Pith. sign in

REVIEW 3 major objections 6 minor 64 references

LiDARTag: A Real-Time Fiducial Tag System for Point Clouds

T0 review · 3 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read LiDARTag makes printed black-and-white tags readable directly from LiDAR point clouds in real time, including in total darkness.

desk verdict First real LiDAR fiducial tag; credible engineering with an open-source release, but the headline accuracy overstates the data and decoding leans on an unquantified intensity assumption. read the letter →

arxiv 1908.10349 v3 pith:RVXNLW47 submitted 2019-08-23 cs.RO cs.CVcs.LG

Jiunn-Kai Huang , Shoutian Wang , Maani Ghaffari , Jessy W. Grizzle This is my paper · ORCID
classification cs.ROcs.CVcs.LG
keywords fiducialmarkersLiDARpointcloudstagdetectionposeestimationreproducingkernelHilbertspaceIDdecodingintensityreturnsreal-timerobotics
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 tries to establish that a printed black-and-white landmark can work as a fiducial tag directly in LiDAR point clouds, without needing a camera. It introduces LiDARTag, a complete pipeline that detects such tags, estimates their pose, and reads their unique ID, running at over 100 Hz on a laptop-class processor. Because LiDAR supplies its own light, the tag would remain detectable in complete darkness and under rapid lighting changes, solving a weakness of camera-based tags like AprilTags. The reported experiments claim millimeter-level translation error, a few degrees of rotation error against motion-capture ground truth, and zero false positives in more than 379,000 scans of public indoor and outdoor datasets. If these claims hold, robots can use the same physical tags for dark-robust localization, multi-sensor calibration, and SLAM landmarks.

What carries the argument

The load-bearing machinery is a three-stage pipeline. Candidate features are edge points found by a distance-gradient operator $\nabla D(p_{i,m})=\|p_{i+l,m}-p_{i,m}\|_2-\|p_{i-l,m}-p_{i,m}\|_2$; these are clustered using signed Manhattan distance in the horizontal plane and ring number vertically, then validated by point-count, payload-edge, and plane-fitting heuristics. Pose estimation minimizes the cost $C(H_T^L(TP))=\sum_i c(\bar{x}_i,\epsilon)+c(\bar{y}_i,d/2)+c(\bar{z}_i,d/2)$, where $c(\lambda,a)$ penalizes points lying outside the template bounds, and the optimization is initialized by a Procrustes problem solved from RANSAC-fitted tag edges. ID decoding lifts the projected points to a continuous function $f(\cdot)=\sum_i \ell(\tilde{p}_i) k(\cdot,\tilde{p}_i)$ in a reproducing kernel Hilbert space, a function space where scattered point sets become smooth functions and inner products measure similarity, using the squared-exponential kernel $k(p_i,p_j)=\sigma^2\exp(-\frac12(p_i-p_j)^T\Lambda(p_i-p_j))$, and compares $f$ by inner product against a precomputed dictionary of tag functions.

What would settle it

Print a LiDARTag, place it in a dark room, and move a LiDAR across distances of 2 to 16 meters and angles up to 45 degrees while measuring the intensity contrast between black and white squares; if decoding accuracy falls as the contrast drops or if a LiDAR with less stable intensity produces wrong IDs, the reliance on stable intensity readings is falsified.

Watch

Extended reading notes

Core claim

The central claim is that a fiducial marker made of black and white squares can be detected, localized, and identified from the sparse, unstructured returns of a LiDAR, using only geometry and intensity. Detection begins with distance-gradient edge points, which are clustered by Manhattan distance and ring number and then validated with tag-family heuristics and a plane-fitting outlier check. Pose is estimated by minimizing an L1-inspired fitting error between the back-projected point cloud and a template of known geometry, initialized by a Procrustes alignment of estimated corners. To decode the ID despite sparse points, the projected point cloud is lifted to a continuous function in a reproducing kernel Hilbert space with a squared-exponential kernel, and the tag's ID is chosen as the dictionary function with the largest inner product; this also settles the 90-degree rotation ambiguity left by pose fitting. The paper reports that this system achieves millimeter and few-degree pose accuracy, runs above 100 Hz, and produces no false positives over the tested public datasets.

Load-bearing premise

The load-bearing premise is that a LiDAR's intensity readings cleanly and consistently separate printed black from printed white across the distances and angles where the tag is used; the paper states this requirement but gives no quantitative characterization of intensity repeatability.

