Pith. sign in

REVIEW 4 major objections 5 minor 28 references

Free-Space Features: Global Localization in 2D Laser SLAM Using Distance Function Maps

T0 review · 4 major / 5 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 Free-space geometry lifts 2D lidar place recognition recall

desk verdict A genuinely new SDF-based free-space descriptor for 2D lidar place recognition that shows large recall gains, but the ablation supporting the free-space mechanism is confounded by feature count. read the letter →

arxiv 1908.01863 v1 pith:LE2KNHDT submitted 2019-08-05 cs.RO

Alexander Millane , Helen Oleynikova , Juan Nieto , Roland Siegwart , César Cadena This is my paper · ORCID
classification cs.RO
keywords globallocalization2DlidarSLAMplacerecognitionsigneddistancefunctionfree-spacefeatureskeypointdetectiondescriptormatchingloopclosure
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 argues that the shape of free space—the open navigable areas between obstacles—carries enough information to recognize places in 2D lidar maps. It represents each submap as a signed distance function, detects keypoints at high-curvature points of this function, and describes them with orientation histograms plus the local distance value. On six trajectories from two public datasets, matching with these free-space features achieves recall at precision 1.0 of 0.18 to 0.99, improving on a Shape Contexts surface-only baseline by 68–177%. The paper concludes that the gain comes from explicitly describing free-space geometry rather than from a stronger detector.

What carries the argument

The signed distance function (SDF) $f: \mathbb{R}^2 \to \mathbb{R}$ maps each point to its signed distance to the nearest surface, representing free and occupied space symmetrically. Keypoints are selected where the determinant of the Hessian of the Gaussian-smoothed SDF is locally maximal—points of high curvature in the distance field—and are classified as maxima, minima, or saddles by the Hessian eigenvalues. Each keypoint is described by a 17-bin histogram of gradient orientations computed in a circular window; the histogram is made rotation-invariant by referencing a dominant orientation, and is augmented by the window's average SDF value and the stationary-point class. Matching then proceeds by nearest-neighbour descriptor lookup with a ratio test and RANSAC over SE(2) transforms. The load-bearing idea is that the SDF makes free-space geometry equally available for description, so descriptors can carry information about open areas rather than only surface points.

What would settle it

Run the same precision-recall evaluation with the distance-threshold ablation while subsampling the kept features so every trial has the same number of descriptors; if recall stops improving with distance from surfaces, the paper's free-space explanation is falsified.

Watch

Extended reading notes

Core claim

The central claim is that place recognition in 2D lidar SLAM is improved by describing the geometry of free space, not just the surfaces of occupied space, and that a signed distance function is a natural representation for doing so. Extracting keypoints with a determinant-of-Hessian detector on the SDF and describing them with gradient-histogram descriptors augmented with the average distance and stationary-point class yields a feature that outperforms a curvature-cluster/Shape Contexts pipeline at the same RANSAC matching step. The reported recall at precision 1.0 rises from 0.27 to 0.45 on the EG trajectory and from 0.36 to 0.99 on PR3, with increases between 68% and 177% across all six trajectories. The authors attribute this performance gap to the inclusion of features in free-space, based on an ablation that removes features farther than a threshold distance from surfaces.

Load-bearing premise

The paper's claim that free-space geometry, rather than merely having more features, drives the improvement rests on the untested assumption that matching recall is not substantially affected by the number of descriptors available in each submap.

Editorial extensions

If this is right

  • On all six evaluated trajectories, the free-space feature beats the Shape Contexts baseline at precision 1.0, so a robot revisiting a mapped area is more likely to recognize it without false positives.
  • The ablation suggests that including free-space regions up to several meters from surfaces improves matching, so SLAM front-ends that already produce SDFs can feed place recognition directly without converting to point clouds.
  • The same descriptor pipeline extends naturally to loop-closure detection and map-based localization; the paper demonstrates 292 submap-submap matches with no false matches in one experiment.
  • Because the SDF is metric, descriptor distances correspond to physical distances, which may make ratio-test matching more reliable than in image-retrieval style pipelines.

