 2025 and TRICKY Challenge 2025, published leaderboard results will be directly comparable across future methods.

Reading between the lines

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

  • The sub-millimeter accuracy claim is only demonstrated in simulation; human annotation error appears as not available in the paper's error budget. A direct way to test the claim is to have several annotators re-label the same real scenes and measure inter-annotator pose spread, because a spread near or above 1 mm would lower the real dataset's effective accuracy below the simulated 0.999 mm.
  • Because ground-truth poses are annotated from anti-reflection-sprayed depth while the released test images are the raw, reflective captures, part of the benchmark's difficulty may come from a train/eval domain shift between sprayed and raw appearances; this is testable by comparing method performance on sprayed versus raw depth of the same scenes.
  • The high proportion of symmetric objects and roughly 22 instances per image makes XYZ-IBD a natural stress test for symmetry-aware pose metrics, so methods that explicitly model symmetries may show larger gains here than on household datasets.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper introduces XYZ-IBD, a multi-view RGB-D benchmark for industrial bin-picking 6D pose estimation. It provides 75 real-world scenes with roughly 22k frames and 273k annotated instances of 15 metallic, symmetric, and texture-less objects, plus a 45k-frame synthetic training set. The authors describe a pipeline using anti-reflection spray, multi-view depth fusion, and semi-automatic pose annotation, and they claim sub-millimeter positional and sub-degree angular annotation accuracy validated through simulation. They benchmark detection, pose estimation, and depth estimation methods, reporting substantial performance drops (e.g., FoundationPose 0.547 BOP AP) relative to household benchmarks.

Significance. If the annotation accuracy and dataset-scale claims are credible, XYZ-IBD fills a real gap: existing industrial pose datasets lack dense bin-picking clutter, high-reflectivity objects, and multi-instance ambiguity. The dataset is already integrated into the BOP Challenge 2025 industrial track and the TRICKY Challenge 2025 depth track, which is a strong sign of community need. The benchmark numbers for detection, pose, and depth provide a new challenging testbed, and the synthetic training set enables a controlled train/ test protocol. The paper's strengths include a multi-sensor setup with three cameras, public release, and concrete baseline experiments. However, the headline sub-millimeter annotation accuracy claim rests on a simulation whose noise parameter is fitted to a single calibration statistic, and several dataset statistics are internally inconsistent; these issues need to be resolved before the paper can serve as a reliable reference.

