Pith. sign in

REVIEW 4 major objections 5 minor 45 references

VKFPos: A Learning-Based Monocular Positioning with Variational Bayesian Extended Kalman Filter Integration

T0 review · 4 major / 5 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read The paper argues that monocular positioning posterior decomposes into absolute and relative pose factors, and fusing the two learned estimators with an extended Kalman filter outperforms temporal pose regression.

desk verdict VKFPos is a workmanlike EKF-based APR+RPR fusion with competitive numbers, but the missing KFNet comparison, a faulty proof, and an unvalidated independence assumption keep it from being a clear advance. read the letter →

arxiv 2501.18994 v1 pith:6C3UIK5B submitted 2025-01-31 cs.CV cs.AI

Jian-Yu Chen , Yi-Ru Chen , Yin-Qiao Chang , Che-Ming Li , Jann-Long Chern , Chih-Wei Huang This is my paper · ORCID
classification cs.CVcs.AI
keywords monocularpositioningabsoluteposeregressionrelativeextendedKalmanfiltervariationalBayesianinferencecovariancelearningcamerarelocalizationse(3)representation
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

VKFPos sets out to solve monocular camera positioning—locating a camera from single or consecutive images—by fusing two learned predictions usually used separately: absolute pose regression (APR), which locates the camera in a global frame but is noisy, and relative pose regression (RPR), which tracks motion between frames but accumulates drift. The paper's central claim is that the posterior distribution of the camera pose factorizes into an APR likelihood and an RPR transition prior, so each branch can be trained independently with its own predicted covariance and then combined optimally by an extended Kalman filter. If this decomposition holds, it turns model-based integration—which the paper finds unstable—into a principled Bayesian calculation cheap enough for real-time use. On the indoor 7-Scenes and outdoor Oxford RobotCar benchmarks, the authors report that the single-shot branch matches state-of-the-art APR accuracy while the temporal output beats temporal APR and RNN-based integration methods in translation and rotation accuracy.

What carries the argument

The load-bearing object is the posterior factorization of Eq. (6), $p(\hat{x}_t | \hat{x}_{t-1}, u_{t,t-1}, z_t) \propto p(z_t | \hat{x}_t) \, p(u_{t,t-1} | \hat{x}_t, \hat{x}_{t-1})$, which turns the extended Kalman filter into a Bayes-rule product of learned factors. Each branch approximates its factor with a Gaussian $q(\cdot)$ on $\mathfrak{se}(3)$ whose diagonal covariance is predicted by the network; the per-sample negative-log-likelihood losses of Eq. (14) and Eq. (15) train mean and covariance together. In the EKF, the prediction step composes the previous estimate with the relative pose through a manifold update $\oplus$, and the correction step treats the absolute pose as an identity measurement. The learned covariances enter directly as $\Sigma_{u,t-1}$ in the covariance prediction and as $\Sigma_{z,t}$ in the Kalman gain, so the filter's trust in each branch is learned from data.

What would settle it

On a held-out trajectory, compare the relative-pose branch's predicted covariance $\Sigma_{u,t-1}$ with the empirical covariance of its residuals against ground-truth relative poses. If the predicted uncertainty is systematically off-scale or the residuals have non-zero mean, the EKF-corrected trajectory should degrade relative to the APR-only branch, which would show that the learned control-input model is not statistically consistent enough for the filter to do the work the paper claims.

Watch

Extended reading notes

Core claim

At the heart of the paper is Theorem 1: when the current state is conditioned on the latest measurement, the monocular positioning posterior $p(x_t | x_{t-1}, u_{t,t-1}, z_t)$ is proportional to $p(z_t | x_t) \, p(u_{t,t-1} | x_t, x_{t-1})$. The paper reads this as a decomposition of the positioning problem into an absolute-pose likelihood and a relative-pose transition, and it realizes each factor as a neural network that outputs a Gaussian mean and covariance on the Lie algebra $\mathfrak{se}(3)$. The branches are trained by minimizing their negative log-likelihood losses, so the covariances are learned rather than hand-tuned and then serve directly as the measurement noise and process noise in an extended Kalman filter. With the APR prediction as measurement and the RPR prediction as control input, the EKF produces a temporally smoothed pose; the reported consequence is single-shot accuracy on par with AtLoc and temporal accuracy that beats MapNet, AtLoc+, ViPR, and LSTM-KF on both benchmark datasets.