Reading between the lines

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

  • A natural next test is to hold the number of descriptors constant while varying free-space inclusion; the paper's ablation varies the distance threshold, which changes feature count and spatial distribution along with free-space content, so the mechanism could be tested more cleanly.
  • The same signed-distance representation could be used for global localization in 3D, where SDFs are already common in dense reconstruction; this is the paper's stated future direction but is not evaluated here.
  • If free-space shape is indeed distinctive, then maps with highly structured open areas—such as warehouse aisles or corridors—may benefit most from this approach; this is an inference, not a result reported in the paper.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. This paper proposes a global localization method for 2D laser SLAM based on features extracted from signed distance function (SDF) submaps. The detector selects keypoints at high curvature of the SDF using the determinant of the Hessian, classifies them as maxima, minima, or saddles, and describes them with a gradient orientation histogram plus a weighted average SDF value. Place recognition is performed by nearest-neighbor descriptor matching with the ratio test, followed by RANSAC for geometric verification. The authors evaluate on six trajectories from the Deutsches Museum and PR2 Willow Garage datasets, reporting recall at precision 1.0 of 0.45, 0.18, 0.30, 0.85, 0.76, and 0.99, corresponding to stated improvements of roughly 68--177% over an in-house Shape Contexts baseline. A second experiment varies the maximum distance of keypoints from surfaces to argue that free-space geometry drives the improvement, and a final experiment demonstrates loop closure and localization with 292 matches and no false positives.

Significance. If the central attribution to free-space were established, this would be a useful contribution: it introduces a representation that exploits free-space structure in 2D maps, reports concrete gains over a previously competitive descriptor on public datasets, and provides a clearly described evaluation protocol with public data and a parameter table. The distance-function formulation is clean, and the paper makes explicit, falsifiable claims about the role of free-space. However, the mechanistic claim is not yet supported because the ablation changes feature count as well as feature content, and the comparative claim is weakened by the absence of variance estimates and by an asymmetric input representation for the baseline. These issues are fixable with additional controlled experiments, so the work has solid potential but needs revision.

major comments (4)
  1. [Sec. V-B, Fig. 7] The free-space ablation varies d_threshold in {2.0, 1.5, 1.0, 0.5} m, which simultaneously changes how far keypoints may lie from surfaces and how many keypoints are extracted in each submap. Because RANSAC matching success probability increases with the number of tentative correspondences, the monotonic improvement in Fig. 7 is consistent with a pure feature-count effect even if the added descriptors carry no extra geometric information. Therefore the statement that this evaluation 'demonstrates that the performance of our proposal is due to the use of free-space, and not an advantage in the descriptive power of the keypoint' is not supported. Please add a control that fixes the number of features per submap across d_threshold conditions (e.g., by capping or randomly subsampling descriptors to a common count) and re-analyze; without such a control, the central mechanistic claim of the paper is not established.
  2. [Sec. V-A, Table II and Fig. 5] The precision-recall curves are generated from a single random selection of 1000 submap pairs, and no repeated trials, confidence intervals, or statistical tests are reported. Since both the pair selection and RANSAC are stochastic, the reported recall differences (e.g., 0.99 vs. 0.36 on PR3) should be accompanied by variance estimates such as repeated subsampling or bootstrap confidence intervals. Without these, the magnitudes of the claimed improvements are not statistically grounded, and it is unclear whether the ranking of methods is stable across random subsamples.
  3. [Sec. V-A, baseline construction] The comparison with Shape Contexts is not fully apples-to-apples. The proposed method uses the full occupancy-grid submaps produced by Cartographer, while the Shape Contexts pointcloud is produced by aggregating temporally sub-sampled scans to approximately 30 scans per submap. This asymmetry can reduce the density of the baseline pointcloud and therefore the number and quality of baseline keypoints, potentially inflating the relative improvement. Please either provide the baseline with an equivalently complete pointcloud (or justify the subsampling), and report feature counts for both methods to demonstrate that the comparison is not driven by input data quantity.
  4. [Sec. V-A, parameter tuning] The grid search used to set parameters for both methods is described only as maximizing recall at precision 1.0 on 'a separate localization experiment,' without specifying which data, protocol, or split this refers to. In addition, Table I does not list all parameters of the Curvature Clusters detector (e.g., curvature thresholds and clustering parameters). Please provide the full tuning protocol and parameter values, and state clearly whether any of the six evaluation trajectories were used, directly or indirectly, during parameter selection.
minor comments (5)
  1. [Sec. IV-B] The statement that the SDF is 'smooth, in the sense that it is differentiable almost everywhere' conflates smoothness with almost-everywhere differentiability; the distance function has gradient discontinuities across the medial axis and at surface boundaries, and the Hessian used for detection is computed on a smoothed grid. Please rephrase to avoid this imprecision.
  2. [Sec. V-A] The percentage increases quoted in the text ('69%, 140%, 68%, 157%, 160%, 177%') do not exactly match the rounded values in Table II (for example, 0.18 vs. 0.07 is a ~157% increase, not 140%). If these percentages are computed from unrounded recall values, please state that in the text.
  3. [Sec. V-B] The phrase 'we perform several trails' should be 'trials'.
  4. [Sec. V-A, related work] The baseline [18] is from 2009; the characterization of Shape Contexts as 'state-of-the-art' should be justified with respect to more recent 2D place-recognition methods, or the claim should be softened to 'competitive in the evaluation of [18]'.
  5. [General] The paper does not report runtime, memory usage, or feature counts for the proposed method and baseline. Adding these would substantially help readers assess practical applicability in a SLAM context.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported gains are measured against an external baseline and no fitted quantity is renamed as a prediction.

