RASR: Range-Aware Scale Recovery for Metric UAV Navigation

Chenxu Wang; Fangwei Ye; Fuhui Zhou; Hongtao Liang; Jiahuan Ji; Qihui Wu; Xinyu Shao; Yiyao Wan

REVIEW 3 major objections 4 minor 33 references

RASR: Range-Aware Scale Recovery for Metric UAV Navigation

T0 review · 3 major / 4 minor · reviewed 2026-08-02 · deepseek-v4-flash

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

Pith's one-line read Range-aware residual correction lowers metric distance error in UAV navigation beyond global scale calibration.

desk verdict Range-aware stacking gives a real but unvalidated gain on PairUAV; the range proxy that routes everything never gets checked, so treat the headline numbers as provisional. read the letter →

arxiv 2607.09815 v2 pith:J3AZBOHX submitted 2026-07-10 cs.RO cs.CV

Hongtao Liang , Xinyu Shao , Chenxu Wang , Yiyao Wan , Jiahuan Ji , Fangwei Ye , Fuhui Zhou , Qihui Wu This is my paper

classification cs.ROcs.CV

keywords UAVnavigationmetricdistanceestimationrange-awarescalerecoverypairwisegeometryMASt3Rglobalcalibrationrelativeerrorimage-goal

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 claims that the metric distance error left over after fitting a single global scale to pairwise 3D geometry is not random: it varies systematically with how far away the goal is. To exploit that pattern, RASR builds a range proxy from the geometry itself, splits the calibration data into seven proxy buckets, and learns different mixtures of four candidate distance estimates per bucket, then applies a final regime-specific affine correction. On the PairUAV benchmark, this yields a total error of 0.003189, lower than the 0.003424 of global scale calibration alone, and on a scene-disjoint test set it cuts relative distance error from 0.004033 to 0.002605. The implication is that frozen dense stereo geometry contains enough metric signal for navigation once the residual scale errors are corrected per range, not globally.

What carries the argument

The central mechanism is the range proxy s_i = |h_1(z_i)|, a scalar predicted from the 422-dimensional dense-tabular summary of frozen MASt3R geometry. Calibration-set cut points turn this proxy into seven buckets, and within each bucket a fixed convex combination of four candidate distance heads is optimized against the relative-error objective. A second stage groups pairs by the candidate spread and bucket into regimes that select an affine correction before 2.5 m grid alignment. The proxy substitutes for unavailable ground-truth distance, making the correction range-aware while remaining computable at inference from a single image pair.

What would settle it

On the scene-disjoint set, compute the Spearman rank correlation between the range proxy s_i and the true distance d_i. If the correlation is low (say below 0.8) or if bucket assignment accuracy falls sharply compared with the calibration set, the reported 32–88% per-bin error reductions should be re-examined. Alternatively, train the same pipeline with a deliberately miscalibrated proxy (e.g., a proxy with flipped bucket ordering on one third of pairs) and show the gain is erased.

Watch

Extended reading notes

Core claim

RASR establishes that a single global affine scale calibration of MASt3R pairwise geometry leaves systematic, range-dependent distance residuals, and that these residuals can be corrected without ground-truth distance at inference. The correction is conditioned on a predicted range proxy s_i = |h_1(z_i)|, the absolute output of a chosen distance head, which routes each image pair into one of seven buckets. Inside each bucket, four candidate distances produced by structurally different heads are combined with fixed convex weights optimized on 367,416 calibration pairs, and a final quantile-based regime chooses an affine correction plus 2.5 m command-grid rounding. The method's official PairUA

Load-bearing premise

The method assumes the range proxy s_i = |h_1(z_i)|, fitted on the calibration set, preserves the ordering of true distance on unseen scenes; if this ordering breaks down across scenes, the per-bucket mixtures and corrections fitted on calibration data will not transfer.