Load-bearing premise

The relative-pose branch's predicted means and covariances are trustworthy enough to serve as the filter's motion model, even though that branch is trained only to minimize its own likelihood loss and is never calibrated for filtering.

Editorial extensions

If this is right

  • A single-shot APR branch trained with covariance prediction can match attention-based APR baselines, so adding uncertainty outputs does not sacrifice accuracy.
  • Temporal positioning with EKF fusion beats temporal APR methods (MapNet, AtLoc+) and RNN-based integration (LSTM-KF, ViPR), indicating the learned filter, not extra trainable recurrency, is the effective ingredient.
  • Because the APR and RPR branches are trained independently, one branch can be swapped or upgraded without retraining the entire system.
  • The predicted covariances give a confidence estimate for each pose, which can inform downstream decisions in navigation, loop closure, and sensor fusion.
  • EKF-based fusion is computationally lighter than pose-graph optimization, making the accuracy gain available in real-time applications.

Reading between the lines

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

  • This decomposition suggests a general recipe: any learned measurement model and any learned motion model that output calibrated Gaussians could be fused in the same way, independent of the particular APR and RPR architectures.
  • A natural stress test is to replace the learned RPR branch with a visual odometry or inertial motion estimate; the EKF equations do not depend on how the relative pose is generated.
  • The diagonal-covariance assumption is a simplifying restriction; permitting correlations between translation and rotation might change how much the correction step trusts each branch.
  • Part of the reported gain may come from the covariance weighting acting as learned loss weights; isolating that effect would require ablating covariance prediction while keeping the EKF integration fixed.
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

4 major / 5 minor

Summary. The paper proposes VKFPos, a monocular camera positioning method that combines an Absolute Pose Regression (APR) branch, which regresses the 6DoF pose of a single image and its covariance, with a Relative Pose Regression (RPR) branch, which regresses the relative pose between consecutive images and its covariance. The two branches are trained independently with negative log-likelihood losses, and their outputs are fused at inference time through an Extended Kalman Filter (EKF) in which the RPR output acts as the control input and the APR output acts as the measurement. The authors claim that the posterior of the positioning problem decomposes into APR and RPR components, that the covariance predictions improve generalization, and that the EKF integration yields superior temporal positioning accuracy on the 7-Scenes and Oxford RobotCar datasets compared with existing temporal APR and model-based integration methods.

Significance. If the claims hold, VKFPos would offer a practical and lightweight integration of learned absolute and relative pose estimators, with the important practical virtue of being trainable without a filtering objective and of releasing source code. The covariance prediction is a useful ingredient, and the experimental results on two standard benchmarks are encouraging. However, the paper's central theoretical justification contains a flawed proof, the key conditional-independence assumption is not validated, and the experimental comparison omits the most closely related Kalman-filter baseline. The contribution is therefore potentially valuable but is not yet established at the level claimed.

