Pith. sign in

REVIEW 4 major objections 5 minor 2 cited by

A sparse, landmark-free map of stored images and graph edges can match explicit 3D models for multi-session visual navigation, at far lower storage and construction cost.

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

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

OpenNavMap shows that an image graph plus on-demand 3D reconstruction can match structure-based maps for visual localization and navigation.

T0 review reviewed 2026-08-03 challenge →

load-bearing objection A competent, honest systems paper whose Map-Free and 360Loc results give the core idea real grounding; the multi-session ATE claims rest on ground truth too close to the thing being measured. the 4 major comments →

arxiv 2601.12291 v2 pith:TUDCHNZ7 submitted 2026-01-18 cs.RO cs.CV

OpenNavMap: Multi-Session Appearance-Based Topometric Mapping for Scalable Visual Navigation

Jianhao Jiao , Changkun Liu , Jingwen Yu , Boyi Liu , Qianyi Zhang , Yue Wang , Dimitrios Kanoulas This is my paper
classification cs.RO cs.CV
keywords topometric mappingstructure-free SLAMmulti-session mappingcollaborative localizationvisual localizationvisual place recognition3D geometric foundation modelsimage-goal navigation
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The reading

OpenNavMap tries to establish that a sparse, landmark-free topometric map — just images, poses, and three kinds of connectivity edges — can do what explicit 3D models do for multi-session visual navigation, at a fraction of the storage and construction cost. Instead of storing geometry, the system calls a pretrained 3D geometric foundation model on demand to generate dense pointmaps for a query and a few overlapping reference images, then refines the pose with a confidence-calibrated optimization. The paper reports an average translation error of 0.62 m on a public map-free relocalization benchmark, over 80% of queries localized within 1 meter and 10 degrees using only two reference images, and an absolute trajectory error below 3 m over a 15.7 km multi-session map built without depth sensors. A sympathetic reader would care because, if true, crowd-sourced phone and street-view imagery could maintain robot maps for years without heavy reconstruction pipelines.

Core claim

The central discovery is that a sparse observational map can match or exceed a globally consistent 3D map for localization and navigation, provided geometry is generated on demand and filtered by calibrated confidence. The paper shows that pairwise image geometry from a pretrained geometric foundation model, combined with dynamic-programming sequence matching and geometric verification, yields precise relative poses — 0.62 m average translation on the map-free benchmark, with over 80% of queries within 1 m and 10° using two references. It further shows the same machinery can merge 15.7 km of multi-session, multi-device data into one graph with absolute trajectory error under 3 m, and that th

What carries the argument

The carrying object is a three-layer topometric graph: covisibility edges connect overlapping images, odometry edges carry relative poses and covariances, and traversability edges encode feasible motion. Geometry is deliberately absent; when a query arrives, a 3D geometric foundation model predicts dense pointmaps for the query and a handful of co-visible references, and a global optimization with residual-calibrated confidence weighting recovers the query pose and a per-pose covariance. Four mechanisms carry the argument: DP-based sequence matching over a difference matrix that allows jumps between trajectory segments; RANSAC-based geometric verification; the confidence-calibrated optimizat

Load-bearing premise

The headline accuracy numbers rest on ground-truth poses produced by a cloud-based SLAM service that also supplies the local VIO priors used to build the submaps, so if that ground truth carries the same drift the system is meant to correct, the 3 m and 0.62 m figures are not independent measurements of world accuracy.

What would settle it

Recompute the 15.7 km map-merging and metric-localization experiments using independent ground truth (surveyed markers, differential GPS, or a laser-scanned map) instead of the cloud SLAM poses; if the absolute trajectory error exceeds 3 m or average translation error rises above the reported 0.62 m under those independent references, the central claim that sparse maps match 3D models is weakened.

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

If this is right

  • If the claim holds, building a navigable map no longer requires structure-from-motion or LiDAR; a set of overlapping images with phone-level odometry is enough, because geometry is synthesized only when a localization query needs it.
  • Localization accuracy becomes mostly a property of the geometric foundation model and the calibration filter, not the density of the map: two reference images already yield sub-meter precision in the tested scenes, so maps can be grown cheaply.
  • Multi-session data from phones, vehicle cameras, and street-view panoramas can be merged into one globally consistent graph, so coverage can extend over time without rebuilding a 3D model — the paper demonstrates this on a 15.7 km map with ATE below 3 m.
  • The fused topometric graph supports image-goal navigation directly: 12 simulated and physical robot runs reached image-specified goals, suggesting the representation is enough for closed-loop visual navigation in GNSS-denied settings.
  • Node culling based on image quality, temporal difference, and information gain cuts map size by roughly 20% with minimal accuracy loss, which matters for long-term deployment.

Where Pith is reading between the lines

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

  • If the system's logic is extended, the real long-term bottleneck is the on-demand geometry model: as that model becomes faster and more robust to appearance change, the whole system's accuracy and cost improve without changing the map representation.
  • Because the paper itself notes that DP matching indexes traversal sequences rather than graph topology, one natural next step is to replace sequence indexing with graph- or location-indexed matching, which should improve recall for short, disconnected crowd-sourced clips.
  • The calibrated confidence score attached to every loop-closure pose could be exposed to the planner as an uncertainty estimate, enabling risk-aware navigation rather than serving only as a filter before pose graph optimization.
  • Given that the reported ATE relies on cloud-SLAM ground truth, an independent survey-based evaluation would be the prudent check before treating the sub-3 m multi-session claim as settled.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes OpenNavMap, a structure-free topometric mapping and multi-session localization system. A three-layer graph (covisibility, odometry, traversability) stores images with poses, while geometry is recovered on demand by the MASt3R foundation model. Submaps are aligned through a coarse-to-fine pipeline: DP-based sequence matching, geometric verification, global pose-only optimization with confidence-map calibration, and pose-graph optimization. Cross-device input and probabilistic node culling are added for lifelong operation, and the map is used for image-goal navigation. Experiments are reported on Map-Free, GZ-Campus, 360Loc, and self-collected Aria-glasses datasets, with claims of state-of-the-art metric-localization accuracy, sub-3 m ATE over 15.7 km, and 12 successful navigation trials.