major comments (3)
  1. [Section 3.3, Tables 3 and 4] The sub-millimeter annotation accuracy claim is not established. The validation is self-referential: the Gaussian depth noise level σ=0.26 mm is chosen so that the simulated multi-view calibration RMSE (0.248 mm) matches the real-world calibration RMSE (Table 4), and the same simulation is then used to report the mean pose error of 0.999 mm (Table 3 'Overall'). Table 3 explicitly lists 'Depth fusion TSDF' and 'Manually annotate Human, ICP' as N/A, so human annotation error is never quantified on real data. The simulation assumes real sensor noise is homogeneous Gaussian and that manual coarse alignment followed by ICP behaves identically on synthetic and real data; reflective structured-light depth with missing pixels, flying pixels, and anti-reflection spray residue can violate both assumptions, and ICP can converge to different local minima for symmetric parts. Matching a single aggregate RMSE does not validate the full annotation chain. The authors should either provide independent real-world validation (e.g., compare a subset of annotations against a coordinate measuring machine or a high-resolution scanner) or, at minimum, state that the <1 mm value is the simulated pose error under an assumed noise model rather than a measured real-world accuracy. Because 'high-precision' and 'sub-millimeter accuracy' are central contributions, this issue is load-bearing.
  2. [Section 3.2 and Table 1] The dataset scale statistics are internally inconsistent. The text states that 50 viewpoints are sampled per scene and that the dataset contains 75 real-world scenes, which would give 3,750 frames, not 'over 22k frames.' The paper also states 'average of 22 instances per image' and 'approximately 273k annotated instances'; however, 22 × 22k = 484k, not 273k. If instead '22 instances per scene' is intended, then the total is 75 × 22 = 1,650 physical instances, and the per-view annotation count depends on visibility, which is not explained. The authors should provide a precise accounting of the number of views per scene, the number of physical object instances, and how the 273k annotated-instance count is derived. This affects the headline dataset-size claims and the benchmark's credibility.
  3. [Section 4.3, Table 5] The claim that state-of-the-art methods 'degrade sharply' compared to household benchmarks is not supported by any in-paper numerical comparison. The authors report absolute AP values on XYZ-IBD (e.g., FoundationPose 0.547, SurfEmb 0.266) but do not report the same methods' scores on a household benchmark under the same evaluation protocol, despite explicitly comparing to 'existing benchmarks based on household objects' in the text. Without this direct comparison, the degradation claim is an assertion rather than a result. The authors should add a comparative table (e.g., BOP-format AP on YCB-V or T-LESS under identical settings) or qualify the claim as a qualitative observation. This is central to the paper's motivation of showing that existing benchmarks are saturated while industrial scenarios remain unsolved.
minor comments (5)
  1. [Table 2] The row labeled 'PhoCal [19]' cites reference [19] (HAMMER), but PhoCal is reference [24] in the bibliography.
  2. [Table 5] The header 'SAMDGDRNet' appears to be a concatenation of 'SAM-6D' and 'GDRNet'; the column should be labeled 'GDRNet' and the corresponding mAP value (0.296) attributed to GDRNet, not to a combined method.
  3. [Figure 2 caption] 'Anti-reflection Detph' is a typo; it should be 'Anti-reflection Depth.'
  4. [Section 4.1] The headings 'Object 2D Detection Metics' and 'Object 6D Pose Estimation Metics' misspell 'Metrics'; the latter also uses 'model-based 6D object detection' where 'pose estimation' would be clearer.
  5. [Supplementary A.2] The list of supported tasks repeats 'Model-based 2D detection on unseen objects' twice; one entry should likely be 'Model-based 2D detection on seen objects.'

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the sub-millimeter annotation claim rests on a calibrated simulation, not on an equation that reduces to its own inputs.

full rationale

The paper's precision claim is supported by Section 3.3, where Gaussian depth noise (sigma=0.26 mm) is tuned so that the simulated multi-view calibration RMSE (0.248 mm) matches the real-world calibration RMSE. This is a parameter calibration, not a circular reduction: the resulting 6D pose error (0.999 mm positional, 0.432 degrees angular) is a downstream output of a full simulated annotation pipeline (multi-view fusion, manual adjustment, multi-scale ICP) and is numerically distinct from the fitted noise magnitude and the calibration RMSE. Nothing in the derivation defines 'real annotation accuracy' as the simulation output by construction; the simulation is explicitly presented as a validation proxy. Missing quantification of human annotator error and real-sensor non-Gaussianity (Table 3 lists 'Manually annotate Human, ICP' as N/A) is a validity/correctness limitation, not evidence that the claim is equivalent to its inputs. The benchmark results in Section 4 compare external methods on the released data and are self-contained. Self-citations ([20], [24]) are contextual and not load-bearing for the annotation-accuracy argument. Therefore no step in the paper's derivation chain reduces to its own inputs by definition or by fitted-parameter identity.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The benchmark's value stands on the accuracy of pose labels and the realism of the scenes. The headline '<1 mm' accuracy depends on a simulated error model with a noise parameter fitted to one calibration statistic, and human annotation error is unquantified. The CAD-matching and static-scene assumptions are standard for this type of dataset but are not independently verified here.