Editorial extensions

If this is right

  • A robot can find and identify a tagged object in total darkness, because detection relies on the LiDAR's own light rather than ambient illumination.
  • The same printed tag can serve both camera and LiDAR pipelines, making LiDAR-camera calibration and multi-sensor fusion easier.
  • The detector can run at 100 Hz, faster than current LiDAR frame rates, so it will not bottleneck a robot's perception pipeline.
  • The reported zero false positives on large indoor and outdoor datasets supports using LiDARTags in cluttered and crowded scenes.
  • One detector can handle tags of different physical sizes in the same scene, unlike the single-size assumption common in camera fiducial systems.

Reading between the lines

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

  • Because detection uses geometric edge and plane cues rather than learned categories, the same pipeline should generalize to other planar patterns with sufficient intensity contrast; this is an untested extension.
  • The dictionary inner-product decoder could in principle recognize partially occluded tags if the kernel were spatially localized, but the paper does not test occlusion.
  • The main assumption, stable intensity, could be turned into a calibration benchmark: a LiDARTag could serve as a standard target for measuring intensity repeatability across distance and angle.
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

3 major / 6 minor

Summary. The paper introduces LiDARTag, a fiducial marker system designed for LiDAR point clouds. The tag is an AprilTag-derived planar marker printed on a rigid object, and the system detects potential tags via distance-gradient features and clustering, estimates pose by minimizing an L1-inspired fitting error between the point cloud and a template, and decodes the tag ID by lifting the sparse projected point cloud to a continuous function in an RKHS with intensity as the label. Experiments with a 32-beam Velodyne ULTRA Puck, motion-capture ground truth, and two public datasets (Google Cartographer, Honda H3D) are used to support claims of real-time performance (above 100 Hz), millimeter/degree-level pose accuracy, 99.7% ID decoding accuracy, and zero false positives in over 379,000 scans. The implementation is released in C++/ROS.

Significance. If the claims hold, LiDARTag is the first practical fiducial tag system for LiDAR point clouds, and the paper's open-source implementation plus validation against external motion capture and large public datasets would make it a valuable building block for calibration, multi-sensor fusion, SLAM loop closures, and lighting-invariant landmark tracking. The pose-estimation method and RKHS decoding stage are clearly described and built on the authors' prior work, but the external validation gives the core system credibility beyond self-consistency.

major comments (3)
  1. [Abstract; Section VII-A, Table I] The abstract's claim of 'millimeter error in translation and a few degrees in rotation' is not supported across the reported operating range. Table I shows face-on translation errors of 10.13 mm at 4.29 m and 16.23 mm at 5.90 m, and rotation errors of 10.48 degrees at 13.87 m face-on and 15.92 degrees at 14.08 m rotated. The mean face-on translation error is 6.891 mm, which is not millimeter-level. Please qualify the accuracy statement to the range where it holds, report per-distance conditions and error distributions, or provide a tolerance-based reliability metric so that the user-facing claims match the evidence.
  2. [Section VI, Eq. (21); Section I; Table I] The ID-decoding stage relies on intensity being a stable proxy for printed black-and-white reflectivity, as acknowledged in Section I ('LiDARs with stable (good) intensity readings are required'), but the paper provides no quantitative characterization of intensity repeatability across distance, incidence angle, or sensor model. The intensity kernel in Eq. (21) with length-scale l_I=10 can separate black and white payload cells only if the intensity distributions have a sufficient margin; the wrong-ID events at 13.87 m (1 of 35 scans) and 14.08 m (2 of 49 scans) indicate that this margin degrades at range. Please report intensity histograms or class-separation margins for the tested distances and angles, and analyze how the margin relates to decoding failures.
  3. [Remark 8; Section VII-C, Table V] Because decoding failure is used as a validity gate for clusters in Remark 8, the zero-false-positive results on the Google Cartographer and Honda H3D datasets are coupled to the intensity-based decoding behavior of Section VI. If the intensity statistics of those public datasets happen to be benign, the evaluation may not expose failures caused by low or unstable intensity contrast between black and white regions. The paper should either include stress tests with simulated or real adversarial intensity statistics, or explicitly state that false-positive rejection was validated only under the intensity conditions present in those datasets.
minor comments (6)
  1. [Section III-B] The requirement that the first LiDAR ring hitting the tag be above 3/4 of the tag is stated without experimental validation; please explain how this condition was enforced in the data collection and how much performance degrades when it is violated.
  2. [Section IV-B, Eq. (2)] The clearance assumption τ = t√2/4 is a practical constraint that may cause false linkages in cluttered scenes; please discuss how restrictive this is and whether the clustering is robust when the clearance is not met.
  3. [Remark 7, Eq. (22)] The intensity-to-depth scaling t/(2(d+4)Imax) in Eq. (22) is introduced without derivation or citation; please justify the factor of two and explain how Imax is estimated in practice.
  4. [Section VII-B, Table II] Table II reports 'Fill In Clusters' as 0.00 ms, which seems inconsistent with a nontrivial pipeline step; please either report the actual cost or remove the column to avoid confusing readers.
  5. [Section VII-A] The text says a 1.2-m target placed at 16 meters and rotated by 45 degrees is the detection limit, but the furthest distance in Table I is about 14 meters; please reconcile these statements.
  6. [Table I] In the summary rows, 'No. Scans' appears to be the mean number of scans at each distance rather than the total; please define the quantity explicitly so the wrong-ID ratio is not misinterpreted.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: pose, decoding, and false-positive claims are validated against external motion capture and public datasets; self-citations are methodological and not load-bearing.

