SCREP: Scene Coordinate Regression and Evidential Learning-based Perception-Aware Trajectory Generation

Guilherme V. Cavalheiro; Juyeop Han; Lukas Lao Beyer; Sertac Karaman

REVIEW 1 major objections 4 minor 41 references

SCREP: Scene Coordinate Regression and Evidential Learning-based Perception-Aware Trajectory Generation

T0 review · 1 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Steering a camera at low-uncertainty scene coordinates improves indoor drone localization, cutting average rotation RMSE by 30.8 to 41.3 percent.

desk verdict A credible integration of evidential SCR uncertainty with yaw planning; average gains hold up, but the causal mechanism rests on an untested spatial-correlation assumption and some consistency claims outrun the data. read the letter →

arxiv 2507.07467 v3 pith:PLL4NSYW submitted 2025-07-10 cs.RO

Juyeop Han , Lukas Lao Beyer , Guilherme V. Cavalheiro , Sertac Karaman This is my paper · ORCID

classification cs.RO

keywords scenecoordinateregressionevidentiallearningperception-awareplanningtrajectoryoptimizationuncertaintyquantificationvisuallocalizationUAVnavigationreceding-horizon

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 argues that a small drone can localize itself more accurately indoors—without GPS—if it actively points its camera at parts of the scene whose 3D coordinates its neural network is most confident about, rather than at visually rich features or straight ahead. It couples a scene-coordinate-regression network trained with evidential learning, which outputs per-pixel uncertainty as a single closed-form entropy value, with a receding-horizon trajectory optimizer whose field-of-view cost rewards keeping low-entropy scene coordinates in view. The claimed payoff is concrete: in photorealistic simulation the method cuts average translation RMSE by 4.9–23.5% and rotation RMSE by 30.8–41.3% relative to three baselines, and hardware-in-the-loop tests confirm the pipeline runs in real time. If right, entropy-aware gaze control makes learning-based absolute localization accurate enough for onboard autonomous indoor flight without expensive feature-map reconstruction.

What carries the argument

The load-bearing object is E-SCRNet, a scene-coordinate-regression network trained with deep evidential regression, which outputs—for each pixel—the Normal-Inverse-Gamma hyperparameters ($\gamma, \lambda, \alpha, \beta$) of a distribution over the predicted 3D coordinate. From these, the network's predictive distribution is a Student-t with location $\gamma$, and the predictive entropy $H_\theta = \mathbb{E}_{v\sim p(v|m_\theta)}[-\log p(v|m_\theta)]$ has a closed form; this single scalar combines aleatoric and epistemic uncertainty. The entropy map enters a receding-horizon B-spline trajectory optimizer through a differentiable field-of-view indicator $F(c_i, v_j)$ and an exponential entropy weight $\exp(-a_{\mathrm{fov}} H_{\theta,j})$, so that yaw is steered toward reliable coordinates while position tracking, velocity/acceleration limits, and smoothness are enforced by separate costs. The sequential position-then-yaw optimization and the fixed-lag smoother are what make the scheme fast enough for real-time flight.

What would settle it

Measure how well the entropy seen from a future viewpoint (0.8 s ahead) matches the entropy predicted by the current map: if the two do not correlate, the planner is steering using stale information. Alternatively, feed the optimizer a deliberately wrong or permuted entropy map; if localization accuracy does not degrade, the entropy signal is not the active ingredient.

Watch

Extended reading notes

Core claim

On its own terms, this paper establishes that the uncertainty already produced by an evidential scene-coordinate-regression network can be used directly as a planning signal to improve localization. E-SCRNet predicts, for each pixel, the hyperparameters of a Normal-Inverse-Gamma distribution over the scene coordinate; marginalizing gives a Student-t predictive distribution whose predictive entropy $H_\theta$ (Eq. 3) summarizes both aleatoric and epistemic uncertainty. The trajectory optimizer maximizes a weighted field-of-view score $C_{\mathrm{fov}} = -\sum_i \sum_j \exp(-a_{\mathrm{fov}} H_{\theta,j}) F(c_i, v_j)$, so low-entropy coordinates pull the camera yaw toward them and high-entropy coordinates are filtered out. A fixed-lag smoother fuses the resulting low-rate SCR poses with high-rate IMU data. The paper's experimental claim is that this entropy-weighted steering outperforms forward-facing, uncertainty-blind, and Fisher-information-based yaw strategies, and that entropy is a better scalar uncertainty proxy than aleatoric or epistemic uncertainty alone. The paper also claims to be the first active-perception work using scene coordinate regression.