Significance. If the claims hold, the paper is a significant step toward replacing dense 3D maps with sparse, image-only topometric representations for long-term visual navigation. The architecture is clearly specified and modular, the evaluation includes external benchmarks (Map-Free, GZ-Campus, 360Loc) with pretrained models used without fine-tuning, and the map-size analysis is concrete and useful. The real-robot navigation demonstrations and the code/data release commitment are also strengths. However, the load-bearing multi-session consistency claim currently rests on ground truth that is not independent of the system being evaluated, and the headline 0.62 m number is not actually reported in the experimental section. These issues do not invalidate the architecture but they do limit the current evidence for the central claims.

major comments (4)
  1. [Abstract / §I.C / §VII.B] The abstract and introduction claim a 0.62 m average translation error on Map-Free, but §VII.B reports only Precision@[1m,10°] and AUC curves (Fig. 10). No table or derivation gives the 0.62 m value, nor states the number of reference views, the averaging protocol, or the standard deviation. This number is part of the paper's headline contribution and must be backed by a concrete result in the main text.
  2. [§VI-B1 / Table IV] The ATE ground truth for the 15.7 km self-collected dataset is generated by Meta's cloud-based SLAM service, while the submaps themselves are built from VIO poses in the same Aria ecosystem. The two may therefore share estimation lineage and drift modes, making the reported <3 m ATE a measure of agreement with a related estimator rather than an independent global-accuracy evaluation. The paper states that the reference has only been 'qualitatively validated' at city scale. Please provide an independent reference for at least a subset (RTK/PPK GNSS, surveyed markers, LiDAR SLAM) or quantify the uncertainty/correlation of the cloud-SLAM reference with the local VIO poses. This is load-bearing for the multi-session scalability claim.
  3. [Algorithm 1 / Eq. (2)] There is an inconsistency between the DP recurrence in Eq. (2) and the implementation in Algorithm 1. Eq. (2) uses a single jump offset Δ from the previous row, while Algorithm 1 (lines 15–20) allows a jump to any k with |k−i| ≥ Δ, with the same constant penalty λ. The latter makes arbitrarily long jumps as cheap as Δ-length jumps, which can materially affect the matched path and the resulting loop-closure candidates. Please clarify which formulation is actually used, correct the pseudocode or the equation, and report sensitivity to Δ and λ.
  4. [§IV-B2 / §IV-B3 / Table IV] The PGO in Eq. (7) is weighted by the CCM-derived covariance Σ = diag[(W_i,i · W_i,j)^-2] introduced in §IV-B2. The paper asserts that the mean CCM reflects pose accuracy and uses it to filter loop closures in Fig. 14, but no calibration analysis connects the CCM to true pose error, and no ablation shows that the form of Σ improves the final ATE. Without this, the role of confidence calibration in the <3 m result is not established. Please add a quantitative evaluation of CCM calibration and the effect of the covariance model on PGO output.
minor comments (5)
  1. [Abstract / §VI-B1] The abstract and §I.C say '19 km dataset', while §VI-B1 says the Aria data total 'over 18.7 km'. Please harmonize these numbers.
  2. [§IV-A1 / §IV-D2] The covisibility strength v_Cij is defined but never used quantitatively; edge augmentation in §IV-D2 uses keypoint counts. Please clarify whether v_Cij is consumed anywhere or is only a conceptual annotation.
  3. [§IV-B2 / Eq. (6)] The notation W is used for the per-point robust weight in Eq. (6), but then Σ is defined via 'Wi,i · Wi,j' without a clear definition of these indices. This should be rewritten for readability.
  4. [§VII-E3 / Fig. 18] The real-world navigation experiments report a single 160 m Lab run with 312 s and multiple repetitions, but there is no quantitative summary table of success rates, path lengths, or completion times across all 12 trials. A compact table would strengthen the navigation claim.
  5. [§VI-C1] The metric 'Average@1' is defined but never reported or referenced in the results. Either use it in the tables or remove the definition.

Circularity Check

0 steps flagged

No significant circularity: central results rest on external benchmarks and pre-trained models; self-citations are contextual and the cloud-SLAM ground-truth issue is an evaluation caveat, not a circular derivation.

full rationale

The paper's central quantitative claims are not forced by its inputs. Metric localization is evaluated on public Map-Free (65 scenes), GZ-Campus, and 360Loc datasets with pre-trained MASt3R and no fine-tuning ('all learning-based models were utilized without additional fine-tuning', Sec. VI-A), so Eq. (5)'s pose optimization is not fit to the target errors. The >80% Precision@[1m,10°] with two references and the 0.62m average translation error are external-benchmark results. The multi-session ATE claim (Table IV) is the only headline tied to self-collected data; there, GT comes from Meta's cloud SLAM service (Sec. VI-B1) and the submaps use Aria VIO priors (Sec. IV-A2). This is a real independence/validity concern, but it is not a circularity: the PGO output is not computed from the GT, and the GT's accuracy is supported by an external city-scale study [69]. Self-citations (LiteVLoc [14], GV-Bench [54], MS-Mapping [60]) are related-work/component citations rather than load-bearing premises; no uniqueness theorem or ansatz is imported from the authors' own prior work. No equation-level reduction, no fitted parameter renamed as a prediction, and no self-defined target was found.

