Pith. sign in

REVIEW 3 major objections 5 minor 50 references

Camera-Based Localization and Enhanced Normalized Mutual Information

T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read For camera-based vehicle localization, the paper shows that matching should weight image tiles by the inverse of their noise power, and that the same noise-aware idea improves normalized mutual information matching.

desk verdict GIP2D is a clean ML result; ENMI rests on an independence assumption that contradicts the paper's own model and is not tested on real data. read the letter →

arxiv 2412.16137 v1 pith:CH4MGAX4 submitted 2024-12-20 cs.CV eess.SPstat.AP

Vishnu Teja Kunde , Jean-Francois Chamberland , Siddharth Agarwal This is my paper · ORCID
classification cs.CVeess.SPstat.AP
keywords camera-basedlocalizationnormalizedmutualinformationperspectivetransformationnoise-awarematchingweightedinnerproductautonomousdrivingmaximumlikelihooddetectionimage
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

Camera-based vehicle localization normally matches a captured image against sections of a global map using the standard Euclidean distance or normalized mutual information (NMI). The paper argues that both baselines are suboptimal because the camera's perspective transformation makes distant road tiles noisier than near ones, and neither method accounts for this uneven noise. Under a Gaussian noise model, the paper derives the maximum-likelihood match as a weighted inner product in which each tile is weighted by the inverse of its total noise variance. It then introduces enhanced NMI (ENMI) versions that spread probability mass in the joint distribution according to per-tile noise uncertainty. Numerical simulations show uniform reductions in misclassification probability for the weighted and ENMI schemes across noise levels and road-texture correlations.

What carries the argument

The central mechanism is the Jacobian of the pinhole-camera perspective projection, $\det(J) = \frac{f^2 h}{(\bar{y} \cos\theta + h\sin\theta)^3}$, which determines how a road tile of fixed physical area maps to a focal-plane area $\tilde{A}_{k,j}$ that shrinks with distance. Because sensor noise variance within a tile is $N_0/\tilde{A}_{k,j}$, distant tiles carry more noise power; this per-tile variance enters the generalized inner product weight $G_{k+(j-1)N_w, k+(j-1)N_w} = \frac{1}{2\sigma_i^2 + N_0/\tilde{A}_{k,j}}$ and controls how much probability mass ENMI spreads around each observed value.

What would settle it

Run the matching comparison on real road imagery with per-tile noise variances estimated from data, then repeat with deliberately wrong variance estimates; if the weighted GIP2D and ENMI2D schemes do not beat Euclidean distance and standard NMI when the noise model is misspecified, the claimed uniform gains would not hold in practice.

Watch

Extended reading notes

Core claim

The paper's central claim is that matching performance in camera-based localization improves when the algorithm uses the noise structure induced by the camera's perspective geometry. For the maximum-likelihood criterion, the optimal location estimate is $\hat{\ell} = \arg\min_{\ell} \sum_{k,j} \frac{(y^\ell_{k,j} - y_{k,j})^2}{2\sigma_i^2 + N_0/\tilde{A}_{k,j}}$, where $\sigma_i^2$ is the variance of intrinsic road-surface noise and $N_0/\tilde{A}_{k,j}$ is the sensor noise variance of tile $(k,j)$. This is equivalent to minimizing a generalized inner product with a diagonal weight matrix. The same principle motivates the enhanced normalized mutual information algorithms, ENMI1D and ENMI2D, which assign probability mass to value pairs according to the posterior distribution of the underlying signal rather than placing full mass on observed values. The paper maintains, and its simulations support, that these noise-aware schemes uniformly outperform their naive counterparts.

Load-bearing premise

The derivation assumes that the algorithm knows the per-tile noise variances $\sigma_i^2$ and $N_0/\tilde{A}_{k,j}$ exactly and that both the captured image and the map section are independent Gaussian noisy observations of the same underlying road intensity; if these variances are misspecified or the noise is not Gaussian, the weights and ENMI spreads become miscalibrated.