Load-bearing premise

The planner assumes that the uncertainty map seen from the current viewpoint remains a good predictor of what the camera will see over the 0.8-second planning horizon; the paper supports this only with qualitative evidence, so if spatial correlation of uncertainty is weak, the yaw optimization may steer toward views that are not actually reliable.

Editorial extensions

If this is right

If the central claim holds, uncertainty-aware camera gaze becomes a practical lever for improving absolute localization accuracy in GPS-denied flight, with the largest gains in rotation error (average RMSE down 30.8–41.3% versus baselines).
Because the entropy map is computed in a single forward pass, the approach avoids the multi-pass cost of Bayesian neural networks and can run online: SCR and PnP take 62.2 ms, smoothing 0.8 ms, and optimization 28.3 ms.
Entropy, not aleatoric or epistemic uncertainty separately, is the recommended planning signal: the paper shows that mean and standard deviation of scene-coordinate error grow with entropy but not consistently with the other two measures.
The fixed-lag smoother makes the low-rate, sometimes outlier-prone SCR pose stream usable for control, producing a drift-free high-rate estimate; this is a direct corollary of the fused-pose results.
The method is the first active-perception planner built on scene coordinate regression, so it opens SCR-based localization to perception-aware planning rather than only to passive pose estimation.

Reading between the lines

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

An immediate extension suggested, but not tested, by the paper is applying the same entropy-weighted field-of-view cost to other learning-based pose estimators, such as absolute pose regressors or NeRF-augmented localization, since the cost only needs per-pixel or per-region uncertainty.
The decoupling of yaw from velocity means the same gaze could serve dual purposes—localization and, say, obstacle or semantic mapping—at no extra trajectory cost; the authors list obstacle avoidance as future work.
The 0.8-second planning horizon is only as good as the spatial correlation of uncertainty; in scenes with sharp occlusions or moving objects, the current-view entropy map may not predict the next-view map, so the claimed gains may shrink outside static structured environments.
A conservative transfer check would be to measure realized entropy at the planned future viewpoint; if the yaw optimizer does not actually lower next-view entropy relative to the baseline, the localization improvement should disappear, which would localize the mechanism precisely.

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.

Desk Editor's Note

read the letter

Short version: this is a solid engineering paper and a genuine first: coupling an evidential SCR pose estimator with a perception-aware yaw optimizer, using predictive entropy as the planning signal. The Vanilla baseline (same optimizer, no uncertainty weighting) is the right ablation, and the simulation suite is decent: eight Blackbird trajectories across two sites, four yaw strategies, plus a block bootstrap whose confidence interval for the mean improvement over Vanilla excludes zero. The reported runtimes are plausible for real-time use. Given the subfield's habit of benchmarking only feature- or VIO-based planners, this is a useful addition and I don't see a load-bearing math error.

The soft spots are real but not fatal. The biggest is Assumption A2: the planner's cost in Eq. (11) uses entropy maps from the current pose to select viewpoints up to 0.8 s ahead, which only works if entropy is spatially correlated enough that those maps predict future views. The paper supports that with Fig. 2 and a supplementary video, both qualitative. That is not enough for a claim that low-entropy steering *causes* the localization gains; it could be correlated with the improvement. A quantitative check — autocorrelation of entropy over the horizon, or a prediction-error analysis — would close this. The stress-test note calls this the least supported load-bearing condition, and I agree.

Second, the text says the method 'consistently outperforms' baselines, but Table Ib shows 'sid' where Vanilla's RMSE (34.8 cm) beats Ours (40.5 cm). The average claim survives, the consistency claim does not. The hardware-in-the-loop experiment is also a single trajectory, simulated camera, and only against Forward, so it demonstrates integration feasibility, not planner superiority.

Minor: no code or models released, and the entropy-vs-error calibration (Fig. 3) comes from the same simulator and environments as the evaluation. That limits, but does not erase, the evidence.