Axiom & Free-Parameter Ledger

7 free parameters · 6 axioms · 2 invented entities

The central results depend on many unreported hyperparameters, on strong assumptions about VIO submaps and MASt3R pointmaps, and on trusting a commercial cloud-SLAM service as ground truth. The two invented constructs (three-layer graph, CCM) are tested on external data, which limits but does not remove the burden.

free parameters (7)
  • Keyframe selection thresholds = 3.9 m / 60 deg
    Sec. VI-A1: fixed thresholds determine which frames become graph nodes; no ablation or justification for exact values.
  • DP matching parameters (velocity set V, jump threshold Delta, jump penalty lambda) = not specified
    Sec. IV-B1 / Algorithm 1: these control sequence matching and are never reported.
  • RANSAC inlier threshold for geometric verification = not specified
    Sec. IV-B1: a pair is accepted only if inliers exceed a threshold, but the threshold is not given.
  • Geman-McClure outlier scale mu = not specified
    Sec. IV-B2, Eq. (6): controls confidence-weight calibration and is not reported.
  • IQA minimum threshold and culling probability threshold = not specified
    Sec. IV-C / IV-D: images below an IQA threshold are filtered and nodes below a probability threshold are culled; neither threshold is given.
  • Pose covariance heuristic = diag[(W_i*W_j)^-2]
    Sec. IV-B2: ad hoc covariance for PGO factors, neither learned nor externally validated.
  • Topological correctness tolerance = [7.5 m, 75 deg]
    Sec. VI-C1: chosen 'to reflect practical navigation requirements'; directly defines all topological recall/precision numbers.
axioms (6)
  • domain assumption Each device's VIO provides scale-aware, locally consistent poses for submap construction.
    Sec. IV-A2 assumes VIO/SLAM outputs are locally consistent enough for graph construction and PGO.
  • domain assumption Consecutive keyframes are traversable.
    Sec. VI-A1 explicitly states that local connectivity between sequential keyframes is assumed traversable when building the TrG.
  • domain assumption Meta Aria cloud-based SLAM poses are accurate enough to serve as ground truth.
    Sec. VI-B1: all self-collected GT poses come from this service, with only qualitative prior validation cited.
  • domain assumption MASt3R predicts reliable pointmaps, intrinsics, and confidence maps for uncalibrated cross-device image pairs.
    Sec. IV-A3 and IV-B2: the whole metric localization stage rests on the pretrained GFM's generalization.
  • domain assumption Geographically close images are visually similar enough for descriptor-based retrieval.
    Sec. IV-B1: cosine similarity of VPR descriptors is used as the basis for topological matching.
  • standard math Nonlinear least-squares and PGO converge to a good local minimum.
    Sec. IV-B2 / IV-B3 rely on gradient descent and GTSAM's Levenberg-Marquardt solver without convergence guarantees.
invented entities (2)
  • Three-layer topometric map (covisibility, odometry, traversability graphs) independent evidence
    purpose: Unified sparse representation supporting localization, pose-graph optimization, and path planning.
    Validated on external benchmarks (Map-Free, GZ-Campus) and in navigation trials; the packaging is new even if the layers are known.
  • Calibrated Confidence Map (CCM) independent evidence
    purpose: Reweight per-point confidence by residuals during optimization and estimate pose covariance for loop-closure filtering.
    AUC curves on Map-Free and GZ-Campus show CCM separates correct from incorrect pose estimates better than raw confidence.

reviewed 2026-08-03 · how reviews work

0 comments
Cite this review

Pith. "Pith review of OpenNavMap: Multi-Session Appearance-Based Topometric Mapping for Scalable Visual Navigation." pith.science (2026). https://pith.science/paper/TUDCHNZ7

@misc{pith2026260112291,
  author       = {Pith},
  title        = {Pith review of: OpenNavMap: Multi-Session Appearance-Based Topometric Mapping for Scalable Visual Navigation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TUDCHNZ7}},
  note         = {Machine review of arXiv:2601.12291}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Scalable and maintainable maps are fundamental to large-scale navigation and the long-term deployment of robots in real-world environments. However, conventional maps that explicitly maintain dense geometry or 3D landmarks incur high storage and maintenance costs, while the core challenge of scaling to multi-session mapping is visual localization under sparse viewpoint overlap, temporal appearance shifts, and cross-device variance. To address this, we propose OpenNavMap, a lightweight, landmark-free topometric mapping system that organizes image nodes into covisibility, odometry, and traversability graphs and delegates local geometry recovery to 3D geometric foundation models (GFMs) on demand. For localization, dynamic-programming-based sequence matching narrows candidate correspondences for a GFM, reducing global estimation to a lightweight, pose-only optimization; for mapping, a lifelong pipeline fuses multi-session, multi-device data via cross-device merging and node culling. On a 19km dataset across four real-world environments, \methodname attains a state-of-the-art $0.62$m translation error on the Map-Free benchmark, bounds the absolute trajectory error below $3$m across 15.7km without depth sensors, and completes $12$ autonomous image-goal visual navigation tasks on both simulated and physical robots. Code and datasets will be made publicly available at https://rpl-cs-ucl.github.io/OpenNavMap_page.

Figures

Figures reproduced from arXiv: 2601.12291 by Boyi Liu, Changkun Liu, Dimitrios Kanoulas, Jianhao Jiao, Jingwen Yu, Qianyi Zhang, Yue Wang.