major comments (4)
  1. [Theorem 1 and Eqs. (8)-(11)] The proof of Theorem 1 is not mathematically correct. In Eq. (8), the denominator should be the conditional density p(z_t | u_{t,t-1}, x_{t-1}), not the marginal p(z_t), because the posterior being decomposed conditions on u_{t,t-1} and x_{t-1}. In Eq. (10), the factor p(x_{t-1}, x_t) is dropped without justification; the line 'p(x_t, x_{t-1}, u) ∝ p(u|x_t, x_{t-1})' is not a valid conditioning step unless additional assumptions are stated. As written, the chain of equalities and proportionalities does not establish Eq. (6). The authors should provide a correct Bayesian derivation or explicitly present Eq. (6) as a modeling assumption rather than a theorem.
  2. [Eq. (9) and the shared-image architecture] The factorization in Eq. (6) relies on the conditional independence p(z_t | u_{t,t-1}, x_{t-1}, x_t) = p(z_t | x_t), meaning that the APR measurement noise and the RPR control noise are independent given the true state. In VKFPos, the APR branch takes the current image I_t, while the RPR branch takes (I_{t-1}, I_t); because both branches depend on I_t, their regression residuals will in general be correlated through scene content, lighting, dynamic objects, and network biases. This is not merely a technicality: correlated process and measurement noise makes the EKF misspecified, the Kalman gain suboptimal, and the state covariance potentially overconfident. The paper provides no empirical check of this independence assumption and no measure of cross-correlation between the APR and RPR residuals. This is load-bearing for the claim that the decomposition is theoretically founded.
  3. [Training/inference mismatch and covariance calibration] The APR and RPR branches are trained to minimize their respective NLL losses in Eqs. (14)-(15), but at inference the predicted covariances Σ_z and Σ_u are used directly as the measurement and process noise covariances in the EKF update Eqs. (19)-(25). There is no check that the predicted covariances are statistically calibrated for filtering, e.g., via normalized innovation squared or consistency tests. If the RPR covariance is miscalibrated, the EKF correction can degrade rather than improve the trajectory; the paper reports only final pose errors, not filter consistency. The authors should add a quantitative assessment of covariance calibration and, ideally, an ablation that replaces the learned covariances with fixed constants to show that the learned values are what drive the improvement.
  4. [Tables 1-3 and baseline choice] The experimental section does not compare against KFNet (Zhou et al. 2020), which is cited and discussed in the Introduction as the most closely related learning-based EKF integration method for temporal camera relocalization. Because KFNet is the natural baseline for the central claim 'outperforms ... model-based integration methods', its omission leaves the comparative claim incomplete. Additionally, Tables 1 and 3 report no error bars, standard deviations, or number of runs, so it is unclear whether the reported differences, some of which are small, are statistically meaningful. Finally, Table 3's caption states 'translation(cm)' while the values are clearly in meters (e.g., 4.7m), which should be corrected.
minor comments (5)
  1. [Figure 4] The caption spells 'AtLoc+' as 'AtLot+', which should be corrected.
  2. [Eq. (16)] The equation contains an extraneous comma and a misplaced plus sign: 'f (ˆxt−1, ut,t−1), +N (0, Σut,t−1 )' should read 'f (x_{t-1}, u_{t,t-1}) + N(0, Σ_{u_{t,t-1}})'.
  3. [Tables 1 and 2] Some table entries contain repeated or concatenated numbers (e.g., '0.10m0.10m0.10m, 4.07°4.07°4.07°' and '5.27°5.27°5.27°'), which appears to be a formatting artifact and should be cleaned.
  4. [References] The reference 'Charles, A. 2018. Kalman Filtering: A Bayesian Approach' is incomplete; it lacks a venue, volume, or publisher information.
  5. [Introduction] The sentence 'The results of performance evaluations ... delivering superior accuracy' in the contribution list is immediately followed by a footnote marker and a URL; the footnote should be placed at the end of the sentence without disrupting the list.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the APR/RPR decomposition is a modeling assumption, and learned covariances are shared quantities rather than fitted to the final trajectory result.