Bottom line: people working on perception-aware planning and SCR-based relocalization will get value from this. It deserves a serious referee rather than a desk reject; the right path is a revision that adds the A2 evidence, reports per-trajectory statistics, and tempers the consistency language. I would encourage you to engage with it on those terms.

Referee Report

1 major / 4 minor

Summary. The paper presents SCREP, a receding-horizon perception-aware trajectory planner for indoor UAV flight. The method uses an evidential-learning scene coordinate regression network (E-SCRNet) that outputs per-pixel 3D scene coordinates along with aleatoric and epistemic uncertainties, summarized as a closed-form predictive entropy. A two-stage B-spline optimizer first plans the position trajectory to track waypoints and then optimizes yaw so that low-entropy scene coordinates remain in the camera frustum; SCR poses are fused with IMU data in a fixed-lag smoother. The approach is evaluated in FlightGoggles on two Blackbird datasets with four baselines (Forward, Vanilla, FIM, and Ours) and in one hardware-in-the-loop experiment. The authors report average translation and rotation RMSE reductions over the baselines and provide a block-bootstrap confidence interval that excludes zero for the mean improvement over the strongest baseline.

Significance. If the causal mechanism is substantiated, this is a useful contribution to perception-aware planning: the paper is, to the best of my knowledge, the first active-perception planner built on scene coordinate regression rather than feature matching or visual-inertial odometry, and the choice of a closed-form entropy cost with a real-time two-stage optimizer is well motivated. The experimental design is also partially strong: the baselines share the same SCR pipeline, isolating the uncertainty-weighting mechanism, and the mean improvement over the Vanilla baseline is supported by a block bootstrap that respects temporal correlation. The runtime figures (62.2 ms for SCR+PnP, 0.8 ms for smoothing, 28.3 ms for optimization) support the real-time claim. The main weakness is that the paper's central causal claim relies on an assumption about the spatial persistence of uncertainty that is only supported qualitatively; this gap is fixable with additional quantitative analysis.

major comments (1)

[Table I(b), 'sid' row; Section V-C text] The text states that the method 'consistently outperforms the baselines' on IMU+SCR localization, but the per-trajectory results contradict this for translation error. In the 'sid' row of Table I(b), Vanilla achieves RMSE 34.8 cm while Ours achieves 40.5 cm; in Table I(a), Vanilla's mean translation error is 22.9 cm versus Ours 30.1 cm. The average-level improvement and the bootstrap interval over Vanilla are credible, but the 'consistently outperforms' claim is overstated. Please replace this phrasing with per-trajectory win/loss counts or per-trajectory confidence intervals, and qualify the conclusion accordingly.

minor comments (4)

[Section IV-B, Eq. (11) and parameter list] The entropy-selection threshold used to choose the n_f scene coordinates in Eq. (11) is not reported in the parameter list after Eq. (5). Please state the threshold value and clarify whether exp(-a_fov H) and the threshold operate on the raw entropy from Eq. (3) or on a normalized version.
[Section V-B and Fig. 3] The claim that entropy is a more informative uncertainty measure than aleatoric or epistemic uncertainty is supported only by visual inspection of Fig. 3. Reporting a numeric correlation coefficient or mutual information between each uncertainty metric and the L2 scene-coordinate error would make this contribution claim more precise.
[Abstract and Section V-C] The abstract phrase 'reduces translation and rotation RMSE by at least 4.9% and 30.8% relative to baselines' is technically relative to the strongest baseline (Vanilla), not to all three baselines; the reductions relative to Forward and FIM are larger. Please rephrase to avoid ambiguity.
[Section V-C] The block bootstrap is reported only for the mean translation improvement over Vanilla. I recommend reporting a similar interval for the rotation improvement and for at least the 'sid' trajectory, since that trajectory is the main counterexample to the consistency claim.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the entropy-weighted FOV cost in Eq. (11) is an independently defined objective, and the reported localization RMSE reductions are not fitted reparametrizations of it.