Figure 1
Figure 1. Figure 1: Conceptual illustration of the structure-free topometric map generated [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Block diagram illustrating the pipeline of the proposed O [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Three types of trajectories occurring in multi-session mapping, with [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: The raw and calibrated confidence maps of the query image are shown [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Illustration of strategies for cross-device localization. (a) shows how a [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Illustration of the node culling strategy for lifelong map maintenance. [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 8
Figure 8. Figure 8: The VNav system is deployed on a real-world ANYmal-D robot, [PITH_FULL_IMAGE:figures/full_fig_p009_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Overview of our self-collected dataset and trajectories captured [PITH_FULL_IMAGE:figures/full_fig_p010_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Evaluation of metric localization performance across three datasets. [PITH_FULL_IMAGE:figures/full_fig_p012_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Sample results of metric localization on three datasets. The accompa [PITH_FULL_IMAGE:figures/full_fig_p013_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Incremental map merging with submaps added in random order. Disconnected submaps, as in (a), do not affect pose graph optimization. Example [PITH_FULL_IMAGE:figures/full_fig_p014_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Estimated trajectory of R2-4 from our multi-sesion mapping system and GT trajectory from Aria glass-provided SLAM service. TABLE IV ABSOLUTE TRAJECTORY ERROR (ATE) ACROSS DIFFERENT SCENARIOS IN OUR SELF-COLLECTED DATASET. Data Distance Time Spans Shuffle Translational ATE [m] Rotational ATE [deg] R0-InOrder 0.6km 6mins N 0.47 0.70 R0-0 Y 0.65 0.86 R0-1 Y 0.27 0.50 R1-InOrder 2.5km 18hours N 1.18 0.44 R1-0… view at source ↗
Figure 15
Figure 15. Figure 15: Experiments on multi-session mapping with heterogeneous devices [PITH_FULL_IMAGE:figures/full_fig_p015_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Global path planning on the topometric map at different mapping stages (a)–(e). The caption for each subfigure indicates the time span of the [PITH_FULL_IMAGE:figures/full_fig_p016_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: The VNav system is evaluated in a simulated home environment [PITH_FULL_IMAGE:figures/full_fig_p016_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: Real-world experiment (Lab) with a quadruped robot. The topometric map structure comprises nodes (blue arrows) and traversable edges (blue [PITH_FULL_IMAGE:figures/full_fig_p017_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: Real-world experiments (Bridge and Building) demonstrate the [PITH_FULL_IMAGE:figures/full_fig_p017_19.png] view at source ↗
Figure 20
Figure 20. Figure 20: Lifelong operation for the incremental map merging in Lab, where [PITH_FULL_IMAGE:figures/full_fig_p018_20.png] view at source ↗

discussion (0)

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

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. RA-LWLM: Retrieval-Augmented In-Context Localization with Wireless Foundation Models

    eess.SP 2026-06 unverdicted novelty 7.0

    RA-LWLM uses retrieval from per-scene databases and in-context learning with a frozen foundation model to achieve cross-scene wireless localization without retraining.

  2. G2G: Exploiting Intra-Group Geometry for Inter-Group Pose Estimation

    cs.CV 2026-06 unverdicted novelty 6.0

    G2G attaches three small trainable modules to frozen backbones and reports state-of-the-art inter-group pose accuracy on four datasets spanning simulation, real cross-season, and sim-to-real transfer using only relati...

Reference graph

Works this paper leans on

82 extracted references · 6 linked inside Pith · cited by 2 Pith papers

  1. [1]

    Robust collaborative visual-inertial slam for mobile augmented reality,

    X. Pan, G. Huang, Z. Zhang, J. Li, H. Bao, and G. Zhang, “Robust collaborative visual-inertial slam for mobile augmented reality,”IEEE Trans. Vis. Comput. Graph., 2024

    2024

  2. [2]

    Map-free visual relocalization: Metric pose relative to a single image,

    E. Arnold, J. Wynn, S. Vicente, G. Garcia-Hernando, A. Monszpart, V . Prisacariu, D. Turmukhambetov, and E. Brachmann, “Map-free visual relocalization: Metric pose relative to a single image,” inProc. Eur . Conf. Comput. Vis., 2022, pp. 690–708

    2022

  3. [3]

    Grounding image matching in 3D with MASt3R,

    V . Leroy, Y . Cabon, and J. Revaud, “Grounding image matching in 3D with MASt3R,” inProc. Eur . Conf. Comput. Vis., 2024, pp. 71–91

    2024

  4. [4]

    DUSt3R: Geometric 3D vision made easy,

    S. Wang, V . Leroy, Y . Cabon, B. Chidlovskii, and J. Revaud, “DUSt3R: Geometric 3D vision made easy,” inProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2024, pp. 20 697–20 709

    2024

  5. [5]

    VGGT: Visual geometry grounded transformer,

    J. Wang, M. Chen, N. Karaev, A. Vedaldi, C. Rupprecht, and D. Novotny, “VGGT: Visual geometry grounded transformer,” inProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2025, pp. 5294–5306

    2025

  6. [6]

    Learning robust autonomous navigation and locomotion for wheeled- legged robots,

    J. Lee, M. Bjelonic, A. Reske, L. Wellhausen, T. Miki, and M. Hutter, “Learning robust autonomous navigation and locomotion for wheeled- legged robots,”Sci. Robot., vol. 9, no. 89, p. eadi9641, 2024

    2024

  7. [7]

    1 year, 1000 km: The oxford robotcar dataset,

    W. Maddern, G. Pascoe, C. Linegar, and P. Newman, “1 year, 1000 km: The oxford robotcar dataset,”Int. J. Robot. Res., vol. 36, no. 1, pp. 3–15, 2017

    2017

  8. [8]

    Structure-from-motion revisited,

    J. L. Sch ¨onberger and J.-M. Frahm, “Structure-from-motion revisited,” inProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2016

    2016

  9. [9]

    Crowd- sourced NeRF: Collecting data from production vehicles for 3D street view reconstruction,

    T. Qin, C. Li, H. Ye, S. Wan, M. Li, H. Liu, and M. Yang, “Crowd- sourced NeRF: Collecting data from production vehicles for 3D street view reconstruction,”IEEE Trans. Intell. Transp. Syst., 2024

    2024

  10. [10]

    Planet dump retrieved from https://planet.osm.org ,

    OpenStreetMap contributors, “Planet dump retrieved from https://planet.osm.org ,” https://www.openstreetmap.org, 2017

    2017

  11. [11]