free parameters (2)
  • Simulated depth noise sigma = 0.26 mm
    Chosen in Section 3.3 so that simulated multi-view calibration RMSE (0.248 mm) matches the real-world calibration RMSE; this noise level drives the simulated pose annotation error of 0.999 mm.
  • Human annotation error = None (treated as negligible)
    Table 3 lists 'Manually annotate Human, ICP N/A', so the simulation implicitly assumes human annotation variability is either zero or exactly reproduced by the automated pipeline, with no independent measurement.
assumptions (5)
  • domain assumption CAD models from industrial partners are micron-accurate and exactly match the physical objects.
    Invoked in Section 3.1 to justify using the same models for real capture and synthetic rendering; if the geometry differs, pose labels and evaluations inherit bias.
  • ad hoc to paper A single Gaussian noise level fitted to the average calibration RMSE captures all relevant sensor, robot, and calibration errors in the simulated validation.
    Section 3.3 uses sigma = 0.26 mm to reproduce the 0.248 mm calibration RMSE; spatial correlation, distortion, and non-Gaussian sensor artifacts are not modeled.
  • domain assumption The anti-reflection spray pass and the raw pass observe the same static object poses.
    Section 3.2 relies on robotic arm repeatability of plus or minus 0.06 mm and static bins to project labels from spray-enhanced depth to raw images.
  • domain assumption Predefined symmetry transformations for each object are correct, so that symmetry-aware evaluation matches the intended annotation pose.
    Section 4.1 uses the symmetry set in MSSD and MSPD metrics; incorrect symmetry definitions would bias both annotation and evaluation.
  • standard math Multi-scale ICP refinement converges to the correct pose from coarse manual initialization.
    Section 3.2 assumes the ICP pipeline reaches the global optimum; local optima remain possible for symmetric or featureless objects.

how reviews work

0 comments
Cite this review

Pith. "Pith review of XYZ-IBD: Benchmarking Robust 6D Object Pose Estimation under Real-World Industrial Complexity." pith.science (2026). https://pith.science/paper/SSPIBCT6

@misc{pith2026250600599,
  author       = {Pith},
  title        = {Pith review of: XYZ-IBD: Benchmarking Robust 6D Object Pose Estimation under Real-World Industrial Complexity},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SSPIBCT6}},
  note         = {Machine review of arXiv:2506.00599}
}
read the original abstract

While current 6D pose estimation benchmarks have reached near-saturation on household objects, they often fail to capture the stochastic and optical complexities of industrial environments. We introduce XYZ-IBD, a high-precision benchmark for object detection and 6D pose estimation specifically designed for industrial bin-picking. XYZ-IBD addresses the domain gap by providing 75 multi-view real-world scenes containing approximately 273k annotated instances of metallic, symmetrical, and specular objects. Unlike existing datasets, our benchmark features high-density stochastic stacking and multi-instance ambiguity, reflecting authentic robotic manipulation challenges. We employ a rigorous multi-stage and semi-automatic annotation pipeline, ensuring sub-millimeter annotation accuracy. The annotations are validated through our designed error quantification scheme, securing the reliability of the annotation quality. In addition to real-world evaluation data, we provide a large-scale complementary synthetic training set that is rendered under a realistic bin-picking simulation. Benchmarking state-of-the-art (SOTA) methods for 2D detection and 6D pose estimation reveals a significant performance degradation compared to standard household benchmarks, highlighting the unsolved challenges of industrial vision. XYZ-IBD establishes a new frontier for robust pose estimation in complex, high-occlusion, and reflective scenarios. The dataset and benchmark are publicly available at https://xyz-ibd.github.io.

Figures

Figures reproduced from arXiv: 2506.00599 by the authors.

Figure 1
Figure 1. Example data from our industrial bin-picking dataset that shows challenging scenes captured by three cameras with different modalities, along with our 6D pose annotations. Raw Image Anti-reflection Image Raw Depth Anti-reflection Detph [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. Example data of the synthetic training data with bin-picking simulation. 1 Introduction The ability to detect, segment, and estimate the 6D pose of objects is critical for robotics applications, particularly in industrial bin-picking scenarios. These tasks demand not only high accuracy but also efficiency to enable real-time operation. While recent advancements in computer vision have signifi￾cantly improved perform… view at source ↗
Figure 4
Figure 4. The real-world industrial data collection pipeline and the annotation error quantification [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figures from the paper (6 more)
Figure 5
Figure 5. Figure 5: (a) shows the data collection setup for the robot arm and sensors; (b) lists the 15 real [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: The collected industrial parts and their corresponding CAD models of XYZ-IBD dataset. [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]
Figure 7
Figure 7. Figure 7: More data samples from the XYZ camera. RGB Depth Pose Lable Instance Mask [PITH_FULL_IMAGE:figures/full_fig_p015_7.png]
Figure 8
Figure 8. Figure 8: More data samples from the RealSense camera. [PITH_FULL_IMAGE:figures/full_fig_p015_8.png]
Figure 9
Figure 9. Figure 9: The instance distribution of the BOP industrial datasets. [PITH_FULL_IMAGE:figures/full_fig_p015_9.png]
Figure 10
Figure 10. Figure 10: Directory structure of the dataset. 16 [PITH_FULL_IMAGE:figures/full_fig_p016_10.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