full rationale

The paper's central claim is that the monocular positioning posterior factorizes into an APR likelihood and an RPR transition. Theorem 1 (Eq. 6) is a standard Bayesian filtering identity; the proof relies on the conditional independence assumption in Eq. (9), which is an explicit EKF assumption rather than a result derived from the data. The proportionality step in Eq. (10) drops p(xhat_{t-1}, xhat_t) without justification, so the theorem is not a rigorous derivation, but this is a correctness or misspecification concern, not circularity. The APR and RPR branches are trained with separate NLL losses (Eqs. 14-15) on their respective ground truths. The predicted covariances appear both in these losses and in the EKF (Eqs. 19, 22), but this is a legitimate shared quantity: the covariances are learned to model each branch's own error distribution, not tuned to minimize final trajectory error. The EKF is applied at inference with standard equations. Experimental comparisons are against external baselines (AtLoc, MapNet, AtLoc+, LSTM-KF, ViPR) on 7-Scenes and RobotCar, so the performance claims are not manufactured by construction. No load-bearing self-citations appear. The reviewer's noted concern about correlated APR/RPR errors due to the shared current image is a genuine misspecification risk, but it is an unvalidated assumption, not circular reasoning.

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

The paper introduces no new physical entities or free constants; it relies on standard deep learning components and a standard EKF. The main assumptions are the Gaussian and diagonal-covariance model for pose distributions, the conditional independence factorization, and the independent training of the two branches.

assumptions (4)
  • domain assumption Pose distributions are Gaussian on se(3) with diagonal covariance matrices (Eq. 3 and Eq. 5).
    This assumption is stated in the absolute and relative pose estimator sections; it is required for the NLL loss functions and for the EKF noise model, but is not validated against the true pose error distribution.
  • standard math The posterior p(xt | xt-1, u, zt) factors as p(zt | xt) p(ut | xt, xt-1) via conditional independence (Theorem 1, Eq. 9-11).
    This is the standard Bayesian filtering identity, but the proof in the paper drops p(xt, xt-1) without justification and uses an incorrect denominator; the conclusion is still standard.
  • domain assumption Process and measurement noises are additive white Gaussian with covariances equal to the predicted covariances (Eq. 16-17).
    The EKF assumes Gaussian noise; the paper maps the learned covariances directly to the noise matrices, assuming the regressors output well-calibrated uncertainties.
  • domain assumption APR and RPR branches can be trained independently to maximize the full posterior (Eq. 12-13 and Sec. Training).
    The paper separates the loss into two independent NLL terms, treating the posterior factorization as a product that can be optimized separately; this ignores possible interactions through the EKF during training.

how reviews work

0 comments
Cite this review

Pith. "Pith review of VKFPos: A Learning-Based Monocular Positioning with Variational Bayesian Extended Kalman Filter Integration." pith.science (2026). https://pith.science/paper/6C3UIK5B

@misc{pith2026250118994,
  author       = {Pith},
  title        = {Pith review of: VKFPos: A Learning-Based Monocular Positioning with Variational Bayesian Extended Kalman Filter Integration},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6C3UIK5B}},
  note         = {Machine review of arXiv:2501.18994}
}
read the original abstract

This paper addresses the challenges in learning-based monocular positioning by proposing VKFPos, a novel approach that integrates Absolute Pose Regression (APR) and Relative Pose Regression (RPR) via an Extended Kalman Filter (EKF) within a variational Bayesian inference framework. Our method shows that the essential posterior probability of the monocular positioning problem can be decomposed into APR and RPR components. This decomposition is embedded in the deep learning model by predicting covariances in both APR and RPR branches, allowing them to account for associated uncertainties. These covariances enhance the loss functions and facilitate EKF integration. Experimental evaluations on both indoor and outdoor datasets show that the single-shot APR branch achieves accuracy on par with state-of-the-art methods. Furthermore, for temporal positioning, where consecutive images allow for RPR and EKF integration, VKFPos outperforms temporal APR and model-based integration methods, achieving superior accuracy.

Figures

Figures reproduced from arXiv: 2501.18994 by the authors.