Editorial extensions

If this is right

Frozen pairwise geometry (MASt3R) suffices for metric distance estimation per image pair once scale and range-dependent residuals are corrected.
Range-aware residual correction transfers to scenes not used for calibration: on the scene-disjoint set it cuts relative distance error from 0.004033 to 0.002605.
Separating scale recovery from task-specific command calibration allows the same core to be reused for other protocols.
Routing based on a predicted proxy outperforms both a global mapping and a best-single-head alternative, showing that adapting to range is the source of gain.

Reading between the lines

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

The paper does not measure how accurately the range proxy tracks true distance; if proxy ordering degrades across scenes, the bucket corrections should degrade. A direct check would be the rank correlation between s_i and true distance on the scene-disjoint set.
The same range-aware stacking idea could be applied to monocular depth models trained with AbsRel objectives, where the same distance-normalized weighting makes range-dependent residuals likely.
The 2.5 m grid alignment is protocol-specific; in a real controller, the residual correction should be retained without the grid rounding, and the error gain on continuous distances remains untested.
Since the descriptor contains scene-agnostic statistics, the approach might be paired with other dense geometry backbones (e.g., DUSt3R) with only recalibration.

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 modest, honest engineering paper with one strong controlled result and one load-bearing assumption that never gets tested. The controlled comparison in Table 2 is the right one: with protocol-specific calibration enabled for both variants, full RASR drops relative distance error from 0.004033 to 0.002605 on a scene-disjoint set. That is a ~35% reduction, and it's a legitimate empirical finding if it holds. The paper also does something sensible structurally: it separates the scale-recovery core from the protocol-specific command calibration, and it evaluates on disjoint scenes rather than just the challenge test set. That's more than many challenge papers do.

The soft spot is the range proxy. Everything downstream — the bucket assignment, the convex weights, the regime-specific affine corrections — is conditioned on s_i = |h1(z_i)|, the absolute output of one fitted head. The paper never measures how well that proxy tracks true distance on unseen scenes, never reports bucket purity, and never tests sensitivity to misrouting. Figure 4 bins by true distance, which shows the range-aware correction helps in diagnosis but says nothing about whether the proxy routes correctly at inference. If h1's ordering degrades across scenes, the whole gain could shrink or reverse. That is a load-bearing gap, not a picky detail.

Other issues are smaller. No error bars or significance tests anywhere; Table 2 is a single scene-disjoint snapshot. Table 1 mixes protocol calibration, so the official 0.003189 vs 0.003424 is not a matched comparison. And the architecture details are thin — four heads with 'three design families' and no layer counts, which makes reproduction harder even with code released.

On the plus side, the weighted-median derivation in Eqs. (2)-(3) is standard and clean, the calibration set is large (367k pairs), and there's no sign of circular reasoning. The paper's own related work correctly places range-conditioned scaling as a known idea in monocular depth, so this is an extension to a new task, not a conceptual breakthrough.

Who gets value: people working on metric scale recovery from frozen geometry, or on the PairUAV-style navigation benchmark. A serious referee should see it. The fixes are straightforward: validate the proxy, add error bars, and report the Table 1 comparison with matched protocol calibration. I'd send it to review with that expectation.

Referee Report

3 major / 4 minor

Summary. The paper presents RASR, a post-processing method for metric distance estimation in image-goal UAV navigation. RASR takes frozen MASt3R pairwise geometry, compresses it into a 422-dimensional descriptor, and uses four distance heads to produce candidate estimates. A range proxy, defined as the absolute output of one of these heads, routes each pair into one of seven buckets; within each bucket, convex weights blend the candidates, and a regime-specific affine correction followed by command-grid rounding produces the final distance. Heading is handled by a separate fixed transform. The authors evaluate on the PairUAV benchmark, reporting a final error of 0.003189 on the official hidden test set and a scene-disjoint ablation in which full RASR reduces relative distance error from 0.004033 to 0.002605 compared with global scale calibration under matched protocol calibration. The central claim is that range-aware residual correction, conditioned on a predicted range proxy, improves metric distance estimation beyond a single global affine scale calibration.