full rationale

The central derivation chain is not circular. The pose is estimated by minimizing an L1-inspired cost (Eqs. 7-9) against a known-geometry template, and the claimed millimeter/degree accuracy is then compared to a 30-camera motion capture system (Table I), not to the cost itself. The ID decoding lifts the point cloud to an RKHS function and takes the largest inner product against a pre-computed tag dictionary (Eqs. 17-21); this is an empirical template-matching rule whose success is scored against known IDs (Table I), so it does not reduce to a fit. The zero false-positive claim is checked on the Google Cartographer and Honda H3D public datasets, which contain no tags, making the rejection external. Fixed design parameters such as l_I=10, kappa=0.05, and the feature thresholds are not fitted to the reported outcomes. Self-citations are present--[8] for the L1-inspired pose cost and [43],[44] for RKHS-inspired function construction--but they are methodological references, not uniqueness theorems, and the paper's conclusions are not derived from those citations alone. The admitted intensity-stability requirement (Sec. I) is a limitation and correctness risk, not a circularity: no quantity used in the derivation is defined in terms of the target result. No circular step can be exhibited with the required specificity, so the appropriate finding is low circularity burden.

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

The central claim rests on stable LiDAR intensity, known planar tag geometry, standard kernel-method mathematics, and the validity of negative test datasets. The main free parameters are detection thresholds and kernel hyperparameters chosen by hand; none are fitted to the reported accuracy numbers, so the system is not tuning to its own benchmarks, but sensitivity to these choices is not characterized.

free parameters (6)
  • Edge gradient threshold ζ = not reported
    In Section IV-A, a point is an edge point if ∇D exceeds ζ, but no numeric value is given; it is a tuned detection threshold affecting cluster extraction.
  • Signal variance σ² in kernel = 1e5
    Set by hand in equation (19); not fitted but chosen without sensitivity analysis.
  • Intensity length-scale l_I = 10
    Set by hand in equation (21) for the intensity kernel; affects ID decoding similarity.
  • Plane fitting outlier threshold κ = 0.05
    Chosen in Section IV-C to reject non-planar clusters; no sensitivity study.
  • Minimum returns per bit heuristic = 5 per bit
    Equation (3) assumes at least five LiDAR returns per payload bit; this is a heuristic lower bound used to filter clusters.
  • First-ring placement fraction = 3/4 of tag
    Section IV-B requires the first beam hitting the tag to be above 3/4 of the target; this placement rule is a design assumption.
assumptions (6)
  • domain assumption LiDAR intensity measurements are stable and reflect the printed reflectivity of surfaces well enough to distinguish black and white payload bits.
    Stated in Section I and used throughout Section VI for ID decoding; the paper does not quantify intensity repeatability.
  • domain assumption The tag is planar and rigidly attached to a 3D object with known geometry, and the template has known dimensions.
    Assumed in Section V for the L1-inspired pose fitting and template definition.
  • standard math The squared exponential kernel is positive definite and the RKHS inner product is a valid similarity measure for intensity-labeled point sets.
    From [54] and [43], used in Section VI; standard kernel-methods background.
  • standard math The AprilTag lexicode family provides sufficient Hamming distance to distinguish tags and reject false positives.
    Coding theory from [3], [20], used to select tag16h6c5 in Section III-B.
  • domain assumption The motion capture system provides accurate ground truth for pose errors.
    Section VII-A treats 30-camera Qualisys system as ground truth without stated accuracy bounds.
  • domain assumption The public Cartographer and H3D datasets contain no LiDARTags, so any detection is a false positive.
    Section VII-C; this is a reasonable negative-set assumption but cannot guarantee absence of patterns resembling tags.

