Pith. sign in

REVIEW 3 major objections 6 minor 23 references

Spatiotemporal Contrastive Learning for Cross-View Video Localization in Unstructured Off-road Terrains

T0 review · 3 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A self-supervised matcher trained on under 30 minutes of driving video keeps off-road vehicles within 25 meters of ground truth 93% of the time, using outdated satellite maps.

desk verdict Good new method for season-robust cross-view video localization, but the global-localization claim outruns the evaluation. read the letter →

arxiv 2506.05250 v1 pith:EUTJ7MWL submitted 2025-06-05 cs.CV cs.RO

Zhiyun Deng , Dongmyeong Lee , Amanda Adkins , Jesse Quattrociocchi , Christian Ellis , Joydeep Biswas This is my paper
classification cs.CVcs.RO
keywords cross-viewlocalizationself-supervisedcontrastivelearningoff-roadnavigationMonteCarlosatelliteimagerytemporalrobustnessvideoGPS-denied
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

The paper tries to establish that cross-view localization in GPS-denied off-road terrain does not need handcrafted models, environment-specific tuning, or large training sets: a self-supervised contrastive network trained on less than 30 minutes of GPS-tagged driving video can match ground-level imagery to outdated satellite maps and keep a particle filter within 25 meters of ground truth for 93% of the time over 12.29 km of unseen trails. The central move is to make the learned representation simultaneously invariant to viewpoint and season while remaining sensitive to heading, so it can resolve the repetitive-vegetation ambiguity that color-based augmentation cannot. If true, this gives off-road robots a practical localization fallback that uses only an uncalibrated monocular camera, standard odometry, and publicly available historical satellite imagery.

What carries the argument

The load-bearing object is MoViX's spatiotemporal contrastive matcher: a dual-branch network, a frozen Vision Transformer backbone plus trainable alignment layers, that maps a ground-view clip and an aerial patch to a similarity score, trained so that similarity tracks geospatial proximity and heading alignment rather than appearance. Three mechanisms carry the argument: pose-dependent positive sampling with temporal-robustness enforcement, which draws hard positive aerial patches across seasons at the true pose and heading; temporal-matched hard negative mining, which restricts negatives to the same season within 5–40 meters and at least 30 degrees away in heading; and a quality-aware temporal aggregator that pools frame embeddings with soft attention weights plus an entropy penalty. At inference these feed a neural-augmented Monte Carlo Localization filter whose likelihood-sharpening temperature is modulated by the kernel-density-estimate entropy of the particle distribution, letting the filter keep multiple hypotheses in ambiguous vegetation and sharpen when reliable evidence arrives.

What would settle it

Initialize the same particle filter uniformly over a large region, for example one kilometer square, instead of with Gaussian noise around a prior pose, and measure the 25-meter and 50-meter success rates; if those rates collapse, the reported numbers depend on the spatial prior rather than on global cross-view matching.

Watch

Extended reading notes

Core claim

On its own terms, MoViX's central claim is that a self-supervised spatiotemporal contrastive objective can learn cross-view features that survive both viewpoint change and seasonal shift. Training pairs a ground-view anchor with hard positive aerial patches sampled from multiple seasons at the true pose and heading, and hard negatives drawn from the same season within 5–40 meters with a heading difference of at least 30 degrees, which forces the encoder to encode geometry and orientation rather than seasonal appearance. A motion-informed sampler selects spatially uniform frames, and a learned soft-attention aggregator weights them by geometric alignment before pooling. At inference the same matcher serves as the measurement model of a 300-particle Monte Carlo filter, with the softmax temperature scaled down by the spatial entropy of the belief. The paper reports that on 12.29 km of test trajectories unseen during training, using satellite imagery from seasons other than the test season, the full system stays within 25 meters of ground truth 93% of the time and within 50 meters 100% of the time, with mean absolute trajectory error around 8–11 meters depending on area.

Load-bearing premise

The method assumes that when the filter starts, the vehicle is already known to be somewhere near its true position; the reported accuracy is measured from that starting guess, not from a completely unknown location.