Significance. If the central claim holds, RASR provides a simple and modular way to turn frozen pairwise geometry into metric navigation commands, which is practically relevant for GNSS-denied UAV navigation. The paper is transparent about the protocol-specific calibration module, uses a frozen backbone, fixes all parameters before hidden-test inference, and releases code. The scene-disjoint diagnostic set is a genuine attempt to measure transfer. However, the evidence currently rests on a single matched evaluation without uncertainty quantification, and the load-bearing range proxy is not validated. The contribution is plausible and potentially useful, but the paper needs additional analysis to support the transferability claim at the level expected for a journal publication.

major comments (3)

[§3.2, Eqs. (4)–(5)] The range proxy s_i = |h_1(z_i)| is the sole routing signal for all per-bucket weights and regime-specific corrections, yet the paper never reports how well this proxy tracks true distance on unseen scenes. The scene-disjoint set provides one transfer point, but a single aggregate error does not establish that the proxy preserves distance ordering across scenes. Please report proxy-to-truth correlation and bucket purity/confusion on both the calibration set and the scene-disjoint set, and add a sensitivity analysis in which the cut points are perturbed or routing is performed with an oracle true-distance proxy. Without this, the improvement in Table 2 could reflect calibration-set-specific ordering that does not generalize.
[Tables 1 and 2] No error bars, confidence intervals, or significance tests are reported for any of the central comparisons. In Table 2 the headline distance-error improvement is 0.004033 to 0.002605; in Table 1 the final-error improvement is 0.003424 to 0.003189. These are small absolute differences, and the official hidden test is a single split. Please provide scene-level or sub-sampled bootstrap confidence intervals and a paired test (e.g., Wilcoxon) over scenes for the Table 2 comparison. Also clarify whether the Table 1 global-scale-calibration row uses the same protocol-specific calibration module as RASR; if not, Table 1 does not isolate the range-aware contribution.
[§4, Table 2 ablation interpretation] The ablation shows that bucketed convex stacking without the protocol-specific calibration module is substantially worse than global scale calibration without that module (0.007944 vs. 0.006656). The full RASR gain therefore appears to depend on the final regime-specific affine correction in Eq. (5), not on range-aware stacking alone. Please analyze the contribution of each component separately (e.g., stacking without regime correction, regime correction on a global combination, and full RASR), and discuss whether the central claim should be attributed to range-aware stacking, the regime-specific correction, or their interaction. This is important for understanding what 'range-aware residual correction' actually contributes beyond global calibration plus a per-bucket affine map.

minor comments (4)

[§3.2, descriptor definition] The 422-dimensional descriptor is only described by counts (144 point-map statistics, 28 confidence statistics, 250 descriptor statistics). Please list the specific statistics used, or provide a reference to code with a stable commit hash, so the method is reproducible without reverse-engineering.
[§3.3, Eq. (6)] The heading transform notation is hard to parse. Define wrap_180 explicitly and state where the constants 1.014 and 1.2 come from (fitted on calibration set?). The term 'fixed heading transform' should clarify whether these constants are learned or manually set.
[Figure 3] The mixture-weight bars are difficult to read from the figure; the numeric labels are small. Consider a table of weights or a larger font.
[Throughout] There are several spacing/typo issues, e.g., 'UA V' instead of 'UAV', 'VGGT and Fast3R' formatting in Section 2, and inconsistent reference spacing. A careful proofread is needed.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: RASR's parameters are fit on calibration data and evaluated on held-out pairs; no step reduces to its inputs by construction.