how reviews work

0 comments
Cite this review

Pith. "Pith review of LiDARTag: A Real-Time Fiducial Tag System for Point Clouds." pith.science (2026). https://pith.science/paper/RVXNLW47

@misc{pith2026190810349,
  author       = {Pith},
  title        = {Pith review of: LiDARTag: A Real-Time Fiducial Tag System for Point Clouds},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RVXNLW47}},
  note         = {Machine review of arXiv:1908.10349}
}
read the original abstract

Image-based fiducial markers are useful in problems such as object tracking in cluttered or textureless environments, camera (and multi-sensor) calibration tasks, and vision-based simultaneous localization and mapping (SLAM). The state-of-the-art fiducial marker detection algorithms rely on the consistency of the ambient lighting. This paper introduces LiDARTag, a novel fiducial tag design and detection algorithm suitable for light detection and ranging (LiDAR) point clouds. The proposed method runs in real-time and can process data at 100 Hz, which is faster than the currently available LiDAR sensor frequencies. Because of the LiDAR sensors' nature, rapidly changing ambient lighting will not affect the detection of a LiDARTag; hence, the proposed fiducial marker can operate in a completely dark environment. In addition, the LiDARTag nicely complements and is compatible with existing visual fiducial markers, such as AprilTags, allowing for efficient multi-sensor fusion and calibration tasks. We further propose a concept of minimizing a fitting error between a point cloud and the marker's template to estimate the marker's pose. The proposed method achieves millimeter error in translation and a few degrees in rotation. Due to LiDAR returns' sparsity, the point cloud is lifted to a continuous function in a reproducing kernel Hilbert space where the inner product can be used to determine a marker's ID. The experimental results, verified by a motion capture system, confirm that the proposed method can reliably provide a tag's pose and unique ID code. The rejection of false positives is validated on the Google Cartographer indoor dataset and the Honda H3D outdoor dataset. All implementations are coded in C++ and are available at: https://github.com/UMich-BipedLab/LiDARTag.

Figures

Figures reproduced from arXiv: 1908.10349 by the authors.

Figure 1
Figure 1. LiDAR-based markers can be used in tandem with camera-based [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. This figure illustrates the unstructured nature of a LiDAR-point-cloud [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. (a) illustrates a LiDARTag consisting of two parts: a 3D object with a rigidly attached, planar fiducial marker where t and h are the marker size and height of the object respectively. (b) shows the marker should be placed inside the yellow region and (c) illustrates an example of AprilTag being used as a LiDARTag. (d) is the sensor setup consisting of a LiDAR, a camera and several motion capture markers. and tag25h… view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: The system contains three parts: tag detection, pose estimation, and tag decoding. The detection step takes an entire LiDAR scan (up to 120,000 points [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 6
Figure 6. Figure 6: This figure shows the initial state of a cluster, in which has only single [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]
Figure 5
Figure 5. Figure 5: Intermediate steps of the LiDARTag system. The system takes a full [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 7
Figure 7. Figure 7: (a) describes the coordinate system of the fiducial marker. (b) indicates that the first beam hitting the LiDARTag should be at least 3/4 above target, outlined as the red region. C. Cluster Validation At this point, we have grouped the features into clusters as shown …
Figure 9
Figure 9. Figure 9: Before decoding, the estimated rotation about the normal axis is only [PITH_FULL_IMAGE:figures/full_fig_p007_9.png]
Figure 10
Figure 10. Figure 10: (a) and (c) are images of the tag placed at 2 and 14 meters away from a Cassie-series robot. (b) and (d) describe the results of projecting the template (green box) from the LiDAR origin to the tag’s returns by the poses of LiDARTag at 2 and 16 meters, respectively. W…
Figure 11
Figure 11. Figure 11: (a) and (c) image a 0.8 and a 0.6 meter tag placed in a cluttered indoor laboratory and a spacious outdoor environment. (b) and (d) show the algorithm successfully detects the two markers of different sizes indicated by cyan boxes. TABLE II: This table averages all th…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

64 extracted references · 61 canonical work pages

  1. [1]

    Artoolkit on the pocketpc platform,

    D. Wagner and D. Schmalstieg, “Artoolkit on the pocketpc platform,” in IEEE International Augmented Reality Toolkit Workshop . IEEE, 2003, pp. 14–15

    work page 2003

  2. [2]

    Artag, a fiducial marker system using digital techniques,

    M. Fiala, “Artag, a fiducial marker system using digital techniques,” in Proc. IEEE Conf. Comput. Vis. Pattern Recog., vol. 2. IEEE, 2005, pp. 590–596

    work page 2005

  3. [3]