Editorial extensions

If this is right

  • A vehicle with noisy odometry and a monocular camera can keep a bounded position error against an outdated satellite map, giving GPS-denied off-road navigation a learned cross-view fallback.
  • Under 30 minutes of GPS-tagged driving data suffices to train the matcher, so per-site deployment reduces to collecting one short labeled drive.
  • Multi-season training with temporal-matched negatives beats aggressive color augmentation, implying that long-term structural appearance change must be learned from data rather than simulated.
  • Entropy-scaled likelihoods let the filter track multiple pose hypotheses through ambiguous vegetation and collapse to one when evidence is strong, reducing the risk of committing to a wrong match.
  • The paper's Baltimore experiments indicate the same matcher transfers across geographic sites and robot platforms with no retuning.

Reading between the lines

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

  • Editorial inference: the experimental protocol measures local tracking from a prior pose, not global localization from an arbitrary start; a uniform-prior experiment would reveal how much of the reported robustness comes from the learned matcher versus filter convergence.
  • Editorial inference: because the camera is uncalibrated and only RGB video plus odometry are required, the same training recipe could be transferred to drones or hand-held video by collecting a short GPS-tagged pass through an area.
  • Editorial inference: the temporal-matched negative rule implies a boundary condition; if the reference satellite imagery's season is absent from the multi-season training set, the model may have no invariant structure to latch onto, so testing on a completely unseen season would bound the method's temporal extrapolation.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper proposes MoViX, a self-supervised cross-view video localization method for off-road, GPS-denied environments. It learns viewpoint- and season-invariant embeddings from GPS-tagged ground video paired with outdated satellite imagery, using pose-dependent positive sampling, temporally matched hard negative mining, motion-informed frame sampling, and a quality-aware temporal aggregator. At inference, MoViX is embedded in a Monte Carlo Localization framework with entropy-guided temperature scaling. The method is evaluated on TartanDrive 2.0 with under 30 minutes of training data and on a real-world Baltimore off-road dataset, reporting success rates of 93% within 25 m and 100% within 50 m in unseen regions with outdated satellite imagery, along with ablations and comparisons to GAMa and GAReT.

Significance. If the reported results hold, MoViX would make a meaningful contribution to off-road visual localization: it learns cross-view, season-invariant representations in a self-supervised manner from a small amount of driving data, integrates a learned observation model into a principled particle-filter pipeline, and demonstrates robustness to temporal mismatch in satellite imagery. The manuscript's strengths include an extensive experimental design with multiple baselines, two test settings, and carefully constructed ablations that isolate the contributions of temporal robustness, temporal encoding, and color augmentation. The authors also provide qualitative likelihood-map analyses and plan to release code, which supports reproducibility. However, the significance of the central claim is currently limited by the evaluation protocol: the experiments assume a strong spatial and orientation prior at inference, so the paper does not yet substantiate the advertised GPS-denied global-localization capability.