full rationale

The derivation chain is a standard calibration/fitting pipeline. Eq. (3) is a textbook weighted-median argument and does not smuggle in the result. The range proxy s_i=|h1(z_i)| is a fitted head's output, but it is used only to define buckets for per-bucket weights (Eq. 4) and regime-specific affine corrections (Eq. 5); all such parameters are fit on the 367,416-pair calibration set and evaluated on scene-disjoint 14,856 pairs plus 2,773,116 hidden pairs, with parameters fixed before inference. The improvement of RASR over global scale calibration is measured on held-out data (Table 2) under matched protocol calibration, so it is not forced by fitting. The only reference to the authors' own prior work ([5]) appears in related-work context and does not carry the central derivation. The concern that the range proxy may not transfer across scenes is a robustness/overfitting risk, not circularity; the paper's scene-disjoint evaluation is evidence of transfer, albeit limited. No quoted step reduces Eq. X to Eq. Y by definition, and no fitted parameter is renamed as a prediction. The Table 1 comparison may be protocol-mismatched, but a matched comparison exists in Table 2, so this is an experimental-design issue, not circularity.

Assumptions & free parameters 7 free parameters · 6 assumptions · 0 invented entities

This is an empirical calibration pipeline. The only theoretical component (weighted-median argument) is standard and does not fix any parameter. All headline numbers are products of parameters fitted to a calibration set, so the load-bearing content is in those fitting choices, not in a first-principles derivation.

free parameters (7)

Global scale calibration slope and intercept (alpha, beta) = not reported
Fitted to 367,416 calibration pairs; baseline mapping from MASt3R distance proxy to meters.
Six range-proxy cut points = 5.6, 14.2, 33.7, 53.8, 77.4, 114.8 m
Determined from calibration set; define seven routing buckets used by Eqs. (4)-(5).
Bucketed convex weights w_{b,m} = four weights per bucket (28 total, simplex-constrained)
Optimized with SLSQP on calibration set to minimize relative error Eq. (2).
Regime affine corrections a_{b,q}, c_{b,q} = not reported
Fit per (bucket, spread-interval, sign) regime in Eq. (5).
Heading transform constants = 1.014, 1.2, 20
Fixed affine/alignment transform for heading in Eq. (6), presumably calibrated on the calibration set.
Weights of four distance heads h1..h4 = not reported
Trained/fitted on calibration set; h1 doubles as the range proxy.
Hand-designed descriptor composition = 422 dims = 144 point-map + 28 confidence + 250 descriptor statistics
Choice of which statistics to extract is manual; no derivation is given.

assumptions (6)

standard math Weighted absolute deviation minimizer is the weighted median
Eqs. (2)-(3) use this to justify distance-dependent weighting; standard result, not original.
domain assumption MASt3R ViT-L/16 produces relative geometry whose scale is recoverable by affine calibration
The whole pipeline assumes frozen pairwise geometry carries a usable metric-scale signal; supported empirically by Table 1 but not guaranteed.
domain assumption Range proxy |h1(z)| orders pairs by true distance well enough for bucket transfer
Cut points and per-bucket corrections are only useful if the proxy is stable across scenes; never measured.
domain assumption Relative-error metric with epsilon floor matches navigation value
The objective in Eq. (2) weights short distances heavily; paper says this is analogous to AbsRel but does not justify it as a navigation safety metric.
domain assumption Calibration, diagnostic, and hidden sets are scene-disjoint with no leakage
The claim's validity depends on the benchmark's split integrity; the paper states disjointness but provides no verification.
ad hoc to paper 2.5 m command grid and 20-degree heading alignment are valid protocol outputs
Q2.5 rounding in Eq. (5) is introduced specifically for the PairUAV output protocol; the paper itself notes it is not physical resolution.

how reviews work

0 comments

Cite this review

Pith. "Pith review of RASR: Range-Aware Scale Recovery for Metric UAV Navigation." pith.science (2026). https://pith.science/paper/J3AZBOHX