full rationale

The paper's derivation chain is empirical rather than definitional. It converts occupancy grids into signed distance functions, detects keypoints with a Hessian-based detector, describes them with a gradient orientation histogram augmented by average SDF value and topology class, and matches submaps using nearest-neighbour lookup, ratio test, and RANSAC. Recall at precision 1.0 is then measured on public datasets (Deutsches Museum and Willow Garage PR2) against Shape Contexts, an external baseline recommended by Bosse and Zlot, with parameters selected on a separate localization experiment. None of these steps defines the reported score in terms of itself, and the comparison is against an independent method. The free-space ablation in Sec. V-B varies d_threshold, which changes both the spatial distribution and the number of keypoints, so the causal claim that free-space geometry drives the improvement is not fully controlled; however, that is an experimental confound and a potential correctness issue, not a circular derivation. Self-citations to the authors' prior work (C-blox, Voxblox, Cadena et al.) are contextual and do not carry the argument. No uniqueness theorem, ansatz citation, or renaming of a known result is used as the load-bearing justification. Therefore no circular step can be exhibited, and the circularity score is 0.

Assumptions & free parameters 7 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new physical entities or conserved quantities. The central claim rests on tuned hyperparameters (listed in Table I plus an unreported Gaussian sigma) and on domain assumptions about ground-truth label quality, the informativeness of free space, and the reliability of the matching pipeline. The main unstated dependency is that the ablation isolates free-space information rather than feature count.

free parameters (7)
  • descriptor radius = 0.8 m
    Feature support window size; grid-searched to maximize recall at precision 1.0 (Table I).
  • number of orientation bins = 17
    Descriptor histogram dimensionality; tuned via grid search (Table I).
  • distance weight = 0.002
    Weight for average SDF value in descriptor; tuned via grid search (Table I).
  • detection threshold = 0.0025
    DoH keypoint detection threshold; tuned via grid search (Table I).
  • matching max ratio = 0.75
    Ratio test threshold for descriptor matching; tuned via grid search (Table I).
  • Shape Context baseline parameters = radius 2.0 m, radial bins 3, angular bins 6, threshold 0.05
    Baseline implementation tuned by the same grid search; in-house baseline may not match the original method's settings.
  • Gaussian smoothing variance sigma^2 = not reported
    Tunable in Eq. (2), but not listed in Table I; affects keypoint detection and is unspecified.
assumptions (4)
  • domain assumption Submaps produced by Cartographer and their poses are reliable ground truth for overlap-based match labels (Sec. V-A).
    The evaluation derives ground-truth positive and negative labels from submap overlap computed from Cartographer poses; if these poses are inaccurate, labels are wrong.
  • domain assumption Features in free space are informative for place recognition (Sec. I).
    This is the motivating hypothesis; the paper tests it but does not derive it.
  • domain assumption SDF generated by thresholding an occupancy grid to binary and then applying a distance transform preserves the geometry relevant to place recognition (Sec. IV-A).
    No analysis of information loss from thresholding occupancy probabilities is given; this is an implementation choice.
  • domain assumption Keypoint classification by Hessian eigenvalues and descriptor matching via nearest-neighbour ratio test plus RANSAC inliers reliably indicate submap overlap (Sec. IV-B, IV-D).
    The pipeline components are adopted from image retrieval literature without formal guarantees in this setting.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Free-Space Features: Global Localization in 2D Laser SLAM Using Distance Function Maps." pith.science (2026). https://pith.science/paper/LE2KNHDT