Figure 1
Figure 1. The architecture of learning-based monocular po [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The upper branch is the relative pose estimator [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. This is a visualization of how EKF does in the [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Temporal positioning trajectory of Map￾Net(Upper), AtLot+(Center), VKFPos(Lower) on Oxford RobotCar Dataset. The ground truth is shown in black lines and the red lines are the prediction, while the start represents the starting point. ments. In particular, VKFPos achie…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

45 extracted references · 36 canonical work pages

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Barshan, B.; and Durrant-Whyte, H. F. 1995. Inertial navigation systems for mobile robots. IEEE transactions on robotics and automation, 11(3): 328--342

    work page 1995

  4. [4]

    Brahmbhatt, S.; Gu, J.; Kim, K.; Hays, J.; and Kautz, J. 2018. Geometry-aware learning of maps for camera localization. In Proceedings of the IEEE conference on computer vision and pattern recognition, 2616--2625

    work page 2018

  5. [5]

    Charles, A. 2018. Kalman Filtering: A Bayesian Approach

    work page 2018

  6. [6]

    Chen, B.; Wan, J.; Shu, L.; Li, P.; Mukherjee, M.; and Yin, B. 2017. Smart factory of industry 4.0: Key technologies, application case, and challenges. Ieee Access, 6: 6505--6519

    work page 2017

  7. [7]

    Chen, S.; Li, X.; Wang, Z.; and Prisacariu, V. A. 2022. Dfnet: Enhance absolute pose regression with direct feature matching. In European Conference on Computer Vision, 1--17. Springer

    work page 2022

  8. [8]

    Clark, R.; Wang, S.; Markham, A.; Trigoni, N.; and Wen, H. 2017. Vidloc: A deep spatio-temporal model for 6-dof video-clip relocalization. In Proceedings of the IEEE conference on computer vision and pattern recognition, 6856--6864

    work page 2017

Show all 45 references
  1. [9]

    Coskun, H.; Achilles, F.; DiPietro, R.; Navab, N.; and Tombari, F. 2017. Long short-term memory kalman filters: Recurrent neural estimators for pose regularization. In Proceedings of the IEEE International Conference on Computer Vision, 5524--5532

    2017

  2. [10]

    J.; Reid, I

    Davison, A. J.; Reid, I. D.; Molton, N. D.; and Stasse, O. 2007. MonoSLAM: Real-time single camera SLAM. IEEE transactions on pattern analysis and machine intelligence, 29(6): 1052--1067

    2007

  3. [11]

    Engel, J.; Sch \"o ps, T.; and Cremers, D. 2014. LSD-SLAM: Large-scale direct monocular SLAM. In European conference on computer vision, 834--849. Springer

    2014

  4. [12]

    O.; and Maximo, M

    Fran c ani, A. O.; and Maximo, M. R. 2023. Transformer-based model for monocular visual odometry: a video understanding approach. arXiv preprint arXiv:2305.06121

    2023 arXiv

  5. [13]

    He, K.; Zhang, X.; Ren, S.; and Sun, J. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, 770--778

    2016

  6. [14]

    Jinyu, L.; Bangbang, Y.; Danpeng, C.; Nan, W.; Guofeng, Z.; and Hujun, B. 2019. Survey and evaluation of monocular visual-inertial SLAM algorithms for augmented reality. Virtual Reality & Intelligent Hardware, 1(4): 386--410

    2019

  7. [15]

    Kendall, A.; and Cipolla, R. 2016. Modelling uncertainty in deep learning for camera relocalization. In 2016 IEEE international conference on Robotics and Automation (ICRA), 4762--4769. IEEE

    2016

  8. [16]

    Kendall, A.; and Cipolla, R. 2017. Geometric loss functions for camera pose regression with deep learning. In Proceedings of the IEEE conference on computer vision and pattern recognition, 5974--5983

    2017

  9. [17]

    Kendall, A.; and Gal, Y. 2017. What uncertainties do we need in bayesian deep learning for computer vision? Advances in neural information processing systems, 30