    AprilTag: A robust and flexible visual fiducial system,

    E. Olson, “AprilTag: A robust and flexible visual fiducial system,” in Proc. IEEE Int. Conf. Robot. and Automation . IEEE, 2011, pp. 3400– 3407

    work page 2011

  4. [4]

    Apriltag 2: Efficient and robust fiducial detec- tion,

    J. Wang and E. Olson, “Apriltag 2: Efficient and robust fiducial detec- tion,” in Proc. IEEE/RSJ Int. Conf. Intell. Robots and Syst. IEEE, 2016, pp. 4193–4198

    work page 2016

  5. [5]

    Flexible layouts for fiducial tags,

    M. Krogius, A. Haggenmiller, and E. Olson, “Flexible layouts for fiducial tags,” in Proc. IEEE/RSJ Int. Conf. Intell. Robots and Syst. IEEE, 2019, pp. 1898–1903

    work page 2019

  6. [6]

    Chromatag: a colored marker and fast detection algorithm,

    J. DeGol, T. Bretl, and D. Hoiem, “Chromatag: a colored marker and fast detection algorithm,” in Proc. IEEE Conf. Comput. Vis. Pattern Recog. , 2017, pp. 1472–1481

    work page 2017

  7. [7]

    Improved structure from motion using fiducial marker matching,

    ——, “Improved structure from motion using fiducial marker matching,” in Proc. European Conf. Comput. Vis. , 2018, pp. 273–288

    work page 2018

  8. [8]

    Improvements to target-based 3D LiDAR to camera calibration,

    J. Huang and J. W. Grizzle, “Improvements to target-based 3D LiDAR to camera calibration,” IEEE Access, vol. 8, pp. 134 101–134 110, 2020

    work page 2020

Show all 64 references
  1. [9]

    Continuous direct sparse visual odometry from RGB-D images,

    M. Ghaffari, W. Clark, A. Bloch, R. M. Eustice, and J. W. Grizzle, “Continuous direct sparse visual odometry from RGB-D images,” in Proc. Robot.: Sci. Syst. Conf. , Freiburg, Germany, June 2019

    2019

  2. [10]

    Extrinsic LiDAR Camera Calibration,

    J.K. Huang and Jessy W. Grizzle, “Extrinsic LiDAR Camera Calibration,” 2019. [Online]. Available: https://github.com/ UMich-BipedLab/extrinsic_lidar_camera_calibration

    2019

  3. [11]

    Real-time loop closure in 2D LIDAR SLAM,

    W. Hess, D. Kohler, H. Rapp, and D. Andor, “Real-time loop closure in 2D LIDAR SLAM,” in Proc. IEEE Int. Conf. Robot. and Automation . IEEE, 2016, pp. 1271–1278

    2016

  4. [12]

    The H3D dataset for full-surround 3D multi-object detection and tracking in crowded urban scenes,

    A. Patil, S. Malla, H. Gang, and Y .-T. Chen, “The H3D dataset for full-surround 3D multi-object detection and tracking in crowded urban scenes,” in Proc. IEEE Int. Conf. Robot. and Automation , 2019

    2019

  5. [13]

    ROS: an open-source Robot Operating System,

    M. Quigley, K. Conley, B. Gerkey, J. Faust, T. Foote, J. Leibs, R. Wheeler, and A. Y . Ng, “ROS: an open-source Robot Operating System,” in ICRA workshop on open source software , 2009

    2009

  6. [14]

    LiDARTag: A Real-Time and Flexible Fiducial Tag for Point Clouds,

    J.K. Huang and Jessy W. Grizzle, “LiDARTag: A Real-Time and Flexible Fiducial Tag for Point Clouds,” 2020. [Online]. Available: https://github.com/UMich-BipedLab/LiDARTag

    2020

  7. [15]

    Automatic reconstruction of wide- area fiducial marker models,

    M. Klopschitz and D. Schmalstieg, “Automatic reconstruction of wide- area fiducial marker models,” in Proc. Int. Symposium on Mixed and Augmented Reality. IEEE, 2007, pp. 71–74

    2007

  8. [16]

    Calibration of RGB camera with velodyne LiDAR

    M. Velas, M. Spanel, Z. Materna, and A. Herout, “Calibration of RGB camera with velodyne LiDAR.” Václav Skala-UNION Agency, 2014

    2014

  9. [17]

    Calibration between color camera and 3D LIDAR instruments with a polygonal planar board,