@misc{pith2026190801863,
  author       = {Pith},
  title        = {Pith review of: Free-Space Features: Global Localization in 2D Laser SLAM Using Distance Function Maps},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LE2KNHDT}},
  note         = {Machine review of arXiv:1908.01863}
}
read the original abstract

In many applications, maintaining a consistent map of the environment is key to enabling robotic platforms to perform higher-level decision making. Detection of already visited locations is one of the primary ways in which map consistency is maintained, especially in situations where external positioning systems are unavailable or unreliable. Mapping in 2D is an important field in robotics, largely due to the fact that man-made environments such as warehouses and homes, where robots are expected to play an increasing role, can often be approximated as planar. Place recognition in this context remains challenging: 2D lidar scans contain scant information with which to characterize, and therefore recognize, a location. This paper introduces a novel approach aimed at addressing this problem. At its core, the system relies on the use of the distance function for representation of geometry. This representation allows extraction of features which describe the geometry of both surfaces and free-space in the environment. We propose a feature for this purpose. Through evaluations on public datasets, we demonstrate the utility of free-space in the description of places, and show an increase in localization performance over a state-of-the-art descriptor extracted from surface geometry.

Figures

Figures reproduced from arXiv: 1908.01863 by the authors.

Figure 1
Figure 1. An example of place-recognition using the proposed method. The query submap (red) from the end of a trajectory is matched against a submap (blue) from the start. Query and match submaps are displayed as distance functions, free-space feature matches are shown in black, and the path taken by the agent in blue. A close up of the matched submaps is shown in (b). be conducted in these areas. The hypothesis motivating th… view at source ↗
Figure 2
Figure 2. An example submap, taken from the Deutsches Museum datasets [6], represented as a occupancy probability grid (a), and an SDF (b). B. Keypoint Detection Keypoint detection aims to extract points which are salient, that is can be reliably re-extracted, and interesting enough to warrant description. In contrast to image data, an SDF is by definition smooth, in the sense that it is differentiable almost everywhere. The … view at source ↗
Figure 4
Figure 4. An example feature extracted from an SDF submap showing (a) the location of the detected feature, (b) weighted gradients in a circular window, and (c), the computed orientation histogram (blue) and the average SDF value of the extracted window (red). shown in Fig. 4a and Fig. 4b. To achieve rotational invariance we use a 36-bin gradient orientation histogram to determine the dominant orientation within this window a… view at source ↗
Figures from the paper (3 more)
Figure 5
Figure 5. Figure 5: Precision-Recall curves for the proposed method (blue) and a comparison method (red) (a combination of curvature-clusters and Shape Contexts [26]). Maps are created using 2D lidar data from the Deutsches Museum dataset [6] (a), (c), (e) and the Willow Garage PR2 datase…
Figure 8
Figure 8. Figure 8: Localization experiment using the proposed method. Submaps created during traversal of the red trajectory are matched reference submaps from the blue trajectory. The experiment results in 292 submap-submap matches of which one is highlighted. submaps and a reference ma…
Figure 7
Figure 7. Figure 7: Precision Recall curves for a range of distances from surface boundaries in which frees-space features were extracted (see Sec. V-B. The plot indicates better matching performance as more free-space is included in the submap description. For comparison, the matching pe…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 25 canonical work pages

  1. [1]

    Past, present, and future of simultaneous localization and mapping: Toward the robust-perception age,

    C. Cadena, L. Carlone, H. Carrillo, Y . Latif, D. Scaramuzza, J. Neira, I. Reid, and J. J. Leonard, “Past, present, and future of simultaneous localization and mapping: Toward the robust-perception age,” IEEE Transactions on Robotics , vol. 32, no. 6, pp. 1309–1332, 2016

    2016

  2. [2]