    2017

  10. [18]

    Kendall, A.; Grimes, M.; and Cipolla, R. 2015. Posenet: A convolutional network for real-time 6-dof camera relocalization. In Proceedings of the IEEE international conference on computer vision, 2938--2946

    2015

  11. [19]

    P.; and Ba, J

    Kingma, D. P.; and Ba, J. 2014. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980

    2014 arXiv

  12. [20]

    Kreuzig, R.; Ochs, M.; and Mester, R. 2019. DistanceNet: Estimating traveled distance from monocular images using a recurrent convolutional neural network. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 0--0

    2019

  13. [21]

    LaViola, J. J. 2003. A comparison of unscented and extended Kalman filtering for estimating quaternion motion. In Proceedings of the 2003 American Control Conference, 2003., volume 3, 2435--2440. IEEE

    2003

  14. [22]

    Li, C.; Lyu, H.; Wu, H.; and Qian, J. 2023. Outdoor Simultaneous Localization and Mapping by Using Millimeter Wave Radar. In IGARSS 2023-2023 IEEE International Geoscience and Remote Sensing Symposium, 4439--4442. IEEE

    2023

  15. [23]

    Li, R.; Wang, S.; Long, Z.; and Gu, D. 2018. Undeepvo: Monocular visual odometry through unsupervised deep learning. In 2018 IEEE international conference on robotics and automation (ICRA), 7286--7291. IEEE

    2018

  16. [24]

    Liashchynskyi, P.; and Liashchynskyi, P. 2019. Grid search, random search, genetic algorithm: a big comparison for NAS. arXiv preprint arXiv:1912.06059

    2019 arXiv

  17. [25]

    Macario Barros, A.; Michel, M.; Moline, Y.; Corre, G.; and Carrel, F. 2022. A comprehensive survey of visual slam algorithms. Robotics, 11(1): 24

    2022

  18. [26]

    Maddern, W.; Pascoe, G.; Linegar, C.; and Newman, P. 2017. 1 year, 1000 km: The Oxford RobotCar dataset. Int. J. Robotics Res., 36(1): 3--15

    2017

  19. [27]

    B.; Guyonneau, R.; Fasquel, J.-B.; Lagrange, S.; and Mercier, F

    Malavazi, F. B.; Guyonneau, R.; Fasquel, J.-B.; Lagrange, S.; and Mercier, F. 2018. LiDAR-only based navigation algorithm for an autonomous agricultural robot. Computers and electronics in agriculture, 154: 71--79

    2018

  20. [28]

    Moreau, A.; Piasco, N.; Tsishkou, D.; Stanciulescu, B.; and de La Fortelle, A. 2022. Coordinet: uncertainty-aware pose regressor for reliable vehicle localization. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2229--2238

    2022

  21. [29]

    Mur-Artal, R.; Montiel, J. M. M.; and Tardos, J. D. 2015. ORB-SLAM: a versatile and accurate monocular SLAM system. IEEE transactions on robotics, 31(5): 1147--1163

    2015

  22. [30]

    Ott, F.; Feigl, T.; Loffler, C.; and Mutschler, C. 2020. ViPR: Visual-odometry-aided pose regression for 6DoF camera localization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 42--43

    2020

  23. [31]

    Z.; DeVito, Z.; Raison, M.; Tejani, A.; Chilamkurthy, S.; Steiner, B.; Fang, L.; Bai, J.; and Chintala, S

    Paszke, A.; Gross, S.; Massa, F.; Lerer, A.; Bradbury, J.; Chanan, G.; Killeen, T.; Lin, Z.; Gimelshein, N.; Antiga, L.; Desmaison, A.; K \" o pf, A.; Yang, E. Z.; DeVito, Z.; Raison, M.; Tejani, A.; Chilamkurthy, S.; Steiner, B.; Fang, L.; Bai, J.; and Chintala, S. 2019. PyTo...