full rationale

The paper's central chain is self-contained: E-SCRNet is trained with the deep-evidential NLL loss in Eq. (1) plus regularizer in Eq. (2), and the entropy in Eq. (3) is a closed-form function of the evidential hyperparameters, not of the evaluation RMSE. The planner minimizes C_fov in Eq. (11) with entropy weights exp(-a_fov H), while pose accuracy is measured through PnP-RANSAC and a fixed-lag smoother; no parameter of Eq. (11) is fitted to the reported RMSE. The 'Vanilla' baseline (fixed weight 1.0, no entropy filtering) is a genuine ablation that isolates the entropy mechanism. Assumptions A1 and A2 are empirical calibration/spatial-correlation claims: A1 is examined in Section V-B (Fig. 3), and A2 is explicitly supported only qualitatively (Fig. 2 and video), which is a weakness in evidence, not a circular derivation. The self-citations ([3], [31], [39], [40]) are prior-method, dataset, or related-work references and are not load-bearing; no uniqueness theorem or ansatz is imported to force the result. The claim that the method 'consistently outperforms' baselines is contradicted by the 'sid' row in Table Ib (Ours 40.5 cm vs Vanilla 34.8 cm RMSE), but that is an empirical overstatement, not circularity.

Assumptions & free parameters 8 free parameters · 7 assumptions · 0 invented entities

No new physical or conceptual entities are postulated; the entropy map is a summary of the evidential network's existing predictive distribution. The planner's behavior depends on several tuned cost and horizon parameters, and on two explicitly stated assumptions about uncertainty calibration and spatial smoothness.

free parameters (8)

cost weights (lambda_wp, lambda_fov, lambda_eq, lambda_ie, lambda_s) = (1e4, 1e1, 1e3, 1.0, 5.0)
Tuned on a single circular trajectory and transferred to all evaluation trajectories; they directly determine the behavior being compared.
planning horizons T_plan and T_exec = 0.8 s and 0.5 s
Receding-horizon lengths chosen by hand; they set how far ahead the uncertainty map is assumed to be predictive.
number of sampled scene coordinates n_f = 200
Number of low-entropy points randomly selected for each FOV cost evaluation; affects optimization stability and cost fidelity.
soft-frustum smoothing constant s = 5.0
Controls the sharpness of the differentiable visibility indicator in Eq. (9) and hence the gradients of the FOV cost.
entropy weight exponent a_fov = 0.5
Scales the entropy weight in Eq. (11); tuned, not derived from any principle.
entropy selection threshold = not reported
Points are chosen among those below a predefined entropy threshold, but the threshold value is never stated.
evidential loss regularization rho = 1e-2
Regularization coefficient in Eq. (2); affects the scale of predicted uncertainties, which the planner then uses.
B-spline degree and control point count = k=3, 6 control points
Trajectory parameterization choices that bound what shapes the optimizer can express.

assumptions (7)

domain assumption Multirotor dynamics are differentially flat (Mellinger and Kumar [37])
Used in Sections III-IV to recover states and controls from position and yaw and to split optimization into position and yaw subproblems.
standard math Normal-Inverse-Gamma is a conjugate prior for Gaussian likelihood, so the predictive distribution is Student-t
Underlies Eqs. (1)-(3), the closed-form NLL and entropy used as the perception cost.
domain assumption PnP-RANSAC converts per-pixel 3D scene-coordinate predictions into an absolute camera pose
Core of the pose estimation block; assumed to yield drift-free estimates when enough reliable correspondences survive.
domain assumption A1: lower uncertainty implies higher scene-coordinate accuracy
Stated in Section V and validated empirically on one circular trajectory per environment; without it the entropy cost has no relation to localization quality.
domain assumption A2: uncertainty is spatially correlated and varies smoothly over the trajectory
Stated as a key assumption in Section V and needed to justify receding-horizon planning; supported only qualitatively by Fig. 2 and a video.
domain assumption Camera intrinsics and body-to-camera extrinsics are known
Required to build the frustum indicator in Eqs. (7)-(9).
domain assumption An external reference trajectory is provided
The planner tracks waypoints rather than doing global path planning, so its benefits are conditional on a supplied path.

how reviews work

0 comments

Cite this review

Pith. "Pith review of SCREP: Scene Coordinate Regression and Evidential Learning-based Perception-Aware Trajectory Generation." pith.science (2026). https://pith.science/paper/PLL4NSYW