    Google street view: Capturing the world at street level,

    D. Anguelov, C. Dulong, D. Filip, C. Frueh, S. Lafon, R. Lyon, A. Ogale, L. Vincent, and J. Weaver, “Google street view: Capturing the world at street level,”IEEE Computer, vol. 43, no. 6, pp. 32–38, 2010

    2010

  12. [12]

    OrienterNet: Visual localization in 2D public maps with neural match- ing,

    P.-E. Sarlin, D. DeTone, T.-Y . Yang, A. Avetisyan, J. Straub, T. Mal- isiewicz, S. R. Bulo, R. Newcombe, P. Kontschieder, and V . Balntas, “OrienterNet: Visual localization in 2D public maps with neural match- ing,” inProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2023, pp. 21 632–21 642

    2023

  13. [13]

    General place recognition survey: Towards real-world autonomy,

    P. Yin, J. Jiao, S. Zhao, L. Xu, G. Huang, H. Choset, S. Scherer, and J. Han, “General place recognition survey: Towards real-world autonomy,”IEEE Trans. Robot., 2025

    2025

  14. [14]

    LiteVLoc: Map-lite visual localization for image goal navigation,

    J. Jiao, J. He, C. Liu, S. Aegidius, X. Hu, T. Braud, and D. Kanoulas, “LiteVLoc: Map-lite visual localization for image goal navigation,” in Proc. IEEE Int. Conf. Robot. Autom., 2025

    2025

  15. [15]

    Hydra: A real-time spatial perception system for 3D scene graph construction and optimization,

    N. Hughes, Y . Chang, and L. Carlone, “Hydra: A real-time spatial perception system for 3D scene graph construction and optimization,” arXiv preprint arXiv:2201.13360, 2022

    Pith/arXiv arXiv 2022

  16. [16]

    Visual teach and repeat for long-range rover autonomy,

    P. Furgale and T. D. Barfoot, “Visual teach and repeat for long-range rover autonomy,”J. Field Robot., vol. 27, no. 5, pp. 534–560, 2010

    2010

  17. [17]

    Project Aria: A new tool for egocentric multi-modal ai research,

    J. Engelet al., “Project Aria: A new tool for egocentric multi-modal ai research,”arXiv preprint arXiv:2308.13561, 2023

    Pith/arXiv arXiv 2023

  18. [18]

    Carlone, A

    L. Carlone, A. Kim, F. Dellaert, T. Barfoot, and D. Cremers,SLAM Handbook. From Localization and Mapping to Spatial Intelligence, L. Carlone, A. Kim, F. Dellaert, T. Barfoot, and D. Cremers, Eds. Cambridge University Press, 2024

    2024

  19. [19]

    ORB-SLAM: A versatile and accurate monocular SLAM system,

    R. Mur-Artal, J. M. M. Montiel, and J. D. Tardos, “ORB-SLAM: A versatile and accurate monocular SLAM system,”IEEE Trans. Robot., vol. 31, no. 5, pp. 1147–1163, 2015

    2015

  20. [20]

    Fast-LIVO2: Fast, direct lidar-inertial-visual odometry,

    C. Zheng, W. Xu, Z. Zou, T. Hua, C. Yuan, D. He, B. Zhou, Z. Liu, J. Lin, F. Zhuet al., “Fast-LIVO2: Fast, direct lidar-inertial-visual odometry,”IEEE Trans. Robot., 2024

    2024

  21. [21]

    3D gaussian splatting for real-time radiance field rendering,

    B. Kerbl, G. Kopanas, T. Leimk ¨uhler, and G. Drettakis, “3D gaussian splatting for real-time radiance field rendering,”ACM Trans. Graph., vol. 42, no. 4, pp. 1–14, 2023

    2023

  22. [22]

    Graph-based subterranean exploration path planning using aerial and legged robots,

    T. Dang, M. Tranzatto, S. Khattak, F. Mascarich, K. Alexis, and M. Hutter, “Graph-based subterranean exploration path planning using aerial and legged robots,”J. Field Robot., vol. 37, no. 8, pp. 1363–1388, 2020

    2020

  23. [23]

    Rethinking visual geo- localization for large-scale applications,

    G. Berton, C. Masone, and B. Caputo, “Rethinking visual geo- localization for large-scale applications,” inProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2022, pp. 4878–4888

    2022

  24. [24]

    GNM: A general navigation model to drive any robot,

    D. Shah, A. Sridhar, A. Bhorkar, N. Hirose, and S. Levine, “GNM: A general navigation model to drive any robot,” inProc. IEEE Int. Conf. Robot. Autom., 2023, pp. 7226–7233

    2023

  25. [25]

    NeRF in the wild: Neural radiance fields for unconstrained photo collections,

    R. Martin-Brualla, N. Radwan, M. S. Sajjadi, J. T. Barron, A. Doso- vitskiy, and D. Duckworth, “NeRF in the wild: Neural radiance fields for unconstrained photo collections,” inProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2021, pp. 7210–7219

    2021

  26. [26]

    Accelerated coordi- nate encoding: Learning to relocalize in minutes using RGB and poses,

    E. Brachmann, T. Cavallari, and V . A. Prisacariu, “Accelerated coordi- nate encoding: Learning to relocalize in minutes using RGB and poses,” inProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2023, pp. 5044–5053. 20

    2023

  27. [27]

    Vision-only robot navigation in a neural radiance world,

    M. Adamkiewicz, T. Chen, A. Caccavale, R. Gardner, P. Culbertson, J. Bohg, and M. Schwager, “Vision-only robot navigation in a neural radiance world,”IEEE Robot. Autom. Lett., vol. 7, no. 2, pp. 4606–4613, 2022

    2022

  28. [28]