Editorial extensions

If this is right

  • If the Gaussian noise model holds, any optimal camera-based localization should use the generalized inner product weights rather than the Euclidean norm, since the former is the maximum-likelihood criterion.
  • ENMI2D and ENMI1D uniformly outperform standard NMI in the paper's simulations, across a wide range of sensor noise levels and road-texture correlation coefficients.
  • The improvements are software-only: they require no hardware changes and can be deployed as updates to existing localization pipelines.
  • The paper's approach extends in principle to multi-camera systems, LiDAR point clouds, and video sequences, though these extensions are not demonstrated.

Reading between the lines

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

  • A practical deployment would need to estimate per-tile noise variances from data; the paper's simulations assume they are known, so real-world gains will depend on the quality of that estimation.
  • Because the ENMI construction assumes $A$ and $A_\ell$ are conditionally independent given the observations, and that assumption is false when both images show the same road, the posterior spreading may overcount shared intrinsic noise; a variant that models the dependence could improve the estimate further.
  • The weighting formula gives a principled cost function for choosing camera mount height and depression angle, since both enter the signal-to-sensor-noise ratio through $\tilde{A}_{k,j}$; one could optimize mounting geometry to minimize expected localization error.
  • The same noise-aware probability-mass spreading could be applied to other NMI uses, such as medical image registration, wherever scan noise varies spatially.
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 / 5 minor

Summary. The manuscript studies camera-based localization by matching a noisy perspective-transformed camera image against sections of a global map. It derives a maximum-likelihood location estimator under a Gaussian tile-noise model, obtaining a generalized inner product (GIP2D) with per-tile weights determined by the noise powers induced by the perspective geometry (Eq. 20). It then proposes two enhanced normalized mutual information criteria, ENMI1D and ENMI2D, in which the empirical joint distribution is built by spreading probability mass according to per-tile Gaussian posteriors over the underlying signal values (Eqs. 26-28). The paper reports simulations over noise level and spatial correlation showing that GIP2D and ENMI2D uniformly outperform their unweighted counterparts.

Significance. The GIP2D maximum-likelihood derivation is a genuine, non-circular contribution: under the stated Gaussian model, the weighting of each tile by the inverse of its total noise variance is the correct likelihood principle, and the AR-1 robustness check is a useful first step beyond i.i.d. tiles. The ENMI construction, however, rests on a modeling assumption that is internally inconsistent with the paper's own generative model, and the numerical validation is generated from exactly the model whose parameters the algorithms require. If the ENMI independence issue is resolved or the method is explicitly repositioned as a heuristic with appropriate validation, the paper could be a useful signal-processing contribution to image matching; in its current form, the ENMI performance claims are not yet supported.

major comments (3)
  1. [Section IV-B, Eqs. (26)-(28)] ENMI2D constructs the joint distribution over underlying signal values as the product of per-tile marginals p(a_kj | y_kj) * p(a_l_kj | y_l_kj). But Section IV states that if the true location is l, then A = A_l, so under the true-match hypothesis the two underlying signals are the same random variable rather than independent draws. The true joint posterior is concentrated near the diagonal a = a_l and has reduced variance, not the product of two independent Gaussians. Consequently H[A, A_l] in Eq. (28) is not the entropy under the paper's own generative model. Since the candidate sections in Section V-B are generated independently, the product model is exactly correct only for false candidates, so the reported uniform gain of ENMI2D over NMI in Fig. 10 may reflect this modeling inconsistency rather than a genuine benefit of noise-aware mass spreading. The authors should either derive the correct joint posterior for the true-match hypothesis and evaluate the resulting criterion, or explicitly present ENMI as a heuristic and validate it under misspecification and on data not drawn from the product model.
  2. [Section V, Figs. 8-11] The entire empirical evaluation is limited to L = 2 synthetic tile images generated from exactly the Gaussian model of Section III, with the algorithm supplied the true noise variances sigma_i^2 and N0 / A_tilde_kj. The claim that GIP2D and ENMI2D 'uniformly outperform' alternatives is therefore not yet evidence for the stated practical setting of camera-based localization in noisy environments, where these variances are unknown and the model is approximate. The authors should add experiments with estimated or misspecified variances, mismatched noise distributions, realistic image data, and more than two candidates; otherwise the practical claims in the abstract and conclusion should be substantially tempered.
  3. [Abstract and Section IV-B] The abstract claims the proposed methods are 'provably better,' but no theorem is proved for ENMI; Proposition 1 and Corollaries 2-3 establish optimality only for the GIP detector, and the ENMI criterion is introduced heuristically. This overstatement should be removed or replaced with a precise statement of what is proved.
minor comments (5)
  1. [Throughout] There are several typographical errors: 'severly' in Section III, 'assesss' in Section V, 'pinole' in Fig. 1, 'the the horizontal' in Section II, and 'V alue' in Section III. These should be corrected in revision.
  2. [Remark 4] Remark 4 states that after adding a constant to an image, 'the joint entropy stays large'; since the empirical distribution remains concentrated on a shifted diagonal, the intended statement is presumably that the joint entropy stays small. Please correct this.
  3. [Equations (14)-(17), (22)-(24)] The notation for tile areas is inconsistent: A_tilde_kj is used for the focal-plane area, but the tilde is sometimes dropped. Please define A_kj versus A_tilde_kj once and use it consistently.
  4. [Table I] The table entries 'SINR 3 dB (IP), 10 dB (MI)' and 'Signal-to-sensor noise ratio sigma^2/N0 10-80, 45 dB (AR1)' are ambiguous because sigma^2 and the per-figure noise settings are not fully defined. Please clarify which parameter values apply to each figure.
  5. [Section IV-B] The computational cost of ENMI is described only as 'more demanding' than NMI. A brief complexity comparison (per pair of images) would help readers judge the practical deployability of the method.