@misc{pith2026260709815,
  author       = {Pith},
  title        = {Pith review of: RASR: Range-Aware Scale Recovery for Metric UAV Navigation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/J3AZBOHX}},
  note         = {Machine review of arXiv:2607.09815}
}

read the original abstract

A central challenge in image-goal UAV navigation under Global Navigation Satellite System (GNSS) denial is estimating metric distance and heading between current and goal views. Dense pairwise geometry models capture relative scene structure, but without a calibrated metric scale, they cannot directly provide reliable distance estimates for navigation. Although global scale calibration corrects the dominant scale bias, the remaining errors vary systematically with distance. In this paper, Range-Aware Scale Recovery (RASR) is proposed, which complements global scale calibration with range-aware residual correction. RASR encodes pairwise geometry extracted by a frozen Matching And Stereo 3D Reconstruction (MASt3R) backbone as a compact descriptor and separates the scale-recovery core from task-specific command calibration. On the official online evaluation of the UAVs in Multimedia 2026 PairUAV challenge, RASR achieved a total error of 0.003189, achieving a lower total error than global scale calibration alone. The results demonstrate that range-aware residual correction improves metric distance estimation beyond global scale calibration. Code and materials are available at https://github.com/lht-research/rasr-pairuav.

Figures

Figures reproduced from arXiv: 2607.09815 by the authors.

**Figure 1.** Overview of RASR, where frozen pair geometry is summarized into a compact descriptor, calibrated by range-aware scale [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗

**Figure 3.** Mixture weights in each bucket over the four heads. [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

**Figure 4.** Clipped relative distance error in each bucket. True [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗

Review history (2 revisions) →

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

33 extracted references · 3 linked inside Pith

[1]

Dhruv Shah, Ajay Sridhar, Arjun Bhorkar, Noriaki Hirose, and Sergey Levine
[2]

Dhruv Shah, Ajay Sridhar, Nitish Dashora, Kyle Stachowicz, Kevin Black, No- riaki Hirose, and Sergey Levine. 2023. ViNT: A Foundation Model for Visual Navigation. InProceedings of the 7th Conference on Robot Learning (Proceedings of Machine Learning Research, Vol. 229). PMLR, Atlanta, GA, USA, 711–733

2023
[3]

Ajay Sridhar, Dhruv Shah, Catherine Glossop, and Sergey Levine. 2024. NoMaD: Goal Masked Diffusion Policies for Navigation and Exploration. InProceedings of the IEEE International Conference on Robotics and Automation. IEEE, Yokohama, Japan, 63–70

2024
[4]

Yijie Deng, Shuaihang Yuan, and Yi Fang. 2026. AnyImageNav: Any-View Ge- ometry for Precise Last-Meter Image-Goal Navigation. arXiv:2604.05351 [cs.RO]

arXiv 2026
[5]

Hongtao Liang, Yihe Diao, Yuhang Wu, Fuhui Zhou, and Qihui Wu. 2026. Syn- ergetic Empowerment: Wireless Communications Meets Embodied Intelligence. IEEE Communications Magazine64 (2026), 1–8. doi:10.1109/MCOM.001. 2500569

doi:10.1109/mcom.001 2026
[6]

Yaxuan Li, Jiarui Zeng, Shaofei Huang, and Zhedong Zheng. 2026. Last-Meter Precision Navigation for UA Vs: A Diffusion-Refined Aerial Visual Servoing Approach. arXiv:2607.04352 [cs.CV] https://arxiv.org/abs/2607.04352

arXiv 2026
[7]

Fabian Deuser, Yaxuan Li, Tingyu Wang, Yujiao Shi, Anna Bößendörfer, Shaofei Huang, Xiao Pan, Zhedong Zheng, and Roger Zimmermann. 2026. The 4th Work- shop on UA Vs in Multimedia: Capturing the World from a New Perspective. In Proceedings of the 34th ACM International Conference on Multimedia Workshop. Association for Computing Machinery, New York, NY , U...