major comments (3)
  1. [Section IV-C, Tables I and II] The central claim that MoViX achieves 'GPS-denied' localization in 'unseen regions' is not supported by the evaluation protocol. Section IV-A states that particles are 'initialized with Gaussian noise around the prior pose,' and Section IV-B evaluates frame-level matching only within a 150×150 m satellite map centered on the ground-truth location, using a 'rough camera orientation as a prior.' This setup measures the method's ability to correct small pose errors in a locally centered search region, not its ability to localize globally from an arbitrary or unknown initial pose. The Gaussian covariance of the prior, the source of the prior pose at deployment, and its expected error are never specified. If the prior is within a few tens of meters, the reported 93%/100% success rates reflect local tracking and data association rather than global localization. To support the stated claim, the authors should report results with particles initialized uniformly over a large map (e.g., the full test area) or, at minimum, characterize performance as a function of prior uncertainty, and adjust the abstract and conclusion accordingly.
  2. [Section IV-C, end of Section IV] The empirical comparison as presented does not unambiguously support the abstract's statement that MoViX 'outperforms state-of-the-art baselines.' In Table I, in the intra-area same-season setting, GAReT achieves a lower mean ATE (7.2 vs. 8.3) and a higher SR@25 (100% vs. 99.0%) than the full MoViX model. MoViX outperforms GAReT mainly in cross-season and cross-area settings, which is an important result but should be stated with appropriate qualification. Furthermore, Tables I and II report single-run metrics with no standard deviations or multiple seeds. Given that the differences between MoViX and its ablations (and sometimes baselines) are often small, the absence of variance estimates leaves the ordering potentially fragile. The authors should provide results across multiple seeds with means and standard deviations, or at least clearly label the reported numbers as single-run outcomes.
  3. [Section IV-C, last paragraph] The claimed generalization to a real-world off-road dataset is not quantitatively evaluated in the manuscript. After presenting Table II, the text says that MoViX is 'showcased' on the Baltimore dataset 'in the accompanying video,' but no success-rate, ATE, or other numerical results are reported for that dataset. Since this is a central part of the claimed contribution ('We further demonstrate generalization on a real-world off-road dataset from a geographically distinct site with a different robot platform' in the abstract), the authors should include quantitative results, even as a summarized table, for the Baltimore trials. Without such numbers, the generalizability claim rests only on a qualitative demonstration.
minor comments (6)
  1. [Section III-A] In Eq. (8), the symbol p0 is used in the interpolation formula but is not defined in the surrounding text; the notation should be clarified, perhaps as the starting pose of the sampled trajectory segment.
  2. [Section IV-A] The text repeatedly refers to a 'temporal encoding τ' in the loss, but Eq. (6) and the surrounding description do not define how τ enters the objective. Please specify the exact form of the temporal encoding and where it appears in the loss.
  3. [Tables I and II] The inference pipeline states that particles are propagated using 'noisy odometry, injecting additional translation and rotation noise,' but the noise magnitudes and the Gaussian initialization covariance are not specified. Reporting these values is important for reproducibility and for interpreting the sensitivity of the MCL results to the prior.
  4. [References] The table headings contain typographical errors: 'WITHOUTDATEDSATELLITEIMAGEGY' should be 'WITH OUTDATED SATELLITE IMAGERY,' and 'OF'TRAJECTORYESTIMATION' is missing a space. These should be corrected.
  5. [Section IV-A] Reference [12] is cited as page 12606 only, which appears to be a single-page citation for a full paper; please verify the page range or the appropriate citation details.
  6. [Section II] The phrase 'training on less than 30 minutes of driving data' is clear, but the total number of training trajectories and the split into intra-area and cross-area test sets would benefit from a small table or explicit enumeration in the text, since Fig. 3 is not self-explanatory from the caption alone.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the learning signal comes from external GPS/odometry, and evaluation uses held-out seasons and areas; the pose-prior protocol is a deployment caveat, not a circular derivation.

full rationale

MoViX's derivation chain is self-contained. Training correspondences come from time-synchronized camera, odometry, and GPS data, with positive aerial patches selected from the robot's true pose and hard negatives constrained by spatial and orientation margins. The contrastive losses (triplet loss and temporal aggregation loss) are standard objectives that align ground and aerial embeddings without encoding the target localization result. Evaluation is performed on held-out trajectories, unseen areas, and seasonally mismatched satellite imagery, so the reported success rates are not recomputed from the training labels. No load-bearing self-citation appears in the paper: the authors do not invoke their own prior results as a uniqueness theorem or as justification for the central architecture. The particle filter is initialized around a prior pose, and the frame-level map search is centered on ground truth; this is a legitimate evaluation-protocol concern about whether the experiments demonstrate global localization rather than local tracking. However, that concern is about external validity and unspecified deployment assumptions, not about the paper's predictions reducing by construction to its inputs. A prior near the true pose can make localization easier, but the pose prior is not a fitted parameter and the success metrics are not defined in terms of it. Therefore, no circular step meeting the evidentiary standard can be identified, and the appropriate finding is no significant circularity.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

No new physical or latent entities are introduced. The temporal encoding tau is a conditioning label, not a new entity or conserved quantity. The main hand-chosen quantities are contrastive loss hyperparameters, negative mining thresholds, and inference parameters that are partially unreported.