    Y . Park, S. Yun, C. S. Won, K. Cho, K. Um, and S. Sim, “Calibration between color camera and 3D LIDAR instruments with a polygonal planar board,” Sensors, vol. 14, no. 3, pp. 5333–5353, 2014

    2014

  10. [18]

    Caltag: High precision fiducial markers for camera calibration

    B. Atcheson, F. Heide, and W. Heidrich, “Caltag: High precision fiducial markers for camera calibration.” in Vision, Modeling, and Visualization, vol. 10. Citeseer, 2010, pp. 41–48

    2010

  11. [19]

    Comparing artag and artoolkit plus fiducial marker systems,

    M. Fiala, “Comparing artag and artoolkit plus fiducial marker systems,” in Int. Workshop on Haptic Audio Visual Environments and their Applications. IEEE, 2005, pp. 6–pp

    2005

  12. [20]

    Computational methods in coding theory,

    A. Trachtenbert, “Computational methods in coding theory,” Master’s thesis, University of Illinois at Urbana-Champaign, 1996

    1996

  13. [21]

    Rune-tag: A high accuracy fiducial marker with strong occlusion resilience,

    F. Bergamasco, A. Albarelli, E. Rodola, and A. Torsello, “Rune-tag: A high accuracy fiducial marker with strong occlusion resilience,” in Proc. IEEE Conf. Comput. Vis. Pattern Recog. IEEE, 2011, pp. 113–120

    2011

  14. [22]

    Detection and accurate localization of circular fiducials under highly challenging conditions,

    L. Calvet, P. Gurdjos, C. Griwodz, and S. Gasparini, “Detection and accurate localization of circular fiducials under highly challenging conditions,” in Proc. IEEE Conf. Comput. Vis. Pattern Recog. , 2016, pp. 562–570

    2016

  15. [23]

    LFTag: A scalable visual fiducial system with low spatial frequency,

    B. Wang, “LFTag: A scalable visual fiducial system with low spatial frequency,” arXiv preprint arXiv:2006.00842 , 2020

    2006 arXiv

  16. [24]

    Learnable visual markers,

    O. Grinchuk, V . Lebedev, and V . Lempitsky, “Learnable visual markers,” in Proc. Advances Neural Inform. Process. Syst. Conf. , 2016, pp. 4143– 4151

    2016

  17. [25]

    Deep charuco: Dark charuco marker pose estimation,

    D. Hu, D. DeTone, and T. Malisiewicz, “Deep charuco: Dark charuco marker pose estimation,” in Proc. IEEE Conf. Comput. Vis. Pattern Recog., 2019, pp. 8436–8444

    2019

  18. [26]

    Sliding shapes for 3d object detection in depth images,

    S. Song and J. Xiao, “Sliding shapes for 3d object detection in depth images,” in Proc. European Conf. Comput. Vis. Springer, 2014, pp. 634–651

    2014

  19. [27]

    V ote3deep: Fast object detection in 3D point clouds using efficient convolutional neural networks,

    M. Engelcke, D. Rao, D. Z. Wang, C. H. Tong, and I. Posner, “V ote3deep: Fast object detection in 3D point clouds using efficient convolutional neural networks,” in Proc. IEEE Int. Conf. Robot. and Automation. IEEE, 2017, pp. 1355–1361

    2017

  20. [28]

    V oxelnet: End-to-end learning for point cloud based 3D object detection,

    S. Song and J. Xiao, “V oxelnet: End-to-end learning for point cloud based 3D object detection,” in Proc. IEEE Conf. Comput. Vis. Pattern Recog., 2018, pp. 4490–4499. 11

    2018

  21. [29]

    Deep sliding shapes for amodal 3D object detection in RGB-D images,

    ——, “Deep sliding shapes for amodal 3D object detection in RGB-D images,” in Proc. IEEE Conf. Comput. Vis. Pattern Recog. , 2016, pp. 808–816

    2016

  22. [30]

    3D fully convolutional network for vehicle detection in point cloud,

    B. Li, “3D fully convolutional network for vehicle detection in point cloud,” in Proc. IEEE/RSJ Int. Conf. Intell. Robots and Syst. IEEE, 2017, pp. 1513–1518

    2017

  23. [31]

    Edge boxes: Locating object proposals from edges,

    C. L. Zitnick and P. Dollár, “Edge boxes: Locating object proposals from edges,” in Proc. European Conf. Comput. Vis. Springer, 2014, pp. 391–405

    2014

  24. [32]