42 extracted references · 31 canonical work pages

  1. [1]

    Bop challenge 2022 on detection, segmentation and pose estimation of specific rigid objects,

    M. Sundermeyer, T. Hodaˇn, Y . Labbe, G. Wang, E. Brachmann, B. Drost, C. Rother, and J. Matas, “Bop challenge 2022 on detection, segmentation and pose estimation of specific rigid objects,” inCVPR, 2023

    work page 2022

  2. [2]

    Bop challenge 2024 on model-based and model-free 6d object pose estimation,

    N. Van Nguyen, S. Tyree, A. Guo, M. Fourmy, A. Gouda, T. Lee, S. Moon, H. Son, L. Ranftl, J. Tremblay et al., “Bop challenge 2024 on model-based and model-free 6d object pose estimation,”arXiv e-prints, pp. arXiv–2504, 2025

    work page 2024

  3. [3]

    Available: https://bop.felk.cvut.cz/challenges/

    [Online]. Available: https://bop.felk.cvut.cz/challenges/

    work page

  4. [4]

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

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

    work page 2018

  5. [5]

    Homebreweddb: Rgb-d dataset for 6d pose estimation of 3d objects,

    R. Kaskman, S. Zakharov, I. Shugurov, and S. Ilic, “Homebreweddb: Rgb-d dataset for 6d pose estimation of 3d objects,”ICCVW, 2019

    work page 2019

  6. [6]

    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, C. Sahin, F. Manhardt, F. Tombari, T.-K. Kim, J. Matas, and C. Rother, “Bop: Benchmark for 6d object pose estimation,” inECCV, 2018

    work page 2018

  7. [7]

    Recovering 6d object pose and predicting next-best-view in the crowd,

    A. Doumanoglou, R. Kouskouridas, S. Malassiotis, and T.-K. Kim, “Recovering 6d object pose and predicting next-best-view in the crowd,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 3583–3592

    work page 2016

  8. [8]

    T-LESS: An RGB-D dataset for 6D pose estimation of texture-less objects,

    T. Hodaˇn, P. Haluza, Š. Obdržálek, J. Matas, M. Lourakis, and X. Zabulis, “T-LESS: An RGB-D dataset for 6D pose estimation of texture-less objects,”WACV, 2017

    work page 2017

Show all 42 references
  1. [9]

    Learning 6d object pose estimation using 3d object coordinates,

    E. Brachmann, A. Krull, F. Michel, S. Gumhold, J. Shotton, and C. Rother, “Learning 6d object pose estimation using 3d object coordinates,” inECCV, 2014

    2014

  2. [10]

    Foundationpose: Unified 6d pose estimation and tracking of novel objects,

    B. Wen, W. Yang, J. Kautz, and S. Birchfield, “Foundationpose: Unified 6d pose estimation and tracking of novel objects,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 17 868–17 879

    2024

  3. [11]

    Sam-6d: Segment anything model meets zero-shot 6d object pose estimation,

    J. Lin, L. Liu, D. Lu, and K. Jia, “Sam-6d: Segment anything model meets zero-shot 6d object pose estimation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 27 906–27 916

    2024

  4. [12]

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

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

    2021

  5. [13]

    Introducing mvtec itodd - a dataset for 3d object recognition in industry,

    B. Drost, M. Ulrich, P. Bergmann, P. Hartinger, and C. Steger, “Introducing mvtec itodd - a dataset for 3d object recognition in industry,” inICCVW, 2017

    2017

  6. [14]

    Dataset of industrial metal objects,

    P. De Roovere, S. Moonen, N. Michiels, and F. Wyffels, “Dataset of industrial metal objects,”arXiv preprint arXiv:2208.04052, 2022

    2022 arXiv

  7. [15]

    Stereobj-1m: Large-scale stereo image dataset for 6d object pose estimation,

    X. Liu, S. Iwase, and K. M. Kitani, “Stereobj-1m: Large-scale stereo image dataset for 6d object pose estimation,” inProceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 10 870–10 879

    2021

  8. [16]

    Towards co-evaluation of cameras hdr and algorithms for industrial-grade 6dof pose estimation,

    A. Kalra, G. Stoppi, D. Marin, V . Taamazyan, A. Shandilya, R. Agarwal, A. Boykov, T. H. Chong, and M. Stark, “Towards co-evaluation of cameras hdr and algorithms for industrial-grade 6dof pose estimation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Patter...