@misc{pith2026250707467,
  author       = {Pith},
  title        = {Pith review of: SCREP: Scene Coordinate Regression and Evidential Learning-based Perception-Aware Trajectory Generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PLL4NSYW}},
  note         = {Machine review of arXiv:2507.07467}
}

read the original abstract

Autonomous flight in GPS-denied indoor spaces requires trajectories that keep visual-localization error tightly bounded across varied missions. Map-based visual localization methods such as feature matching require computationally intensive map reconstruction and have feature-storage scalability issues, especially for large environments. Scene coordinate regression (SCR) provides an efficient learning-based alternative that directly predicts3D coordinates for every pixel, enabling absolute pose estimation with significant potential for onboard roboticsapplications. We present a perception-aware trajectory planner that couples an evidential learning-based SCR poseestimator with a receding-horizon trajectory optimizer. The optimizer steers the onboard camera toward reliablescene coordinates with low uncertainty, while a fixed-lag smoother fuses the low-rate SCR pose estimates with high-rate IMU data to provide a high-quality, high-rate pose estimate. In simulation, our planner reduces translationand rotation RMSE by at least 4.9% and 30.8% relative to baselines, respectively. Hardware-in-the-loop experiments validate the feasibility of our proposed trajectory planner under close-to-real deployment conditions.

Figures

Figures reproduced from arXiv: 2507.07467 by the authors.

**Figure 2.** Top-down 3-D reconstruction of the stata1 site produced by SCR. (left) RGB-colored points (right) Normalized entropy map. The accumulated scene coordinates and their uncertainties from multiple viewpoints reveal spatially consistent uncertainty distributions across the environment. and acceleration limits: Cie = Xr−1 i=1 X o∈{x,y,z,ψ} [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 4.** Qualitative results of E-SCRNet for a test image. (upper left) [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗

Figures from the paper (2 more)

**Figure 5.** Figure 5: Comparison of baseline planners (Forward, Vanilla, FIM) with the proposed method on “winter” and “ampersand” trajectories. (left) Density distributions of translation and rotation errors of the fixed-lag smoother for Forward (blue), Vanilla (orange), FIM (green), and O…

**Figure 6.** Figure 6: Hardware-in-the-loop experiment. (a) Drone and spaces for [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 40 canonical work pages

[1]

PAMPC: Perception-aware model predictive control for quadrotors

Davide Falanga, Philipp Foehn, Peng Lu, and Davide Scara- muzza. PAMPC: Perception-aware model predictive control for quadrotors. In2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 1–8, 2018

work page 2018
[2]

Perception-aware reced- ing horizon navigation for mavs

Zichao Zhang and Davide Scaramuzza. Perception-aware reced- ing horizon navigation for mavs. In2018 IEEE International Conference on Robotics and Automation (ICRA), pages 2534– 2541, 2018

work page 2018
[3]

Perception-aware trajectory generation for aggressive quadrotor flight using differential flatness

Varun Murali, Igor Spasojevic, Winter Guerra, and Sertac Kara- man. Perception-aware trajectory generation for aggressive quadrotor flight using differential flatness. In2019 American Control Conference (ACC), pages 3936–3943, 2019

work page 2019
[4]

Perception-aware path planning for UA Vs using semantic segmentation

Luca Bartolomei, Lucas Teixeira, and Margarita Chli. Perception-aware path planning for UA Vs using semantic segmentation. In2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 5808–5815, 2020

work page 2020
[5]

Perception- aware time optimal path parameterization for quadrotors

Igor Spasojevic, Varun Murali, and Sertac Karaman. Perception- aware time optimal path parameterization for quadrotors. In2020 IEEE International Conference on Robotics and Automation (ICRA), pages 3213–3219, 2020

work page 2020
[6]

Fisher Information Field: an Efficient and Differentiable Map for Perception-aware Planning

Zichao Zhang and Davide Scaramuzza. Fisher information field: an efficient and differentiable map for perception-aware planning, 2020. arXiv:2008.03324

work page Pith review arXiv 2020
[7]

Dabin Kim, Gyeong Chan Kim, Youngseok Jang, and H. Jin Kim. Topology-guided path planning for reliable visual naviga- tion of mavs. In2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 3117–3124, 2021

work page 2021
[8]

Learning where to look: Self-supervised viewpoint selection for active localization using geometrical information