    Segmentation as selective search for object recognition

    K. E. Van de Sande, J. R. Uijlings, T. Gevers, A. W. Smeulders, et al., “Segmentation as selective search for object recognition.” in Proc. IEEE Int. Conf. Comput. Vis. , vol. 1, no. 2, 2011, p. 7

    2011

  25. [33]

    Cpmc: Automatic object segmentation using constrained parametric min-cuts,

    J. Carreira and C. Sminchisescu, “Cpmc: Automatic object segmentation using constrained parametric min-cuts,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 34, no. 7, pp. 1312–1328, 2011

    2011

  26. [34]

    3D IoU-Net: IoU guided 3D object detector for point clouds,

    J. Li, S. Luo, Z. Zhu, H. Dai, A. S. Krylov, Y . Ding, and L. Shao, “3D IoU-Net: IoU guided 3D object detector for point clouds,” arXiv preprint arXiv:2004.04962, 2020

    2004 arXiv

  27. [35]

    Pointrcnn: 3D object proposal generation and detection from point cloud,

    S. Shi, X. Wang, and H. Li, “Pointrcnn: 3D object proposal generation and detection from point cloud,” in Proc. IEEE Conf. Comput. Vis. Pattern Recog., 2019, pp. 770–779

    2019

  28. [36]

    Vision meets robotics: The KITTI dataset,

    A. Geiger, P. Lenz, C. Stiller, and R. Urtasun, “Vision meets robotics: The KITTI dataset,” Int. J. Robot. Res. , 2013

    2013

  29. [37]

    SEMANTIC3D NET: A new large scale point cloud classification benchmark,

    T. Hackel, N. Savinov, L. Ladicky, J. D. Wegner, K. Schindler, and M. Pollefeys, “SEMANTIC3D NET: A new large scale point cloud classification benchmark,” in ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences , 2017, pp. 91–98

    2017

  30. [38]

    Pedx: Bench- mark dataset for metric 3-D pose estimation of pedestrians in complex urban intersections,

    W. Kim, M. S. Ramanagopal, C. Barto, M.-Y . Yu, K. Rosaen, N. Goumas, R. Vasudevan, and M. Johnson-Roberson, “Pedx: Bench- mark dataset for metric 3-D pose estimation of pedestrians in complex urban intersections,” IEEE Robotics and Automation Letters , vol. 4, no. 2, pp. 1940...

    1940

  31. [39]

    Velodyne Ultra Puck: VLP-32C User Manual,

    Velodyne Lidar, “Velodyne Ultra Puck: VLP-32C User Manual,”

  32. [40]

    Circular data matrix fiducial system and robust image processing for a wearable vision-inertial self-tracker,

    L. Naimark and E. Foxlin, “Circular data matrix fiducial system and robust image processing for a wearable vision-inertial self-tracker,” in Proc. Int. Symposium on Mixed and Augmented Reality . IEEE Computer Society, 2002, p. 27

    2002

  33. [41]

    Cybercode: designing augmented reality environments with visual tags,

    J. Rekimoto and Y . Ayatsuka, “Cybercode: designing augmented reality environments with visual tags,” in Proc. Designing augmented reality environments. ACM, 2000, pp. 1–10

    2000

  34. [42]

    A multi-ring color fiducial system and an intensity-invariant detection method for scalable fiducial-tracking augmented reality,

    Y . Cho, J. Lee, and U. Neumann, “A multi-ring color fiducial system and an intensity-invariant detection method for scalable fiducial-tracking augmented reality,” in Proc. Int. Workshop on Augmented Reality . Citeseer, 1998

    1998

  35. [43]

    Nonparametric continuous sensor registration,

    W. Clark, M. Ghaffari, and A. Bloch, “Nonparametric continuous sensor registration,” arXiv preprint arXiv:2001.04286 , 2020

    2001 arXiv

  36. [44]

    Continuous direct sparse visual odometry from rgb-d images,

    M. Ghaffari, W. Clark, A. Bloch, R. M. Eustice, and J. W. Grizzle, “Continuous direct sparse visual odometry from rgb-d images,” arXiv preprint arXiv:1904.02266, 2019

    1904 arXiv

  37. [45]

    A computational approach to edge detection,

    J. Canny, “A computational approach to edge detection,” in Readings in computer vision. Elsevier, 1987, pp. 184–203

    1987

  38. [46]

    LeGO-LOAM: Lightweight and ground- optimized lidar odometry and mapping on variable terrain,