    Efficient grid-based spatial rep- resentations for robot navigation in dynamic environments,

    B. Lau, C. Sprunk, and W. Burgard, “Efficient grid-based spatial rep- resentations for robot navigation in dynamic environments,” Robotics and Autonomous Systems , vol. 61, no. 10, pp. 1116–1130, 2013

    work page 2013

  3. [3]

    Kinectfusion: real-time 3d reconstruction and interaction using a moving depth camera,

    S. Izadi, D. Kim, O. Hilliges, D. Molyneaux, R. Newcombe, P. Kohli, J. Shotton, S. Hodges, D. Freeman, A. Davison, et al., “Kinectfusion: real-time 3d reconstruction and interaction using a moving depth camera,” in Proceedings of the 24th annual ACM symposium on User interface software and technology , pp. 559–568, ACM, 2011

    work page 2011

  4. [4]

    Adaptively sampled distance fields: A general representation of shape for computer graphics,

    S. F. Frisken, R. N. Perry, A. P. Rockwood, and T. R. Jones, “Adaptively sampled distance fields: A general representation of shape for computer graphics,” in Proceedings of the 27th annual conference on Computer graphics and interactive techniques , pp. 249–254, ACM Press/Addison-Wesley Publishing Co., 2000

    work page 2000

  5. [5]

    Improved techniques for grid mapping with rao-blackwellized particle filters,

    G. Grisetti, C. Stachniss, W. Burgard, et al., “Improved techniques for grid mapping with rao-blackwellized particle filters,” IEEE Transac- tions on Robotics , 2007

    work page 2007

  6. [6]

    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 Robotics and Automation (ICRA), 2016 IEEE International Conference on , pp. 1271–1278, IEEE, 2016

    work page 2016

  7. [7]

    M3rsm: Many-to-many multi-resolution scan matching,

    E. Olson, “M3rsm: Many-to-many multi-resolution scan matching,” in 2015 IEEE International Conference on Robotics and Automation (ICRA), pp. 5815–5821, IEEE, 2015

    work page 2015

  8. [8]

    Incremental mapping of large cyclic environments,

    J.-S. Gutmann and K. Konolige, “Incremental mapping of large cyclic environments,” in Computational Intelligence in Robotics and Automation, 1999. CIRA’99. Proceedings. 1999 IEEE International Symposium on , pp. 318–325, IEEE, 1999

    work page 1999

Show all 28 references
  1. [9]

    Real-time correlative scan matching,

    E. B. Olson, “Real-time correlative scan matching,” in 2009 IEEE International Conference on Robotics and Automation , pp. 4387–4393, IEEE, 2009

    2009

  2. [10]

    Video google: A text retrieval approach to object matching in videos,

    J. Sivic and A. Zisserman, “Video google: A text retrieval approach to object matching in videos,” in null, p. 1470, IEEE, 2003

    2003

  3. [11]

    A new approach to global self-localization with laser range scans in unstructured environments,

    A. Walthelm, “A new approach to global self-localization with laser range scans in unstructured environments,” in Intelligent V ehicle Symposium, 2002. IEEE , vol. 1, pp. 202–208, IEEE, 2002

    2002

  4. [12]

    Flirt-interest regions for 2d range data,

    G. D. Tipaldi and K. O. Arras, “Flirt-interest regions for 2d range data,” in Robotics and Automation (ICRA), 2010 IEEE International Conference on, pp. 3616–3622, IEEE, 2010

    2010

  5. [13]

    Geometrical flirt phrases for large scale place recognition in 2d range data,

    G. D. Tipaldi, L. Spinello, and W. Burgard, “Geometrical flirt phrases for large scale place recognition in 2d range data,” in 2013 IEEE International Conference on Robotics and Automation , pp. 2693–2698, IEEE, 2013

    2013

  6. [14]

    Learning to close loops from range data,

    K. Granstr ¨om, T. B. Sch ¨on, J. I. Nieto, and F. T. Ramos, “Learning to close loops from range data,” The international journal of robotics research, vol. 30, no. 14, pp. 1728–1754, 2011

    2011

  7. [15]