Circularity Check

1 steps flagged · score 4.0 of 10

The GIP maximum-likelihood derivation is genuine, but ENMI is built from an independence ansatz that contradicts the paper's own model, so the reported ENMI gain is partly a self-consistency check of the construction rather than a derived prediction.

  1. other [Section IV-B, Eqs. (26)-(28) and Fig. 6; model assumption stated in Section IV]
    "We assume that, if the true location of the captured image is ℓ, then A = Aℓ. ... Since the global map and the captured images are noisy, the signals ... are distributed as ak,j ∼ N(yk,j, (σi k,j)^2 + (σs k,j)^2) ... aℓ k,j ∼ N(yℓ k,j, (σi k,j)^2) ... ENMI2D[Y, Yℓ] = H[A] + H[Aℓ] / H[A, Aℓ]. ... ˆp(x1, x2) += pa1,3(x1) · paℓ1,3(x2)."

    For the true candidate, the paper's model makes A and Aℓ the same random variable, so the joint posterior is concentrated on the diagonal and is not the product of independent per-tile marginals. The ENMI joint histogram, however, is constructed as a product of independent marginal posteriors, so Eq. (28) is by construction the entropy ratio of that product ansatz rather than a quantity derived from the stated generative model. The reported uniform gain of ENMI2D over NMI in Fig. 10 therefore tests the product-form construction under simulations generated from the same model, with exact noise variances supplied to the algorithm; it does not validate ENMI as a noise-aware posterior criterion derived from first principles.

full rationale

The GIP result is a genuine derivation: Proposition 1 computes the maximum-likelihood location estimate from the Gaussian model, and the weights (2σ_i^2 + N0/Ã_{k,j})^{-1} follow from the likelihood calculation rather than from any fitted parameter. The AR-1 experiment partially relaxes the i.i.d. assumption, adding some independent content. The self-citation in [21] is not load-bearing for any derivation. The circularity concern is confined to the ENMI contribution: Eqs. (26)-(28) define the joint distribution over A and Aℓ as a product of independent marginals, while the paper itself states that A = Aℓ under the true location. Consequently, ENMI is not the posterior under the paper's own model; its performance in Fig. 10 is an internal consistency check of the product-form ansatz, not a demonstration that the enhancement follows from the noise model. The GIP2D maximum-likelihood result stands, but the central ENMI claim is partially circular-by-construction, giving an overall score of 4.