    T. Shan and B. Englot, “LeGO-LOAM: Lightweight and ground- optimized lidar odometry and mapping on variable terrain,” in Proc. IEEE/RSJ Int. Conf. Intell. Robots and Syst. IEEE, 2018, pp. 4758– 4765

    2018

  39. [47]

    Hierarchical clustering schemes,

    S. C. Johnson, “Hierarchical clustering schemes,” Psychometrika, vol. 32, no. 3, pp. 241–254, 1967

    1967

  40. [48]

    Least squares quantization in PCM,

    S. Lloyd, “Least squares quantization in PCM,” IEEE Trans. Inf. Theory, vol. 28, no. 2, pp. 129–137, 1982

    1982

  41. [49]

    Multidimensional binary search trees used for associative searching,

    J. L. Bentley, “Multidimensional binary search trees used for associative searching,” Communications of the ACM , vol. 18, no. 9, pp. 509–517, 1975

    1975

  42. [50]

    The nlopt nonlinear-optimization package,

    S. G. Johnson, “The nlopt nonlinear-optimization package,” 2014. [Online]. Available: https://github.com/stevengj/nlopt

    2014

  43. [51]

    The surveyor’s area formula,

    B. Braden, “The surveyor’s area formula,” The College Mathematics Journal, vol. 17, no. 4, pp. 326–337, 1986

    1986

  44. [52]

    Generalized procrustes analysis,

    J. C. Gower, “Generalized procrustes analysis,” Psychometrika, vol. 40, no. 1, pp. 33–51, 1975

    1975

  45. [53]

    Principal components analysis corrects for strat- ification in genome-wide association studies,

    A. L. Price, N. J. Patterson, R. M. Plenge, M. E. Weinblatt, N. A. Shadick, and D. Reich, “Principal components analysis corrects for strat- ification in genome-wide association studies,” Nature genetics, vol. 38, no. 8, p. 904, 2006

    2006

  46. [54]

    Rasmussen and C

    C. Rasmussen and C. Williams, Gaussian processes for machine learn- ing. MIT press, 2006, vol. 1

    2006

  47. [55]

    A generalized representer theorem,

    B. Schölkopf, R. Herbrich, and A. J. Smola, “A generalized representer theorem,” inInt. Conf. Computational Learning Theory. Springer, 2001, pp. 416–426

    2001

  48. [56]

    Wahba, Spline models for observational data

    G. Wahba, Spline models for observational data . SIAM, 1990

    1990

  49. [57]

    Some results on Tchebycheffian spline functions,

    G. Kimeldorf and G. Wahba, “Some results on Tchebycheffian spline functions,” vol. 33, no. 1, pp. 82–95, 1971

    1971

  50. [58]

    M-Air at the University of Michigan, Ann Arbor,

    “M-Air at the University of Michigan, Ann Arbor,” 2018. [Online]. Available: https://robotics.umich.edu/about/mair/

    2018

  51. [59]

    A class of globally convergent optimization methods based on conservative convex separable approximations,

    K. Svanberg, “A class of globally convergent optimization methods based on conservative convex separable approximations,” SIAM J. on optimization, vol. 12, no. 2, pp. 555–573, 2002

    2002

  52. [60]

    The method of moving asymptotes: a new method for structural optimization,

    ——, “The method of moving asymptotes: a new method for structural optimization,” Int. J. for numerical methods in engineering , vol. 24, no. 2, pp. 359–373, 1987

    1987

  53. [61]

    Metrics for 3D rotations: Comparison and analysis,

    D. Q. Huynh, “Metrics for 3D rotations: Comparison and analysis,” J. Math. Imaging and Vis. , vol. 35, no. 2, pp. 155–164, 2009

    2009

  54. [62]

    Padua, Ed., TBB (Intel Threading Building Blocks)

    D. Padua, Ed., TBB (Intel Threading Building Blocks) . Boston, MA: Springer US, 2011, pp. 2029–2029. [Online]. Available: https://doi.org/10.1007/978-0-387-09766-4_2080

    2011 doi

  55. [63]

    nanoflann: a C++ header-only fork of FLANN, a library for nearest neighbor (NN) with kd-trees,

    J. L. Blanco and P. K. Rai, “nanoflann: a C++ header-only fork of FLANN, a library for nearest neighbor (NN) with kd-trees,” https:// github.com/jlblancoc/nanoflann, 2014. 12

    2014

  56. [2019]

    Available: https://icave2.cse.buffalo.edu/resources/ sensor-modeling/VLP32CManual.pdf

    [Online]. Available: https://icave2.cse.buffalo.edu/resources/ sensor-modeling/VLP32CManual.pdf

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.