Luca Di Giammarino, Boyang Sun, Giorgio Grisetti, Marc Polle- feys, Hermann Blum, and Daniel Barath. Learning where to look: Self-supervised viewpoint selection for active localization using geometrical information. InEuropean Conference on Computer Vision, pages 188–205. Springer, 2024

work page 2024

Show all 41 references

[9]

APACE: Agile and perception-aware trajectory generation for quadrotor flights

Xinyi Chen, Yichen Zhang, Boyu Zhou, and Shaojie Shen. APACE: Agile and perception-aware trajectory generation for quadrotor flights. In2024 IEEE International Conference on Robotics and Automation (ICRA), pages 17858–17864, 2024

2024
[10]

Perception- aware planning for quadrotor flight in unknown and feature- limited environments, 2025

Chenxin Yu, Zihong Lu, Jie Mei, and Boyu Zhou. Perception- aware planning for quadrotor flight in unknown and feature- limited environments, 2025. arXiv:2503.15273

2025 arXiv
[11]

Dabin Kim, Inkyu Jang, Youngsoo Han, Sunwoo Hwang, and H. Jin Kim. Enhancing feature tracking reliability for visual navigation using real-time safety filter, 2025. arXiv:2502.01092

2025 arXiv
[12]

Scene coordinate regression forests for camera relocalization in RGB-D images

Jamie Shotton, Ben Glocker, Christopher Zach, Shahram Izadi, Antonio Criminisi, and Andrew Fitzgibbon. Scene coordinate regression forests for camera relocalization in RGB-D images. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2013

2013
[13]

DSAC - differentiable RANSAC for camera localization

Eric Brachmann, Alexander Krull, Sebastian Nowozin, Jamie Shotton, Frank Michel, Stefan Gumhold, and Carsten Rother. DSAC - differentiable RANSAC for camera localization. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), July 2017

2017
[14]

Visual camera re- localization from RGB and RGB-D images using DSAC.IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(9):5847–5865, 2022

Eric Brachmann and Carsten Rother. Visual camera re- localization from RGB and RGB-D images using DSAC.IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(9):5847–5865, 2022

2022
[15]

Hierarchical scene coordinate classification and regression for visual localization

Xiaotian Li, Shuzhe Wang, Yi Zhao, Jakob Verbeek, and Juho Kannala. Hierarchical scene coordinate classification and regression for visual localization. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recog- nition (CVPR), June 2020

2020
[16]

HSCNet++: Hierarchical scene coordinate classification and regression for visual localization with transformer, 2023

Shuzhe Wang, Zakaria Laskar, Iaroslav Melekhov, Xiaotian Li, Yi Zhao, Giorgos Tolias, and Juho Kannala. HSCNet++: Hierarchical scene coordinate classification and regression for visual localization with transformer, 2023. arXiv:2305.03595

2023 arXiv
[17]

Accelerated coordinate encoding: Learning to re- localize in minutes using rgb and poses

Eric Brachmann, Tommaso Cavallari, and Victor Adrian Prisacariu. Accelerated coordinate encoding: Learning to re- localize in minutes using rgb and poses. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 5044–5053, June 2023

2023
[18]

GLACE: Global local accelerated coordinate encoding

Fangjinhua Wang, Xudong Jiang, Silvano Galliani, Christoph V ogel, and Marc Pollefeys. GLACE: Global local accelerated coordinate encoding. InProceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition (CVPR), pages 21562–21571, June 2024

2024
[19]

R-SCoRe: Revisiting scene coor- dinate regression for robust large-scale visual localization

Xudong Jiang, Fangjinhua Wang, Silvano Galliani, Christoph V ogel, and Marc Pollefeys. R-SCoRe: Revisiting scene coor- dinate regression for robust large-scale visual localization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pag...