free parameters (7)
  • Triplet margin alpha = 0.2
    Eq. (6), chosen by hand; controls the required separation between positive and negative aerial embeddings.
  • Hard negative spatial bounds (d_min, d_max) = 5 to 40 meters
    Eq. (7), chosen to define spatial hardness of negatives; no sensitivity analysis is reported.
  • Heading difference threshold delta_theta = 30 degrees
    Eq. (7), chosen to enforce orientation hardness for negatives.
  • Loss weights lambda_sim and lambda_H = 0.1 each
    Eq. (14), balance direct similarity maximization and attention entropy regularization; values appear chosen without a reported grid search.
  • MCL temperature scaling lambda_base and gamma = not reported
    Eq. (17), control how belief entropy modulates observation likelihood sharpness; required for inference but values are not given.
  • KDE bandwidth h = not reported
    Eq. (15), needed to compute the spatial belief entropy used in Eq. (17); value not specified.
  • Frame attention temperature beta = not reported
    Eq. (11), controls sharpness of the temporal aggregator's soft attention weights; value not specified.
assumptions (5)
  • domain assumption Frozen ViT-B/16 pretrained on ImageNet provides a sufficiently strong visual backbone for cross-view matching in off-road terrain.
    Used throughout; no experiments with other backbone sizes or training regimes are reported.
  • domain assumption Google Maps Static API satellite imagery is accurately georeferenced and the stated season dates are correct.
    Ground-aerial positive construction and all evaluation metrics depend on this georeferencing accuracy.
  • domain assumption TartanDrive 2.0 GPS and odometry ground truth is accurate enough for both training supervision and evaluation.
    Training positives and the reported ATE/SR metrics all rely on this assumed accuracy.
  • standard math Monte Carlo localization with 300 particles and Gaussian-injected motion noise approximates the true posterior in these environments.
    Standard MCL assumption invoked in Sections II-D and IV-A, following Thrun et al.
  • domain assumption The specific historical satellite images used for training and the fall images used for testing are representative of seasonal appearance variation.
    The temporal robustness claim rests on this representativeness; a different seasonal distribution could change results.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Spatiotemporal Contrastive Learning for Cross-View Video Localization in Unstructured Off-road Terrains." pith.science (2026). https://pith.science/paper/EUTJ7MWL

@misc{pith2026250605250,
  author       = {Pith},
  title        = {Pith review of: Spatiotemporal Contrastive Learning for Cross-View Video Localization in Unstructured Off-road Terrains},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EUTJ7MWL}},
  note         = {Machine review of arXiv:2506.05250}
}
read the original abstract

Robust cross-view 3-DoF localization in GPS-denied, off-road environments remains challenging due to (1) perceptual ambiguities from repetitive vegetation and unstructured terrain, and (2) seasonal shifts that significantly alter scene appearance, hindering alignment with outdated satellite imagery. To address this, we introduce MoViX, a self-supervised cross-view video localization framework that learns viewpoint- and season-invariant representations while preserving directional awareness essential for accurate localization. MoViX employs a pose-dependent positive sampling strategy to enhance directional discrimination and temporally aligned hard negative mining to discourage shortcut learning from seasonal cues. A motion-informed frame sampler selects spatially diverse frames, and a lightweight temporal aggregator emphasizes geometrically aligned observations while downweighting ambiguous ones. At inference, MoViX runs within a Monte Carlo Localization framework, using a learned cross-view matching module in place of handcrafted models. Entropy-guided temperature scaling enables robust multi-hypothesis tracking and confident convergence under visual ambiguity. We evaluate MoViX on the TartanDrive 2.0 dataset, training on under 30 minutes of data and testing over 12.29 km. Despite outdated satellite imagery, MoViX localizes within 25 meters of ground truth 93% of the time, and within 50 meters 100% of the time in unseen regions, outperforming state-of-the-art baselines without environment-specific tuning. We further demonstrate generalization on a real-world off-road dataset from a geographically distinct site with a different robot platform.

Figures

Figures reproduced from arXiv: 2506.05250 by the authors.