    Kimera: From SLAM to spatial perception with 3D dynamic scene graphs,

    A. Rosinol, A. Violette, M. Abate, N. Hughes, Y . Chang, J. Shi, A. Gupta, and L. Carlone, “Kimera: From SLAM to spatial perception with 3D dynamic scene graphs,”Int. J. Robot. Res., vol. 40, no. 12-14, pp. 1510–1546, 2021

    2021

  29. [29]

    S-graphs+: Real-time localization and mapping leveraging hierarchical representations,

    H. Bavle, J. L. Sanchez-Lopez, M. Shaheer, J. Civera, and H. V oos, “S-graphs+: Real-time localization and mapping leveraging hierarchical representations,”IEEE Robot. Autom. Lett., vol. 8, no. 8, pp. 4927–4934, 2023

    2023

  30. [30]

    Bridging the appearance gap: Multi-experience localization for long-term visual teach and repeat,

    M. Paton, K. MacTavish, M. Warren, and T. D. Barfoot, “Bridging the appearance gap: Multi-experience localization for long-term visual teach and repeat,” inProc. IEEE/RSJ Int. Conf. Intell. Robots Syst., 2016, pp. 1918–1925

    2016

  31. [31]

    Experience-based navigation for long- term localisation,

    W. Churchill and P. Newman, “Experience-based navigation for long- term localisation,”Int. J. Robot. Res., vol. 32, no. 14, pp. 1645–1661, 2013

    2013

  32. [32]

    Bags of binary words for fast place recognition in image sequences,

    D. G ´alvez-L´opez and J. D. Tardos, “Bags of binary words for fast place recognition in image sequences,”IEEE Trans. Robot., vol. 28, no. 5, pp. 1188–1197, 2012

    2012

  33. [33]

    NetVLAD: CNN architecture for weakly supervised place recognition,

    R. Arandjelovi ´c, P. Gron´at, A. Torii, T. Pajdla, and J. Sivic, “NetVLAD: CNN architecture for weakly supervised place recognition,” inProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2016, pp. 5297–5307

    2016

  34. [34]

    SeqSLAM: Visual route-based naviga- tion for sunny summer days and stormy winter nights,

    M. J. Milford and G. F. Wyeth, “SeqSLAM: Visual route-based naviga- tion for sunny summer days and stormy winter nights,” inProc. IEEE Int. Conf. Robot. Autom., 2012, pp. 1643–1649

    2012

  35. [35]

    Lazy data association for image se- quences matching under substantial appearance changes,

    O. Vysotska and C. Stachniss, “Lazy data association for image se- quences matching under substantial appearance changes,”IEEE Robot. Autom. Lett., vol. 1, no. 1, pp. 213–220, 2015

    2015

  36. [36]

    SeqNet: Learning descriptors for sequence- based hierarchical place recognition,

    S. Garg and M. Milford, “SeqNet: Learning descriptors for sequence- based hierarchical place recognition,”IEEE Robot. Autom. Lett., vol. 6, no. 3, pp. 4305–4312, 2021

    2021

  37. [37]

    Modeling perceptual aliasing in slam via discrete–continuous graphical models,

    P.-Y . Lajoie, S. Hu, G. Beltrame, and L. Carlone, “Modeling perceptual aliasing in slam via discrete–continuous graphical models,”IEEE Robot. Autom. Lett., vol. 4, no. 2, pp. 1232–1239, 2019

    2019

  38. [38]

    SlideSLAM: Sparse, lightweight, decen- tralized metric-semantic slam for multi-robot navigation,

    X. Liu, J. Lei, A. Prabhu, Y . Tao, I. Spasojevic, P. Chaudhari, N. Atanasov, and V . Kumar, “SlideSLAM: Sparse, lightweight, decen- tralized metric-semantic slam for multi-robot navigation,”arXiv preprint arXiv:2406.17249, 2024

    arXiv 2024

  39. [39]

    Kimera-Multi: Robust, distributed, dense metric-semantic slam for multi-robot systems,

    Y . Tian, Y . Chang, F. H. Arias, C. Nieto-Granda, J. P. How, and L. Carlone, “Kimera-Multi: Robust, distributed, dense metric-semantic slam for multi-robot systems,”IEEE Trans. Robot., vol. 38, no. 4, 2022

    2022

  40. [40]

    Complete solution clas- sification for the perspective-three-point problem,

    X. Gao, X. Hou, J. Tang, and H.-F. Cheng, “Complete solution clas- sification for the perspective-three-point problem,”IEEE Trans. Pattern Anal. Mach. Intell., vol. 25, pp. 930–943, 2003

    2003

  41. [41]

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

    M. A. Fischler and R. C. Bolles, “Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography,”Commun. ACM, vol. 24, no. 6, pp. 381–395, 1981

    1981

  42. [42]

    Multi-session slam with differentiable wide- baseline pose optimization,

    L. Lipson and J. Deng, “Multi-session slam with differentiable wide- baseline pose optimization,” inProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2024, pp. 19 626–19 635

    2024

  43. [43]

    CVI-SLAM—collaborative visual- inertial SLAM,

    M. Karrer, P. Schmuck, and M. Chli, “CVI-SLAM—collaborative visual- inertial SLAM,”IEEE Robot. Autom. Lett., vol. 3, no. 4, pp. 2762–2769, 2018

    2018

  44. [44]

    D2SLAM: Decentralized and distributed collaborative visual-inertial SLAM system for aerial swarm,

    H. Xu, P. Liu, X. Chen, and S. Shen, “D2SLAM: Decentralized and distributed collaborative visual-inertial SLAM system for aerial swarm,” IEEE Trans. Robot., 2024

    2024

  45. [45]

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

    A. Kendall, M. Grimes, and R. Cipolla, “PoseNet: A convolutional network for real-time 6-dof camera relocalization,” inProc. IEEE/CVF Int. Conf. Comput. Vis., 2015, pp. 2938–2946

    2015

  46. [46]