2025
[20]

Deep evidential regression

Alexander Amini, Wilko Schwarting, Ava Soleimany, and Daniela Rus. Deep evidential regression. volume 33, pages 14927–14937, 2020

2020
[21]

PoseNet: A convolutional network for real-time 6-DOF camera relocaliza- tion

Alex Kendall, Matthew Grimes, and Roberto Cipolla. PoseNet: A convolutional network for real-time 6-DOF camera relocaliza- tion. InProceedings of the IEEE International Conference on Computer Vision (ICCV), December 2015

2015
[22]

Geometric loss functions for camera pose regression with deep learning

Alex Kendall and Roberto Cipolla. Geometric loss functions for camera pose regression with deep learning. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), July 2017

2017
[23]

Image-based localization using lstms for structured feature correlation

Florian Walch, Caner Hazirbas, Laura Leal-Taixe, Torsten Sat- tler, Sebastian Hilsenbeck, and Daniel Cremers. Image-based localization using lstms for structured feature correlation. In Proceedings of the IEEE International Conference on Computer Vision (ICCV), Oct 2017

2017
[24]

Modelling uncertainty in deep learning for camera relocalization

Alex Kendall and Roberto Cipolla. Modelling uncertainty in deep learning for camera relocalization. In2016 IEEE International Conference on Robotics and Automation (ICRA), pages 4762– 4769, 2016

2016
[25]

CoordiNet: Uncertainty- aware pose regressor for reliable vehicle localization

Arthur Moreau, Nathan Piasco, Dzmitry Tsishkou, Bogdan Stan- ciulescu, and Arnaud de La Fortelle. CoordiNet: Uncertainty- aware pose regressor for reliable vehicle localization. InPro- ceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pages ...

2022
[26]

Understanding the limitations of cnn-based absolute cam- era pose regression

Torsten Sattler, Qunjie Zhou, Marc Pollefeys, and Laura Leal- Taixe. Understanding the limitations of cnn-based absolute cam- era pose regression. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2019

2019
[27]

From coarse to fine: Robust hierarchical localization at large scale

Paul-Edouard Sarlin, Cesar Cadena, Roland Siegwart, and Marcin Dymczyk. From coarse to fine: Robust hierarchical localization at large scale. InProceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition (CVPR), June 2019

2019
[28]

Dropout as a bayesian approximation: Representing model uncertainty in deep learning

Yarin Gal and Zoubin Ghahramani. Dropout as a bayesian approximation: Representing model uncertainty in deep learning. Ininternational conference on machine learning, pages 1050–
[29]

Simple and scalable predictive uncertainty estimation using deep ensembles

Balaji Lakshminarayanan, Alexander Pritzel, and Charles Blun- dell. Simple and scalable predictive uncertainty estimation using deep ensembles. volume 30, 2017

2017
[30]

What uncertainties do we need in bayesian deep learning for computer vision? In I

Alex Kendall and Yarin Gal. What uncertainties do we need in bayesian deep learning for computer vision? In I. Guyon, U. V on Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett, editors,Advances in Neural Information Pro- cessing Systems, volume 30. Cur...

2017
[31]

Cavalheiro, and Sertac Karaman

Juyeop Han, Lukas Lao Beyer, Guilherme V . Cavalheiro, and Sertac Karaman. NVINS: Robust visual inertial navigation fused with nerf-augmented camera pose regressor and uncertainty quantification. In2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)...

2024
[32]

Natural posterior network: Deep bayesian uncertainty for exponential family distributions,

Bertrand Charpentier, Oliver Borchert, Daniel Z ¨ugner, Simon Geisler, and Stephan G ¨unnemann. Natural posterior network: Deep bayesian uncertainty for exponential family distributions,
[33]

Lever- aging neural radiance fields for uncertainty-aware visual local- ization

Le Chen, Weirong Chen, Rui Wang, and Marc Pollefeys. Lever- aging neural radiance fields for uncertainty-aware visual local- ization. In2024 IEEE International Conference on Robotics and Automation (ICRA), pages 6298–6305, 2024

2024
[34]

Osteen, and Nicholas Roy

Siddharth Ancha, Philip R. Osteen, and Nicholas Roy. Deep evidential uncertainty estimation for semantic segmentation un- der out-of-distribution obstacles. In2024 IEEE International Conference on Robotics and Automation (ICRA), pages 6943– 6951, 2024

2024
[35]

Evidential semantic mapping in off-road environments with uncertainty- aware bayesian kernel inference

Junyoung Kim, Junwon Seo, and Jihong Min. Evidential semantic mapping in off-road environments with uncertainty- aware bayesian kernel inference. In2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 1420–1427, 2024

2024
[36]

Osteen, Bernadette Bucher, Stephen Phillips, Jiuguang Wang, Michael Everett, Nicholas Roy, and Jonathan P

Xiaoyi Cai, Siddharth Ancha, Lakshay Sharma, Philip R. Osteen, Bernadette Bucher, Stephen Phillips, Jiuguang Wang, Michael Everett, Nicholas Roy, and Jonathan P. How. EVORA: Deep ev- idential traversability learning for risk-aware off-road autonomy. IEEE Transactions on Roboti...