Figure 1
Figure 1. Overview of our proposed approach, MoViX. (A) A motion [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Network Architecture with A  p [n] t  denoting the aerial patch at p [n] t . Weights are updated by: w [n] t ∝ exp  λt s [n] t  , (4) where λt adjusts the sharpness of the measurement likelihood. The final pose estimate is computed as the weighted average of all particle hypotheses: pˆt = X M n=1 w [n] t p [n] t . (5) While the frame-level localization selects the most likely pose via an argmax (Eq. (1)) over si… view at source ↗
Figure 3
Figure 3. Training, Validation, and Test Split from the TartanDrive 2.0 dataset. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (3 more)
Figure 5
Figure 5. Figure 5: Orientation-conditioned likelihood maps generated by our method [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 4
Figure 4. Figure 4: An orientation-conditioned likelihood map that approximates [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 6
Figure 6. Figure 6: Visualization of frame quality weights from the temporal aggregation [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

23 extracted references · 17 canonical work pages

  1. [1]

    Fine-grained cross- view geo-localization using a correlation-aware homography estimator.Advances in Neural Information Processing Systems, 36:5301–5319, 2023

    Xiaolong Wang, Runsen Xu, Zhuofan Cui, Zeyu Wan, and Yu Zhang. Fine-grained cross- view geo-localization using a correlation-aware homography estimator.Advances in Neural Information Processing Systems, 36:5301–5319, 2023

    2023

  2. [2]

    Boosting 3-dof ground- to-satellite camera localization accuracy via geometry-guided cross-view transformer

    Yujiao Shi, Fei Wu, Akhil Perincherry, Ankit V ora, and Hongdong Li. Boosting 3-dof ground- to-satellite camera localization accuracy via geometry-guided cross-view transformer. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 21516– 21526, 2023

    work page 2023

  3. [3]

    Bevrender: Vision-based cross-view vehicle registration in off-road gnss-denied environment

    Lihong Jin, Wei Dong, Wenshan Wang, and Michael Kaess. Bevrender: Vision-based cross-view vehicle registration in off-road gnss-denied environment. In2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 11032–11039. IEEE, 2024

    work page 2024

  4. [4]

    Bevloc: Cross-view localization and matching via birds-eye-view synthesis

    Christopher Klammer and Michael Kaess. Bevloc: Cross-view localization and matching via birds-eye-view synthesis. In2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 5656–5663. IEEE, 2024

    work page 2024

  5. [5]

    Orienternet: Visual localization in 2d public maps with neural matching

    Paul-Edouard Sarlin, Daniel DeTone, Tsun-Yi Yang, Armen Avetisyan, Julian Straub, Tomasz Malisiewicz, Samuel Rota Bulo, Richard Newcombe, Peter Kontschieder, and Vasileios Balntas. Orienternet: Visual localization in 2d public maps with neural matching. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 21632–21642, 2023

    work page 2023

  6. [6]

    Coming down to earth: Satellite-to-street view synthesis for geo-localization

    Aysim Toker, Qunjie Zhou, Maxim Maximov, and Laura Leal-Taix´e. Coming down to earth: Satellite-to-street view synthesis for geo-localization. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6488–6497, 2021

    work page 2021

  7. [7]

    Wide-area geolocalization with a limited field of view camera

    Lena M Downes, Ted J Steiner, Rebecca L Russell, and Jonathan P How. Wide-area geolocalization with a limited field of view camera. In2023 IEEE International Conference on Robotics and Automation (ICRA), pages 10594–10600. IEEE, 2023

    work page 2023

  8. [8]

    Gama: Cross-view video geo-localization

    Shruti Vyas, Chen Chen, and Mubarak Shah. Gama: Cross-view video geo-localization. In European Conference on Computer Vision, pages 440–456. Springer, 2022

    work page 2022

Show all 23 references
  1. [9]

    Garet: Cross-view video geolocalization with adapters and auto-regressive transformers

    Manu S Pillai, Mamshad Nayeem Rizve, and Mubarak Shah. Garet: Cross-view video geolocalization with adapters and auto-regressive transformers. InEuropean Conference on Computer Vision, pages 466–483. Springer, 2024

    2024

  2. [10]

    A cross-view geo-localization algorithm using uav image and satellite image.Sensors, 24(12):3719, 2024

    Jiqi Fan, Enhui Zheng, Yufei He, and Jianxing Yang. A cross-view geo-localization algorithm using uav image and satellite image.Sensors, 24(12):3719, 2024