    2024

  9. [17]

    Blenderproc: Reducing the reality gap with photorealistic rendering,

    M. Denninger, M. Sundermeyer, D. Winkelbauer, D. Olefir, T. Hodan, Y . Zidan, M. Elbadrawy, M. Knauer, H. Katam, and A. Lodhi, “Blenderproc: Reducing the reality gap with photorealistic rendering,” in International Conference on Robotics: Sciene and Systems(RSS), 2020

    2020

  10. [18]

    Robi: A multi-view dataset for reflective objects in robotic bin-picking,

    J. Yang, Y . Gao, D. Li, and S. L. Waslander, “Robi: A multi-view dataset for reflective objects in robotic bin-picking,” in2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2021, pp. 9788–9795

    2021

  11. [19]

    On the importance of accurate geometry data for dense 3d vision tasks,

    H. Jung, P. Ruhkamp, G. Zhai, N. Brasch, Y . Li, Y . Verdie, J. Song, Y . Zhou, A. Armagan, S. Ilicet al., “On the importance of accurate geometry data for dense 3d vision tasks,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 780–791

    2023

  12. [20]

    SCRREAM : SCan, register, REnder and map: A framework for annotating accurate and dense 3d indoor scenes with a benchmark,

    H. Jung, W. Li, S.-C. Wu, W. Bittner, N. Brasch, J. Song, E. Pérez-Pellitero, Z. Zhang, A. Moreau, N. Navab, and B. Busam, “SCRREAM : SCan, register, REnder and map: A framework for annotating accurate and dense 3d indoor scenes with a benchmark,” inThe Thirty-eight Conference...

    2024

  13. [21]

    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. Springer, 2012, pp. 548–562

    2012

  14. [22]

    6-dof pose estimation of household objects for robotic manipulation: An accessible dataset and benchmark,

    S. Tyree, J. Tremblay, T. To, J. Cheng, T. Mosier, J. Smith, and S. Birchfield, “6-dof pose estimation of household objects for robotic manipulation: An accessible dataset and benchmark,” in2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, ...

    2022

  15. [23]

    Normalized object coordinate space for category-level 6d object pose and size estimation,

    H. Wang, S. Sridhar, J. Huang, J. Valentin, S. Song, and L. J. Guibas, “Normalized object coordinate space for category-level 6d object pose and size estimation,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 2642–2651

    2019

  16. [24]

    Phocal: A multi-modal dataset for category-level object pose estimation with photometrically challenging objects,

    P. Wang, H. Jung, Y . Li, S. Shen, R. P. Srikanth, L. Garattoni, S. Meier, N. Navab, and B. Busam, “Phocal: A multi-modal dataset for category-level object pose estimation with photometrically challenging objects,” inProceedings of the IEEE/CVF conference on computer vision an...

    2022

  17. [25]

    Housecat6d–a large-scale multi-modal category level 6d object pose dataset with household objects in realistic scenarios,

    H. Jung, S.-C. Wu, P. Ruhkamp, H. Schieber, P. Wang, G. Rizzoli, H. Zhao, S. D. Meier, D. Roth, N. Navab et al., “Housecat6d–a large-scale multi-modal category level 6d object pose dataset with household objects in realistic scenarios,”arXiv preprint arXiv:2212.10428, 2022

    2022 arXiv

  18. [26]

    Booster: a benchmark for depth from images of specular and transparent surfaces,

    P. Z. Ramirez, A. Costanzino, F. Tosi, M. Poggi, S. Salti, S. Mattoccia, and L. Di Stefano, “Booster: a benchmark for depth from images of specular and transparent surfaces,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 46, no. 1, pp. 85–102, 2023