2024
[37]

Minimum snap trajectory generation and control for quadrotors

Daniel Mellinger and Vijay Kumar. Minimum snap trajectory generation and control for quadrotors. In2011 IEEE Interna- tional Conference on Robotics and Automation, pages 2520– 2525, 2011

2011
[38]

A convex optimization approach to smooth trajectories for motion planning with car-like robots

Zhijie Zhu, Edward Schmerling, and Marco Pavone. A convex optimization approach to smooth trajectories for motion planning with car-like robots. In2015 54th IEEE conference on decision and control (CDC), pages 835–842. IEEE, 2015

2015
[39]

FlightGoggles: Photorealistic sensor simulation for perception-driven robotics using photogrammetry and virtual re- ality

Winter Guerra, Ezra Tal, Varun Murali, Gilhyun Ryou, and Sertac Karaman. FlightGoggles: Photorealistic sensor simulation for perception-driven robotics using photogrammetry and virtual re- ality. In2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)...

2019
[40]

The blackbird UA V dataset.The International Journal of Robotics Research, 39(10-11):1346– 1364, 2020

Amado Antonini, Winter Guerra, Varun Murali, Thomas Sayre- McCord, and Sertac Karaman. The blackbird UA V dataset.The International Journal of Robotics Research, 39(10-11):1346– 1364, 2020

2020
[41]

Steven G. Johnson. The NLopt nonlinear-optimization package, 2007

2007

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

PAMPC: Perception-aware model predictive control for quadrotors

Davide Falanga, Philipp Foehn, Peng Lu, and Davide Scara- muzza. PAMPC: Perception-aware model predictive control for quadrotors. In2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 1–8, 2018

work page 2018

[2] [2]

Perception-aware reced- ing horizon navigation for mavs

Zichao Zhang and Davide Scaramuzza. Perception-aware reced- ing horizon navigation for mavs. In2018 IEEE International Conference on Robotics and Automation (ICRA), pages 2534– 2541, 2018

work page 2018

[3] [3]

Perception-aware trajectory generation for aggressive quadrotor flight using differential flatness

Varun Murali, Igor Spasojevic, Winter Guerra, and Sertac Kara- man. Perception-aware trajectory generation for aggressive quadrotor flight using differential flatness. In2019 American Control Conference (ACC), pages 3936–3943, 2019

work page 2019

[4] [4]

Perception-aware path planning for UA Vs using semantic segmentation

Luca Bartolomei, Lucas Teixeira, and Margarita Chli. Perception-aware path planning for UA Vs using semantic segmentation. In2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 5808–5815, 2020

work page 2020

[5] [5]

Perception- aware time optimal path parameterization for quadrotors

Igor Spasojevic, Varun Murali, and Sertac Karaman. Perception- aware time optimal path parameterization for quadrotors. In2020 IEEE International Conference on Robotics and Automation (ICRA), pages 3213–3219, 2020

work page 2020

[6] [6]

Fisher Information Field: an Efficient and Differentiable Map for Perception-aware Planning

Zichao Zhang and Davide Scaramuzza. Fisher information field: an efficient and differentiable map for perception-aware planning, 2020. arXiv:2008.03324

work page Pith review arXiv 2020

[7] [7]

Dabin Kim, Gyeong Chan Kim, Youngseok Jang, and H. Jin Kim. Topology-guided path planning for reliable visual naviga- tion of mavs. In2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 3117–3124, 2021

work page 2021

[8] [8]

Learning where to look: Self-supervised viewpoint selection for active localization using geometrical information

Luca Di Giammarino, Boyang Sun, Giorgio Grisetti, Marc Polle- feys, Hermann Blum, and Daniel Barath. Learning where to look: Self-supervised viewpoint selection for active localization using geometrical information. InEuropean Conference on Computer Vision, pages 188–205. Springer, 2024

work page 2024