2026
[8]

Shuzhe Wang, Vincent Leroy, Yohann Cabon, Boris Chidlovskii, and Jerome Revaud. 2024. DUSt3R: Geometric 3D Vision Made Easy. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE/CVF, Seattle, W A, USA, 20697–20709

2024

Show all 33 references

[9]

Vincent Leroy, Yohann Cabon, and Jerome Revaud. 2024. Grounding Image Matching in 3D with MASt3R. InComputer Vision – ECCV 2024. Springer Nature Switzerland, Cham, Switzerland, 71–91

2024
[10]

Wei Yin, Chi Zhang, Hao Chen, Zhipeng Cai, Gang Yu, Kaixuan Wang, Xiaozhi Chen, and Chunhua Shen. 2023. Metric3D: Towards Zero-shot Metric 3D Pre- diction from A Single Image. InProceedings of the IEEE/CVF International Conference on Computer Vision. IEEE/CVF, Paris, France, 9043–9053

2023
[11]

Luigi Piccinelli, Yung-Hsu Yang, Christos Sakaridis, Mattia Segu, Siyuan Li, Luc Van Gool, and Fisher Yu. 2024. UniDepth: Universal Monocular Metric Depth Estimation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE/CVF, Seattle, W A, U...

2024
[12]

Lihe Yang, Bingyi Kang, Zilong Huang, Zhen Zhao, Xiaogang Xu, Jiashi Feng, and Hengshuang Zhao. 2024. Depth Anything V2. InAdvances in Neural In- formation Processing Systems, V ol. 37. Curran Associates, Inc., Vancouver, BC, Canada, 21875–21911

2024
[13]

Sijie Zhu, Mubarak Shah, and Chen Chen. 2022. TransGeo: Transformer Is All You Need for Cross-View Image Geo-Localization. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE/CVF, New Orleans, LA, USA, 1162–1171

2022
[14]

Xiaohan Zhang, Xingyu Li, Waqas Sultani, Yi Zhou, and Safwan Wshah. 2023. Cross-View Geo-Localization via Learning Disentangled Geometric Layout Cor- respondence.Proceedings of the AAAI Conference on Artificial Intelligence37, 3 (2023), 3480–3488

2023
[15]

Fabian Deuser, Konrad Habel, and Norbert Oswald. 2023. Sample4Geo: Hard Neg- ative Sampling for Cross-View Geo-Localisation. InProceedings of the IEEE/CVF International Conference on Computer Vision. IEEE/CVF, Paris, France, 16801– 16810

2023
[16]

Wenmiao Hu, Yichen Zhang, Yuxuan Liang, Yifang Yin, Andrei Georgescu, An Tran, Hannes Kruppa, See-Kiong Ng, and Roger Zimmermann. 2022. Beyond Geo-localization: Fine-grained Orientation of Street-view Images by Cross-view Matching with Satellite Imagery. InProceedings of the 3...

2022
[17]

Li Mi, Chang Xu, Javiera Castillo-Navarro, Syrielle Montariol, Wen Yang, Antoine Bosselut, and Devis Tuia. 2024. ConGeo: Robust Cross-View Geo- Localization Across Ground View Variations. InComputer Vision – ECCV 2024. Springer Nature Switzerland, Cham, Switzerland, 214–230

2024
[18]

Junyan Ye, Honglin Lin, Leyan Ou, Dairong Chen, Zihao Wang, Qi Zhu, Con- ghui He, and Weijia Li. 2025. Where am I? Cross-View Geo-Localization with Natural Language Descriptions. InProceedings of the IEEE/CVF International Conference on Computer Vision. IEEE/CVF, Honolulu, HI,...