    R-SCoRe: Revisiting scene coordinate regression for robust large-scale visual localization,

    X. Jiang, F. Wang, S. Galliani, C. V ogel, and M. Pollefeys, “R-SCoRe: Revisiting scene coordinate regression for robust large-scale visual localization,” inProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2025

    2025

  47. [47]

    LaMAR: Benchmarking localization and mapping for augmented reality,

    P.-E. Sarlin, M. Dusmanu, J. L. Sch ¨onberger, P. Speciale, L. Gruber, V . Larsson, O. Miksik, and M. Pollefeys, “LaMAR: Benchmarking localization and mapping for augmented reality,” inProc. Eur . Conf. Comput. Vis., 2022, pp. 686–704

    2022

  48. [48]

    FusionPortableV2: A unified multi-sensor dataset for generalized SLAM across diverse platforms and scalable environments,

    H. Wei, J. Jiao, X. Hu, J. Yu, X. Xie, J. Wu, Y . Zhu, Y . Liu, L. Wang, and M. Liu, “FusionPortableV2: A unified multi-sensor dataset for generalized SLAM across diverse platforms and scalable environments,” Int. J. Robot. Res., p. 02783649241303525, 2024

    2024

  49. [49]

    CroCoDL: Cross-device collaborative dataset for localization,

    H. Blum, A. Mercurio, J. O’Reilly, T. Engelbracht, M. Dusmanu, M. Pollefeys, and Z. Bauer, “CroCoDL: Cross-device collaborative dataset for localization,” inProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2025, pp. 27 424–27 434

    2025

  50. [50]

    360Loc: A dataset and benchmark for omnidirectional visual local- ization with cross-device queries,

    H. Huang, C. Liu, Y . Zhu, H. Cheng, T. Braud, and S.-K. Yeung, “360Loc: A dataset and benchmark for omnidirectional visual local- ization with cross-device queries,” inProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2024, pp. 22 314–22 324

    2024

  51. [51]

    Xoftr: Cross-modal feature matching transformer,

    ¨O. Tuzcuo˘glu, A. K ¨oksal, B. Sofu, S. Kalkan, and A. A. Alatan, “Xoftr: Cross-modal feature matching transformer,” inProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2024, pp. 4275–4286

    2024

  52. [52]

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

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

    Pith/arXiv arXiv 2025

  53. [53]

    LoRA3D: Low-rank self-calibration of 3D geometric foundation mod- els,

    Z. Lu, H. Yang, D. Xu, B. Li, B. Ivanovic, M. Pavone, and Y . Wang, “LoRA3D: Low-rank self-calibration of 3D geometric foundation mod- els,” inInt. Conf. Learn. Represent., 2025

    2025

  54. [54]

    GV-Bench: Benchmarking local feature matching for geometric verification of long-term loop closure detection,

    J. Yu, H. Ye, J. Jiao, P. Tan, and H. Zhang, “GV-Bench: Benchmarking local feature matching for geometric verification of long-term loop closure detection,” inProc. IEEE/RSJ Int. Conf. Intell. Robots Syst., 2024, pp. 7922–7928

    2024

  55. [55]

    From coarse to fine: Robust hierarchical localization at large scale,

    P.-E. Sarlin, C. Cadena, R. Siegwart, and M. Dymczyk, “From coarse to fine: Robust hierarchical localization at large scale,” inProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2019

    2019

  56. [56]

    borglab/gtsam,

    F. Dellaert and G. Contributors, “borglab/gtsam,” 2022. [Online]. Available: https://github.com/borglab/gtsam)

    2022

  57. [57]

    MUSIQ: Multi- scale image quality transformer,

    J. Ke, Q. Wang, Y . Wang, P. Milanfar, and F. Yang, “MUSIQ: Multi- scale image quality transformer,” inProc. IEEE/CVF Int. Conf. Comput. Vis., 2021, pp. 5148–5157

    2021

  58. [58]

    SLiM: Scalable and lightweight lidar mapping in urban environments,

    Z. Yu, Z. Qiao, W. Liu, H. Yin, and S. Shen, “SLiM: Scalable and lightweight lidar mapping in urban environments,”IEEE Trans. Robot., 2025

    2025

  59. [59]

    VINS-Mono: A robust and versatile monocular visual-inertial state estimator,

    T. Qin, P. Li, and S. Shen, “VINS-Mono: A robust and versatile monocular visual-inertial state estimator,”IEEE Trans. Robot., vol. 34, no. 4, pp. 1004–1020, 2018

    2018

  60. [60]

    MS-Mapping: Multi- session lidar mapping with wasserstein-based keyframe selection,

    X. Hu, J. Wu, J. Jiao, W. Zhang, and P. Tan, “MS-Mapping: Multi- session lidar mapping with wasserstein-based keyframe selection,”arXiv preprint arXiv:2406.02096, 2024

    Pith/arXiv arXiv 2024

  61. [61]

    Watch your STEPP: Semantic traversability estimation using pose projected features,

    S. Ægidius, D. Hadjivelichkov, J. Jiao, J. Embley-Riches, and D. Kanoulas, “Watch your STEPP: Semantic traversability estimation using pose projected features,” inProc. IEEE Int. Conf. Robot. Autom., 2025, pp. 2376–2382

    2025

  62. [62]

    EigenPlaces: Training viewpoint robust models for visual place recognition,

    G. Berton, G. Trivigno, B. Caputo, and C. Masone, “EigenPlaces: Training viewpoint robust models for visual place recognition,” inProc. IEEE/CVF Int. Conf. Comput. Vis., 2023, pp. 11 080–11 090

    2023

  63. [63]

    AnyLoc: Towards universal visual place recognition,

    N. Keetha, A. Mishra, J. Karhade, K. M. Jatavallabhula, S. Scherer, M. Krishna, and S. Garg, “AnyLoc: Towards universal visual place recognition,”IEEE Robot. Autom. Lett., vol. 9, no. 2, pp. 1286–1293, 2023