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

No constants are fitted to data; all simulation parameters are chosen values for a synthetic study, and the algorithm inputs (sigma_i^2, N0 / A_tilde) are assumed known rather than estimated. The main axioms are the planar-road pinhole model and the independent Gaussian noise model, plus the ENMI-specific independence assumption that is questionable under the true-match hypothesis.

assumptions (5)
  • domain assumption Road surface is planar and camera obeys pinhole geometry with known height and depression angle.
    Used in Section II-A to derive the Jacobian and tile area formula (13)/(14).
  • domain assumption Road intensity is piecewise constant on square tiles, with independent signal values across tiles; intrinsic and sensor noises are zero-mean Gaussian and independent.
    Used in Section III to define SINR and SSNR and in Proposition 1 to derive the weighted ML detector.
  • domain assumption Under the true location, the captured image and the corresponding map section share the same underlying signal A; map noise and sensor noise are additive Gaussian with known variances.
    Central to the likelihood in Section IV-A and to ENMI in Eqs (26)-(27).
  • standard math A priori vehicle location is uniform over L candidate sections.
    Used in Proposition 1 to reduce MAP to ML.
  • ad hoc to paper For ENMI, the posterior distributions of A and A_l for a tile are multiplied as if independent, even under the true-match hypothesis.
    Section IV-B and Fig. 6: the joint empirical distribution is built from the product of per-image marginals, ignoring the common underlying road value.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Camera-Based Localization and Enhanced Normalized Mutual Information." pith.science (2026). https://pith.science/paper/CH4MGAX4

@misc{pith2026241216137,
  author       = {Pith},
  title        = {Pith review of: Camera-Based Localization and Enhanced Normalized Mutual Information},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CH4MGAX4}},
  note         = {Machine review of arXiv:2412.16137}
}
read the original abstract

Robust and fine localization algorithms are crucial for autonomous driving. For the production of such vehicles as a commodity, affordable sensing solutions and reliable localization algorithms must be designed. This work considers scenarios where the sensor data comes from images captured by an inexpensive camera mounted on the vehicle and where the vehicle contains a fine global map. Such localization algorithms typically involve finding the section in the global map that best matches the captured image. In harsh environments, both the global map and the captured image can be noisy. Because of physical constraints on camera placement, the image captured by the camera can be viewed as a noisy perspective transformed version of the road in the global map. Thus, an optimal algorithm should take into account the unequal noise power in various regions of the captured image, and the intrinsic uncertainty in the global map due to environmental variations. This article briefly reviews two matching methods: (i) standard inner product (SIP) and (ii) normalized mutual information (NMI). It then proposes novel and principled modifications to improve the performance of these algorithms significantly in noisy environments. These enhancements are inspired by the physical constraints associated with autonomous vehicles. They are grounded in statistical signal processing and, in some context, are provably better. Numerical simulations demonstrate the effectiveness of such modifications.

Figures

Figures reproduced from arXiv: 2412.16137 by the authors.

Figure 1
Figure 1. This notional diagram shows the (standard) alignmen [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The above figure illustrates the coordinate system of [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. The fundamental triangles that govern the projectio [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: The left diagram shows an arbitrary grid of equally si [PITH_FULL_IMAGE:figures/full_fig_p003_4.png]
Figure 5
Figure 5. Figure 5: This diagram shows how the probability mass is added t [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: The diagram shows how the probability mass is added to [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: The diagram shows how the probability mass is added to [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 8
Figure 8. Figure 8: This plot compares the performance of the three inner [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 9
Figure 9. Figure 9: Roads surfaces often have features that are spatiall [PITH_FULL_IMAGE:figures/full_fig_p008_9.png]
Figure 11
Figure 11. Figure 11: This graph tracks the performance of the three mutua [PITH_FULL_IMAGE:figures/full_fig_p009_11.png]
Figure 10
Figure 10. Figure 10: The two enhance NMI schemes take advantage of the une [PITH_FULL_IMAGE:figures/full_fig_p009_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 48 canonical work pages

  1. [1]

    Heuristi c approaches in robot path planning: A survey,

    T. T. Mac, C. Copot, D. T. Tran, and R. De Keyser, “Heuristi c approaches in robot path planning: A survey,” Robotics and Autonomous Systems , vol. 86, pp. 13–28, 2016

    work page 2016

  2. [2]