    2024

  3. [11]

    Accurate 3-dof camera geo-localization via ground-to-satellite image matching.IEEE transactions on pattern analysis and machine intelligence, 45(3):2682–2697, 2022

    Yujiao Shi, Xin Yu, Liu Liu, Dylan Campbell, Piotr Koniusz, and Hongdong Li. Accurate 3-dof camera geo-localization via ground-to-satellite image matching.IEEE transactions on pattern analysis and machine intelligence, 45(3):2682–2697, 2022

    2022

  4. [12]

    Tartandrive 2.0: More modalities and better infrastructure to further self-supervised learning research in off-road driving tasks

    Matthew Sivaprakasam, Parv Maheshwari, Mateo Guaman Castro, Samuel Triest, Micah Nye, Steve Willits, Andrew Saba, Wenshan Wang, and Sebastian Scherer. Tartandrive 2.0: More modalities and better infrastructure to further self-supervised learning research in off-road driving ta...

    2024

  5. [13]

    MIT Press, 2005

    Sebastian Thrun, Wolfram Burgard, and Dieter Fox.Probabilistic Robotics. MIT Press, 2005

    2005

  6. [14]

    Patch-netvlad: Multi-scale fusion of locally-global descriptors for place recognition

    Stephen Hausler, Sourav Garg, Ming Xu, Michael Milford, and Tobias Fischer. Patch-netvlad: Multi-scale fusion of locally-global descriptors for place recognition. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 14141–14152, 2021

    2021

  7. [15]

    Boosting contrastive self-supervised learning with false negative cancellation

    Tri Huynh, Simon Kornblith, Matthew R Walter, Michael Maire, and Maryam Khademi. Boosting contrastive self-supervised learning with false negative cancellation. InProceedings of the IEEE/CVF winter conference on applications of computer vision, pages 2785–2795, 2022

    2022

  8. [16]

    Spatiotemporal contrastive video representation learning

    Rui Qian, Tianjian Meng, Boqing Gong, Ming-Hsuan Yang, Huisheng Wang, Serge Belongie, and Yin Cui. Spatiotemporal contrastive video representation learning. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 6964–6974, 2021

    2021

  9. [17]

    Kld-sampling: Adaptive particle filters.Advances in neural information processing systems, 14, 2001

    Dieter Fox. Kld-sampling: Adaptive particle filters.Advances in neural information processing systems, 14, 2001

    2001

  10. [18]

    Google maps static api, 2024

    Google LLC. Google maps static api, 2024. https://developers.google.com/maps/ documentation/maps-static

    2024

  11. [19]

    Improved deep metric learning with multi-class n-pair loss objective.Advances in neural information processing systems, 29, 2016

    Kihyuk Sohn. Improved deep metric learning with multi-class n-pair loss objective.Advances in neural information processing systems, 29, 2016

    2016

  12. [20]

    Tartanvo: A generalizable learning-based vo

    Wenshan Wang, Yaoyu Hu, and Sebastian Scherer. Tartanvo: A generalizable learning-based vo. InConference on Robot Learning, pages 1761–1772. PMLR, 2021

    2021

  13. [21]

    Super odometry: Imu-centric lidar-visual-inertial estimator for challenging environments

    Shibo Zhao, Hengrui Zhang, Peng Wang, Lucas Nogueira, and Sebastian Scherer. Super odometry: Imu-centric lidar-visual-inertial estimator for challenging environments. In2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 8729–

  14. [22]

    An image is worth 16x16 words: Transformers for image recognition at scale.arXiv preprint arXiv:2010.11929, 2020

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, et al. An image is worth 16x16 words: Transformers for image recognition at scale.arXiv preprint arXiv:...

    2010 arXiv

  15. [23]

    Dinov2: Learning robust visual features without supervision.arXiv preprint arXiv:2304.07193, 2023

    Maxime Oquab, Timoth´ee Darcet, Th´eo Moutakanni, Huy V o, Marc Szafraniec, Vasil Khali- dov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, et al. Dinov2: Learning robust visual features without supervision.arXiv preprint arXiv:2304.07193, 2023

    2023 arXiv

Pith tools

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