    2023

  64. [64]

    Megaloc: One retrieval to place them all,

    G. Berton and C. Masone, “Megaloc: One retrieval to place them all,” in Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2025, pp. 2861– 2867

    2025

  65. [65]

    Falco: Fast likelihood- based collision avoidance with extension to human-guided navigation,

    J. Zhang, C. Hu, R. G. Chadha, and S. Singh, “Falco: Fast likelihood- based collision avoidance with extension to human-guided navigation,” J. Field Robot., vol. 37, no. 8, pp. 1300–1313, 2020

    2020

  66. [66]

    Matterport3D: Learning from RGB- D data in indoor environments,

    A. Chang, A. Dai, T. Funkhouser, M. Halber, M. Niebner, M. Savva, S. Song, A. Zeng, and Y . Zhang, “Matterport3D: Learning from RGB- D data in indoor environments,” inProc. Int. Conf. 3D Vis., 2017, pp. 667–676

    2017

  67. [67]

    iPlanner: Imperative path planning,

    F. Yang, C. Wang, C. Cadena, and M. Hutter, “iPlanner: Imperative path planning,”arXiv preprint arXiv:2302.11434, 2023

    arXiv 2023

  68. [68]

    State estimation for legged robots-kinematics, inertial sensing, and computer vision,

    M. Bloesch, “State estimation for legged robots-kinematics, inertial sensing, and computer vision,” Ph.D. dissertation, ETH Zurich, 2017

    2017

  69. [69]

    Benchmarking egocentric visual-inertial SLAM at city scale,

    A. Krishnan, S. Liu, P.-E. Sarlin, O. Gentilhomme, D. Caruso, M. Monge, R. Newcombe, J. Engel, and M. Pollefeys, “Benchmarking egocentric visual-inertial SLAM at city scale,” inProc. IEEE/CVF Int. Conf. Comput. Vis., 2025, pp. 25 207–25 217

    2025

  70. [70]

    MARViN: Mobile AR dataset with visual-inertial data,

    C. Liu, Y . Zhao, and T. Braud, “MARViN: Mobile AR dataset with visual-inertial data,” inProc. IEEE Conf. Virtual Real. 3D User Inter- faces Abstr . Workshops, 2024, pp. 532–538

    2024

  71. [71]

    Supplementary Materials of OpenNavMap: Structure-Free Topometric Mapping via Large-Scale Collaborative Localization,

    J. Jiaoet al., “Supplementary Materials of OpenNavMap: Structure-Free Topometric Mapping via Large-Scale Collaborative Localization,” 2025

    2025

  72. [72]

    A tutorial on quantitative trajectory evaluation for visual (-inertial) odometry,

    Z. Zhang and D. Scaramuzza, “A tutorial on quantitative trajectory evaluation for visual (-inertial) odometry,” inProc. IEEE/RSJ Int. Conf. Intell. Robots Syst., 2018, pp. 7244–7251

    2018

  73. [73]

    End-to-end learning of deep visual representations for image retrieval,

    A. Gordo, J. Almazan, J. Revaud, and D. Larlus, “End-to-end learning of deep visual representations for image retrieval,”Int. J. Comput. Vis., vol. 124, no. 2, pp. 237–254, 2017. 21

    2017

  74. [74]

    DISK: Learning local features with policy gradient,

    M. Tyszkiewicz, P. Fua, and E. Trulls, “DISK: Learning local features with policy gradient,” inAdv. Neural Inf. Process. Syst., vol. 33, 2020, pp. 14 254–14 265

    2020

  75. [75]

    SuperPoint: Self- supervised interest point detection and description,

    D. DeTone, T. Malisiewicz, and A. Rabinovich, “SuperPoint: Self- supervised interest point detection and description,” inProc. IEEE Conf. Comput. Vis. Pattern Recognit. Workshops, 2018, pp. 224–236

    2018

  76. [76]

    LightGlue: Local feature matching at light speed,

    P. Lindenberger, P.-E. Sarlin, and M. Pollefeys, “LightGlue: Local feature matching at light speed,” inProc. IEEE/CVF Int. Conf. Comput. Vis., 2023, pp. 17 627–17 638

    2023

  77. [77]

    Reloc3r: Large-scale training of relative camera pose regression for generalizable, fast, and accurate visual localization,

    S. Dong, S. Wang, S. Liu, L. Cai, Q. Fan, J. Kannala, and Y . Yang, “Reloc3r: Large-scale training of relative camera pose regression for generalizable, fast, and accurate visual localization,” inProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2025, pp. 16 739–16 752

    2025

  78. [78]

    MV-Dust3r+: Single-stage scene reconstruction from sparse views in 2 seconds,

    Z. Tang, Y . Fan, D. Wang, H. Xu, R. Ranjan, A. Schwing, and Z. Yan, “MV-Dust3r+: Single-stage scene reconstruction from sparse views in 2 seconds,” inProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2025, pp. 5283–5293

    2025

  79. [79]

    LoFTR: Detector-free local feature matching with transformers,

    J. Sun, Z. Shen, Y . Wang, H. Bao, and X. Zhou, “LoFTR: Detector-free local feature matching with transformers,” inProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2021, pp. 8922–8931

    2021

  80. [80]

    SceneGraphLoc: Cross-modal coarse visual localization on 3D scene graphs,

    Y . Miao, F. Engelmann, O. Vysotska, F. Tombari, M. Pollefeys, and D. B. Bar ´ath, “SceneGraphLoc: Cross-modal coarse visual localization on 3D scene graphs,” inProc. Eur . Conf. Comput. Vis., 2024, pp. 127– 150

    2024

Showing first 80 references.

This paper was first reviewed by deepseek-v4-flash on August 3, 2026.