    A survey of the state-of-the-art localizat ion techniques and their potentials for autonomous vehicle applications,

    S. Kuutti, S. Fallah, K. Katsaros, M. Dianati, F. Mccullo ugh, and A. Mouzakitis, “A survey of the state-of-the-art localizat ion techniques and their potentials for autonomous vehicle applications, ” Internet of Things Journal, vol. 5, no. 2, pp. 829–846, 2018

    work page 2018

  3. [3]

    A s urvey of autonomous driving: Common practices and emerging technol ogies,

    E. Y urtsever, J. Lambert, A. Carballo, and K. Takeda, “A s urvey of autonomous driving: Common practices and emerging technol ogies,” IEEE Access , vol. 8, pp. 58 443–58 469, 2020

    work page 2020

  4. [4]

    A s urvey of deep learning applications to autonomous vehicle control,

    S. Kuutti, R. Bowden, Y . Jin, P . Barber, and S. Fallah, “A s urvey of deep learning applications to autonomous vehicle control,” Transactions on Intelligent Transportation Systems , vol. 22, no. 2, pp. 712–733, 2020

    work page 2020

  5. [5]

    4D automotive radar sensing for au tonomous vehicles: A sparsity-oriented approach,

    S. Sun and Y . D. Zhang, “4D automotive radar sensing for au tonomous vehicles: A sparsity-oriented approach,” IEEE Journal of Selected Topics in Signal Processing , vol. 15, no. 4, pp. 879–891, 2021

    2021

  6. [6]

    Autonomous vehicles on the edge: A survey on autonomous vehicle racing,

    J. Betz, H. Zheng, A. Liniger, U. Rosolia, P . Karle, M. Beh l, V . Krovi, and R. Mangharam, “Autonomous vehicles on the edge: A survey on autonomous vehicle racing,” Open Journal of Intelligent Transportation Systems, vol. 3, pp. 458–488, 2022

    work page 2022

  7. [7]

    Mapping the world in 3D,

    B. Schwarz, “Mapping the world in 3D,” Nature Photonics, vol. 4, no. 7, pp. 429–430, 2010

    work page 2010

  8. [8]

    Developments in moder n GNSS and its impact on autonomous vehicle architectures,

    N. Joubert, T. Reid, and F. Noble, “Developments in moder n GNSS and its impact on autonomous vehicle architectures,” in Intelligent V ehicles Symposium (IV) . IEEE, 2020, pp. 2029–2036

    work page 2020

Show all 50 references
  1. [9]

    Deep-urba n unaided precise global navigation satellite system vehicle positi oning,

    T. E. Humphreys, M. J. Murrian, and L. Narula, “Deep-urba n unaided precise global navigation satellite system vehicle positi oning,” Intelligent Transportation Systems Magazine , vol. 12, no. 3, pp. 109–122, 2020

    2020

  2. [11]

    In creased accuracy for fast moving LiDARS: Correction of distorted point cloud s,

    T. Renzler, M. Stolz, M. Schratter, and D. Watzenig, “In creased accuracy for fast moving LiDARS: Correction of distorted point cloud s,” in International Instrumentation and Measurement Technolog y Conference (I2MTC). IEEE, 2020

    2020

  3. [12]

    Lidar for autonomous drivi ng: The principles, challenges, and trends for automotive lidar an d perception systems,

    Y . Li and J. Ibanez-Guzman, “Lidar for autonomous drivi ng: The principles, challenges, and trends for automotive lidar an d perception systems,” IEEE Signal Processing Magazine , vol. 37, no. 4, pp. 50–61, 2020

    2020

  4. [13]

    MIMO radar for ad vanced driver-assistance systems and autonomous driving: Advant ages and challenges,

    S. Sun, A. P . Petropulu, and H. V . Poor, “MIMO radar for ad vanced driver-assistance systems and autonomous driving: Advant ages and challenges,” IEEE Signal Processing Magazine , vol. 37, no. 4, pp. 98– 117, 2020

    2020

  5. [14]