    2023

  19. [27]

    Hand-eye calibration using dual quaternions,

    K. Daniilidis, “Hand-eye calibration using dual quaternions,”The International Journal of Robotics Research, vol. 18, no. 3, pp. 286–298, 1999

    1999

  20. [28]

    Available: https://sites.google.com/view/iccv25tricky/home

    [Online]. Available: https://sites.google.com/view/iccv25tricky/home

  21. [29]

    Microsoft coco: Common objects in context,

    T.-Y . Lin, M. Maire, S. Belongie, J. Hays, P. Perona, D. Ramanan, P. Dollár, and C. L. Zitnick, “Microsoft coco: Common objects in context,” inComputer vision–ECCV 2014: 13th European conference, zurich, Switzerland, September 6-12, 2014, proceedings, part v 13. Springer, 201...

    2014

  22. [30]

    Open challenges in deep stereo: the booster dataset,

    P. Z. Ramirez, F. Tosi, M. Poggi, S. Salti, S. Mattoccia, and L. Di Stefano, “Open challenges in deep stereo: the booster dataset,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 21 168–21 178

    2022

  23. [31]

    Ntire 2024 challenge on hr depth from images of specular and transparent surfaces,

    P. Z. Ramirez, F. Tosi, L. Di Stefano, R. Timofte, A. Costanzino, M. Poggi, S. Salti, S. Mattoccia, Y . Zhang, C. Wuet al., “Ntire 2024 challenge on hr depth from images of specular and transparent surfaces,” inProceedings of the IEEE/CVF Conference on Computer Vision and Patt...

    2024

  24. [32]

    Yolox: Exceeding yolo series in 2021,

    Z. Ge, S. Liu, F. Wang, Z. Li, and J. Sun, “Yolox: Exceeding yolo series in 2021,”arXiv preprint arXiv:2107.08430, 2021

    2021 arXiv

  25. [33]

    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,” inICCV, 2023

    2023

  26. [34]

    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,” inCVPR, 2022. 11

    2022

  27. [35]

    You only look once: Unified, real-time object detection,

    J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You only look once: Unified, real-time object detection,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 779–788

    2016

  28. [36]

    Segment anything,

    A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.-Y . Loet al., “Segment anything,”arXiv preprint arXiv:2304.02643, 2023

    2023 arXiv

  29. [37]

    Dinov2: Learning robust visual features without supervision,

    M. Oquab, T. Darcet, T. Moutakanni, H. V o, M. Szafraniec, V . Khalidov, P. Fernandez, D. Haziza, F. Massa, A. El-Noubyet al., “Dinov2: Learning robust visual features without supervision,”arXiv preprint arXiv:2304.07193, 2023

    2023 arXiv

  30. [38]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Min- derer, G. Heigold, S. Gellyet al., “An image is worth 16x16 words: Transformers for image recognition at scale,”arXiv preprint arXiv:2010.11929, 2020

    2010 arXiv

  31. [39]

    Geotransformer: Fast and robust point cloud registration with geometric transformer,

    Z. Qin, H. Yu, C. Wang, Y . Guo, Y . Peng, S. Ilic, D. Hu, and K. Xu, “Geotransformer: Fast and robust point cloud registration with geometric transformer,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 45, no. 8, pp. 9806–9821, 2023

    2023

  32. [40]

    Depth anything v2,

    L. Yang, B. Kang, Z. Huang, Z. Zhao, X. Xu, J. Feng, and H. Zhao, “Depth anything v2,”Advances in Neural Information Processing Systems, vol. 37, pp. 21 875–21 911, 2024

    2024

  33. [41]

    Available: https://cmp.felk.cvut.cz/sixd/workshop_2025/

    [Online]. Available: https://cmp.felk.cvut.cz/sixd/workshop_2025/

  34. [42]

    Gdrnpp: A geometry-guided and fully learning-based object pose estimator,

    X. Liu, R. Zhang, C. Zhang, G. Wang, J. Tang, Z. Li, and X. Ji, “Gdrnpp: A geometry-guided and fully learning-based object pose estimator,”IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025. 12 Supplementary Material XYZ-IBD: High-precision Bin-picking Datase...

    2025

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.