    2019

  24. [32]

    Shavit, Y.; Ferens, R.; and Keller, Y. 2021. Learning multi-scene absolute pose regression with transformers. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 2733--2742

    2021

  25. [33]

    Shotton, J.; Glocker, B.; Zach, C.; Izadi, S.; Criminisi, A.; and Fitzgibbon, A. 2013. Scene coordinate regression forests for camera relocalization in RGB-D images. In Proceedings of the IEEE conference on computer vision and pattern recognition, 2930--2937

    2013

  26. [34]

    Shu, F.; Lesur, P.; Xie, Y.; Pagani, A.; and Stricker, D. 2021. SLAM in the field: An evaluation of monocular mapping and localization on challenging dynamic agricultural environment. In Proceedings of the IEEE/CVF winter conference on applications of computer vision, 1761--1771

    2021

  27. [35]

    Sola, J.; Deray, J.; and Atchuthan, D. 2018. A micro Lie theory for state estimation in robotics. arXiv preprint arXiv:1812.01537

    2018 arXiv

  28. [36]

    Tan, W.; Liu, H.; Dong, Z.; Zhang, G.; and Bao, H. 2013. Robust monocular SLAM in dynamic environments. In 2013 IEEE International Symposium on Mixed and Augmented Reality (ISMAR), 209--218. IEEE

    2013

  29. [37]

    Valada, A.; Radwan, N.; and Burgard, W. 2018. Deep auxiliary learning for visual localization and odometry. In 2018 IEEE international conference on robotics and automation (ICRA), 6939--6946. IEEE

    2018

  30. [38]

    N.; Kaiser, .; and Polosukhin, I

    Vaswani, A.; Shazeer, N.; Parmar, N.; Uszkoreit, J.; Jones, L.; Gomez, A. N.; Kaiser, .; and Polosukhin, I. 2017. Attention is all you need. Advances in neural information processing systems, 30

    2017

  31. [39]

    X.; Zhao, P.; Trigoni, N.; and Markham, A

    Wang, B.; Chen, C.; Lu, C. X.; Zhao, P.; Trigoni, N.; and Markham, A. 2020. Atloc: Attention guided camera localization. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, 10393--10401

    2020

  32. [40]

    Wang, S.; Clark, R.; Wen, H.; and Trigoni, N. 2017. Deepvo: Towards end-to-end visual odometry with deep recurrent convolutional neural networks. In 2017 IEEE international conference on robotics and automation (ICRA), 2043--2050. IEEE

    2017

  33. [41]

    W.-Y.; Chang, H.-Y.; and Liang, J.-C

    Wu, H.-K.; Lee, S. W.-Y.; Chang, H.-Y.; and Liang, J.-C. 2013. Current status, opportunities and challenges of augmented reality in education. Computers & education, 62: 41--49

    2013

  34. [42]

    Xue, F.; Wang, X.; Yan, Z.; Wang, Q.; Wang, J.; and Zha, H. 2019. Local supports global: Deep camera relocalization with sequence enhancement. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 2841--2850

    2019

  35. [43]

    Yang, N.; Wang, R.; Gao, X.; and Cremers, D. 2018. Challenges in monocular visual odometry: Photometric calibration, motion bias, and rolling shutter effect. IEEE Robotics and Automation Letters, 3(4): 2878--2885

    2018

  36. [44]

    Zhou, B.; He, Y.; Huang, W.; Yu, X.; Fang, F.; and Li, X. 2022. Place recognition and navigation of outdoor mobile robots based on random Forest learning with a 3D LiDAR. Journal of Intelligent & Robotic Systems, 104(4): 72

    2022

  37. [45]

    Zhou, L.; Luo, Z.; Shen, T.; Zhang, J.; Zhen, M.; Yao, Y.; Fang, T.; and Quan, L. 2020. Kfnet: Learning temporal camera relocalization using kalman filtering. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 4919--4928

    2020

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.