    Emergent visua l sensors for autonomous vehicles,

    Y . Li, J. Moreau, and J. Ibanez-Guzman, “Emergent visua l sensors for autonomous vehicles,” IEEE Transactions on Intelligent Transportation Systems, vol. 24, no. 5, pp. 4716–4737, 2023

    2023

  6. [15]

    Planning an d decision- making for autonomous vehicles,

    W. Schwarting, J. Alonso-Mora, and D. Rus, “Planning an d decision- making for autonomous vehicles,” Annual Review of Control, Robotics, and Autonomous Systems , vol. 1, pp. 187–210, 2018

    2018

  7. [16]

    Computer vi sion for au- tonomous vehicles: Problems, datasets and state of the art,

    J. Janai, F. G¨ uney, A. Behl, and A. Geiger, “Computer vi sion for au- tonomous vehicles: Problems, datasets and state of the art, ” F oundations and Trends® in Computer Graphics and Vision , vol. 12, no. 1–3, pp. 1–308, 2020

    2020

  8. [17]

    City- scale localiza- tion for cameras with known vertical direction,

    L. Sv¨ arm, O. Enqvist, F. Kahl, and M. Oskarsson, “City- scale localiza- tion for cameras with known vertical direction,” IEEE Trans. on Pattern Analysis and Machine Intelligence , vol. 39, no. 7, pp. 1455–1461, 2016

    2016

  9. [18]

    Simultaneous localiz ation and map- ping: Part I,

    H. Durrant-Whyte and T. Bailey, “Simultaneous localiz ation and map- ping: Part I,” Robotics & Automation Magazine , vol. 13, no. 2, pp. 99–110, 2006

    2006

  10. [19]

    Simultaneous localiz ation and map- ping (SLAM): Part II,

    T. Bailey and H. Durrant-Whyte, “Simultaneous localiz ation and map- ping (SLAM): Part II,” Robotics & Automation Magazine , vol. 13, no. 3, pp. 108–117, 2006

    2006

  11. [20]

    WiFi-SLAM using Gau ssian process latent variable models,

    B. Ferris, D. Fox, and N. Lawrence, “WiFi-SLAM using Gau ssian process latent variable models,” in International Joint Conferences on Artificial Intelligence (IJCAI) , 2007

    2007

  12. [21]

    Localization requirements for autonomous v ehicles,

    T. G. Reid, S. E. Houts, R. Cammarata, G. Mills, S. Agarwa l, A. V ora, and G. Pandey, “Localization requirements for autonomous v ehicles,” arXiv preprint arXiv:1906.01061 , 2019

    1906 arXiv

  13. [22]

    A solution to the simultaneous locali zation and map building (SLAM) problem,

    M. W. M. G. Dissanayake, P . Newman, S. Clark, H. F. Durran t- Whyte, and M. Csorba, “A solution to the simultaneous locali zation and map building (SLAM) problem,” IEEE Transactions on Robotics and Automation , vol. 17, no. 3, pp. 229–241, 2001

    2001

  14. [23]

    DP-SLAM: Fast, robust simultan eous localiza- tion and mapping without predetermined landmarks,

    A. Eliazar and R. Parr, “DP-SLAM: Fast, robust simultan eous localiza- tion and mapping without predetermined landmarks,” in International Joint Conferences on Artificial Intelligence , vol. 3. Acapulco, Mexico, 2003, pp. 1135–1142

    2003

  15. [24]

    Simulta neous lo- calization and map building in large-scale cyclic environm ents using the Atlas framework,

    M. Bosse, P . Newman, J. Leonard, and S. Teller, “Simulta neous lo- calization and map building in large-scale cyclic environm ents using the Atlas framework,” The International Journal of Robotics Research , vol. 23, no. 12, pp. 1113–1139, 2004

    2004

  16. [25]

    The graph SLAM algorithm wi th appli- cations to large-scale mapping of urban structures,

    S. Thrun and M. Montemerlo, “The graph SLAM algorithm wi th appli- cations to large-scale mapping of urban structures,” The International Journal of Robotics Research , vol. 25, no. 5-6, pp. 403–429, 2006

    2006

  17. [26]