    Deep learning for 2d scan matching and loop closure,

    J. Li, H. Zhan, B. M. Chen, I. Reid, and G. H. Lee, “Deep learning for 2d scan matching and loop closure,” in Intelligent Robots and Systems (IROS), 2017 IEEE/RSJ International Conference on , pp. 763–768, IEEE, 2017

    2017

  8. [16]

    Simultaneous localization and map building in large-scale cyclic environments using the atlas framework,

    M. Bosse, P. Newman, J. Leonard, and S. Teller, “Simultaneous localization and map building in large-scale cyclic environments using the atlas framework,” The International Journal of Robotics Research , vol. 23, no. 12, pp. 1113–1139, 2004

    2004

  9. [17]

    C-blox: A scalable and consistent tsdf-based dense mapping approach,

    A. Millane, Z. Taylor, H. Oleynikova, J. Nieto, R. Siegwart, and C. Cadena, “C-blox: A scalable and consistent tsdf-based dense mapping approach,” in 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pp. 995–1002, IEEE, 2018

    2018

  10. [18]

    Keypoint design and evaluation for place recognition in 2d lidar maps,

    M. Bosse and R. Zlot, “Keypoint design and evaluation for place recognition in 2d lidar maps,” Robotics and Autonomous Systems , vol. 57, no. 12, pp. 1211–1224, 2009

    2009

  11. [19]

    V oxblox: Incremental 3d euclidean signed distance fields for on- board mav planning,

    H. Oleynikova, Z. Taylor, M. Fehr, R. Siegwart, and J. Nieto, “V oxblox: Incremental 3d euclidean signed distance fields for on- board mav planning,” in Intelligent Robots and Systems (IROS), 2017 IEEE/RSJ International Conference on , pp. 1366–1373, IEEE, 2017

    2017

  12. [20]

    2d-sdf-slam: A signed distance function based slam frontend for laser scanners,

    J.-D. Fossel, K. Tuyls, and J. Sturm, “2d-sdf-slam: A signed distance function based slam frontend for laser scanners,” in Intelligent Robots and Systems (IROS), 2015 IEEE/RSJ International Conference on , pp. 1949–1955, IEEE, 2015

    2015

  13. [21]

    Multi-robot local- ization and mapping based on signed distance functions,

    P. Koch, S. May, M. Schmidpeter, M. K ¨uhn, C. Pfitzner, C. Merkl, R. Koch, M. Fees, J. Martin, D. Ammon, et al. , “Multi-robot local- ization and mapping based on signed distance functions,” Journal of Intelligent & Robotic Systems , vol. 83, no. 3-4, pp. 409–428, 2016

    2016

  14. [22]

    A linear time algorithm for computing exact euclidean distance transforms of binary images in arbitrary dimensions,

    C. R. Maurer, R. Qi, and V . Raghavan, “A linear time algorithm for computing exact euclidean distance transforms of binary images in arbitrary dimensions,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 25, no. 2, pp. 265–270, 2003

    2003

  15. [23]

    Surf: Speeded up robust features,

    H. Bay, T. Tuytelaars, and L. Van Gool, “Surf: Speeded up robust features,” in European conference on computer vision , pp. 404–417, Springer, 2006

    2006

  16. [24]

    Histograms of oriented gradients for human detection,

    N. Dalal and B. Triggs, “Histograms of oriented gradients for human detection,” in Computer Vision and Pattern Recognition, 2005. CVPR

    2005

  17. [25]

    Object recognition from local scale-invariant features,

    D. G. Lowe, “Object recognition from local scale-invariant features,” in Computer vision, 1999. The proceedings of the seventh IEEE international conference on , vol. 2, pp. 1150–1157, Ieee, 1999

    1999

  18. [26]

    Shape matching and object recognition using shape contexts,

    S. Belongie, J. Malik, and J. Puzicha, “Shape matching and object recognition using shape contexts,” tech. rep., California University, San Diego, La Jolla, Dept. of Computer Science and Engineering, 2002

    2002

  19. [27]

    An object-based semantic world model for long-term change detection and semantic querying,

    J. Mason and B. Marthi, “An object-based semantic world model for long-term change detection and semantic querying,” in 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems , pp. 3851– 3858, IEEE, 2012

    2012

  20. [2005]

    IEEE Computer Society Conference on , vol. 1, pp. 886–893, IEEE, 2005

    2005

Pith tools

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