    Parallel tracking and mapping f or small AR workspaces,

    G. Klein and D. Murray, “Parallel tracking and mapping f or small AR workspaces,” in International Symposium on Mixed and Augmented Reality. IEEE, 2007, pp. 1–10

    2007

  18. [27]

    LSD-SLAM: Large- scale di- rect monocular SLAM,

    J. Engel, T. Sch¨ ops, and D. Cremers, “LSD-SLAM: Large- scale di- rect monocular SLAM,” in European Conference on Computer Vision . Springer, 2014, pp. 834–849

    2014

  19. [28]

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

    R. Mur-Artal, J. M. M. Montiel, and J. D. Tard´ os, “ORB-S LAM: A versatile and accurate monocular SLAM system,” IEEE Transactions on Robotics , vol. 31, no. 5, pp. 1147–1163, 2015

    2015

  20. [29]

    Robust vehicle localization in urban en- vironments using probabilistic maps,

    J. Levinson and S. Thrun, “Robust vehicle localization in urban en- vironments using probabilistic maps,” in International Conference on Robotics and Automation . IEEE, 2010, pp. 4372–4378

    2010

  21. [30]

    Towards f ully autonomous driving: Systems and algorithms,

    J. Levinson, J. Askeland, J. Becker, J. Dolson, D. Held, S. Kammel, J. Z. Kolter, D. Langer, O. Pink, V . Pratt, V . Pratt, M. Sokolsky, G . Stanek, D. Stavens, A. Teichman, M. Werling, and S. Thrun, “Towards f ully autonomous driving: Systems and algorithms,” in Intelligent V...

    2011

  22. [31]

    Simultane ous localization and mapping: A survey of current trends in autonomous drivin g,

    G. Bresson, Z. Alsayed, L. Y u, and S. Glaser, “Simultane ous localization and mapping: A survey of current trends in autonomous drivin g,” IEEE Trans./ on Intelligent V ehicles, vol. 2, no. 3, pp. 194–220, 2017

    2017

  23. [32]

    Image registration using mul tiresolution frequency domain correlation,

    S. Kr¨ uger and A. Calway, “Image registration using mul tiresolution frequency domain correlation,” in British Machine Vision Conference . BMV A Press, 1998, pp. 1–10

    1998

  24. [33]

    Markov localization using cor relation,

    K. Konolige and K. Chou, “Markov localization using cor relation,” in International Joint Conferences on Artificial Intelligenc e Organization , vol. 99, 1999, pp. 1154–1159

    1999

  25. [34]

    Toward automated driving in cities using close-to-market sensors: An overview of the v-charge proje ct,

    P . Furgale, U. Schwesinger, M. Rufli, W. Derendarz, H. Gr immett, P . M¨ uhlfellner, S. Wonneberger, J. Timpner, S. Rottmann, B . Li, B. Schmidt, T. N. Nguyen, E. Cardarelli, S. Cattani, S. Br¨ un ing, S. Horstmann, M. Stellmacher, H. Mielenz, K. K¨ oser, M. Beer mann, C. H¨ a...

    2013

  26. [35]

    Project autovision: Localization and 3d scene perception for an autonomous vehicle with a multi-camera system,

    L. Heng, B. Choi, Z. Cui, M. Geppert, S. Hu, B. Kuan, P . Liu , R. Nguyen, Y . C. Y eo, A. Geiger, G. H. Lee, M. Pollefeys, and T. Sattler, “Project autovision: Localization and 3d scene perception for an autonomous vehicle with a multi-camera system,” in IEEE International Co...

    2019

  27. [36]

    Detection of signals by informat ion theoretic criteria,

    M. Wax and T. Kailath, “Detection of signals by informat ion theoretic criteria,” IEEE Transactions on acoustics, speech, and signal process ing, vol. 33, no. 2, pp. 387–392, 1985

    1985

  28. [37]

    Parameter estimation of sup erimposed signals using the em algorithm,

    M. Feder and E. Weinstein, “Parameter estimation of sup erimposed signals using the em algorithm,” IEEE Transactions on acoustics, speech, and signal processing , vol. 36, no. 4, pp. 477–489, 1988

    1988

  29. [38]

    Model based processing of signals: A state space approach,

    B. Rao and K. Arun, “Model based processing of signals: A state space approach,” Proceedings of the IEEE , vol. 80, no. 2, pp. 283–309, 1992

    1992

  30. [39]

    Generalized homomorp hic and adaptive order statistic filters for the removal of impulsiv e and signal- dependent noise,

    R. Ding and A. V enetsanopoulos, “Generalized homomorp hic and adaptive order statistic filters for the removal of impulsiv e and signal- dependent noise,” IEEE transactions on circuits and systems , vol. 34, no. 8, pp. 948–955, 1987

    1987

  31. [40]

    Detection algorithms for hyp erspectral imaging applications,

    D. Manolakis and G. Shaw, “Detection algorithms for hyp erspectral imaging applications,” Signal processing magazine , vol. 19, no. 1, pp. 29–43, 2002

    2002

  32. [41]

    Distributed and decentralize d multicamera tracking,

    M. Taj and A. Cavallaro, “Distributed and decentralize d multicamera tracking,” IEEE Signal Processing Magazine , vol. 28, no. 3, pp. 46–58, 2011

    2011

  33. [42]

    Learning sparsifying t ransforms,

    S. Ravishankar and Y . Bresler, “Learning sparsifying t ransforms,” IEEE Transactions on Signal Processing , vol. 61, no. 5, pp. 1072–1086, 2012

    2012

  34. [43]

    Suppression of acoustic noise in speech using spectral subtrac- tion,

    S. Boll, “Suppression of acoustic noise in speech using spectral subtrac- tion,” IEEE Transactions on acoustics, speech, and signal process ing, vol. 27, no. 2, pp. 113–120, 1979

    1979

  35. [44]

    Speech enhancement based on a priori signal to noise estimation,

    P . Scalart et al. , “Speech enhancement based on a priori signal to noise estimation,” in 1996 IEEE International Conference on Acoustics, 11 Speech, and Signal Processing Conference Proceedings , vol. 2. IEEE, 1996, pp. 629–632

    1996

  36. [45]

    A multi-band spectral subtrac tion method for enhancing speech corrupted by colored noise

    S. Kamath and P . Loizou, “A multi-band spectral subtrac tion method for enhancing speech corrupted by colored noise.” in International Conference on Acoustics, Speech and Signal Processing (ICA SSP), vol. 4. IEEE, 2002, p. 4164

    2002

  37. [46]

    A new approach to identify the na ture of the noise affecting an image,

    K. Chehdi and M. Sabri, “A new approach to identify the na ture of the noise affecting an image,” in International Conference on Acoustics, Speech and Signal Processing (ICASSP) , vol. 3. IEEE, 1992, pp. 285– 288

    1992

  38. [47]

    Joint demosaicing and deno ising,

    K. Hirakawa and T. W. Parks, “Joint demosaicing and deno ising,” IEEE Transactions on Image Processing , vol. 15, no. 8, pp. 2146–2157, 2006

    2006

  39. [48]

    Homogeneity localization using particle filters with application to noise estimation,

    M. Ghazal and A. Amer, “Homogeneity localization using particle filters with application to noise estimation,” IEEE Transactions on image processing, vol. 20, no. 7, pp. 1788–1796, 2010

    2010

  40. [49]

    Image deno ising using the higher order singular value decomposition,

    A. Rajwade, A. Rangarajan, and A. Banerjee, “Image deno ising using the higher order singular value decomposition,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 35, no. 4, pp. 849–862, 2012

    2012

  41. [50]

    Imag e registration by maximization of combined mutual information and gradien t informa- tion,

    J. P . W. Pluim, J. B. A. Maintz, and M. A. Viergever, “Imag e registration by maximization of combined mutual information and gradien t informa- tion,” in Medical Image Computing and Computer-Assisted Interventi on (MICCAI), S. L. Delp, A. M. DiGoia, and B. Jaramaz, Eds. Spri...

    2000

  42. [51]

    Medical imag e registration using mutual information,

    F. Maes, D. V andermeulen, and P . Suetens, “Medical imag e registration using mutual information,” Proceedings of the IEEE , vol. 91, no. 10, pp. 1699–1722, 2003

    2003

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.