2025
[19]

Zhedong Zheng, Yunchao Wei, and Yi Yang. 2020. University-1652: A Multi-view Multi-source Benchmark for Drone-based Geo-localization. InProceedings of the 28th ACM International Conference on Multimedia. Association for Computing Machinery, New York, NY , USA, 1395–1403

2020
[20]

Yuxiang Ji, Boyong He, Zhuoyue Tan, and Liaoni Wu. 2025. Game4Loc: A UA V Geo-Localization Benchmark from Game Data.Proceedings of the AAAI Conference on Artificial Intelligence39, 4 (2025), 3913–3921

2025
[21]

Ming Dai, Jianhong Hu, Jiedong Zhuang, and Enhui Zheng. 2022. A Transformer- Based Feature Segmentation and Region Alignment Method for UA V-View Geo- Localization.IEEE Transactions on Circuits and Systems for Video Technology 32, 7 (2022), 4376–4389

2022
[22]

Ming Dai, Enhui Zheng, Zhenhua Feng, Qi Lei, Jiedong Zhuang, and Wankou Yang. 2024. Vision-Based UA V Self-Positioning in Low-Altitude Urban Environ- ments.IEEE Transactions on Image Processing33 (2024), 493–508

2024
[23]

Jianyuan Wang, Minghao Chen, Nikita Karaev, Andrea Vedaldi, Christian Rup- precht, and David Novotny. 2025. VGGT: Visual Geometry Grounded Trans- former. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE/CVF, Nashville, TN, USA, 5294–5306

2025
[24]

Liang, Mikael Henaff, Hao Tang, Ang Cao, Joyce Chai, Franziska Meier, and Matt Feiszli

Jianing Yang, Alexander Sax, Kevin J. Liang, Mikael Henaff, Hao Tang, Ang Cao, Joyce Chai, Franziska Meier, and Matt Feiszli. 2025. Fast3R: Towards 3D Reconstruction of 1000+ Images in One Forward Pass. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Re...

2025
[25]

Junyi Zhang, Charles Herrmann, Junhwa Hur, Varun Jampani, Trevor Darrell, Forrester Cole, Deqing Sun, and Ming-Hsuan Yang. 2025. MonST3R: A Simple Approach for Estimating Geometry in the Presence of Motion. International Conference on Learning Representations. arXiv:2410.03825 [cs.CV]

2025 arXiv
[26]

Efros, and Angjoo Kanazawa

Qianqian Wang, Yifei Zhang, Aleksander Holynski, Alexei A. Efros, and Angjoo Kanazawa. 2025. Continuous 3D Perception Model with Persistent State. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recog- nition. IEEE/CVF, Nashville, TN, USA, 10510–10522

2025
[27]

Siyan Dong, Shuzhe Wang, Shaohui Liu, Lulu Cai, Qingnan Fan, Juho Kannala, and Yanchao Yang. 2025. Reloc3r: Large-Scale Training of Relative Camera Pose Regression for Generalizable, Fast, and Accurate Visual Localization. InProceed- ings of the IEEE/CVF Conference on Computer...

2025
[28]

Mu Hu, Wei Yin, Chi Zhang, Zhipeng Cai, Xiaoxiao Long, Hao Chen, Kaixuan Wang, Gang Yu, Chunhua Shen, and Shaojie Shen. 2024. Metric3D v2: A Versatile Monocular Geometric Foundation Model for Zero-shot Metric Depth and Surface Normal Estimation.IEEE Transactions on Pattern Ana...

2024
[29]

Lihe Yang, Bingyi Kang, Zilong Huang, Xiaogang Xu, Jiashi Feng, and Heng- shuang Zhao. 2024. Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE/CVF, Seattle, W A, USA, 10...

2024
[30]

Bingxin Ke, Anton Obukhov, Shengyu Huang, Nando Metzger, Rodrigo Caye Daudt, and Konrad Schindler. 2024. Repurposing Diffusion-Based Image Genera- tors for Monocular Depth Estimation. InProceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition. IEEE/C...

2024
[31]

Paul-Edouard Sarlin, Daniel DeTone, Tomasz Malisiewicz, and Andrew Rabi- novich. 2020. SuperGlue: Learning Feature Matching with Graph Neural Net- works. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE/CVF, Seattle, W A, USA, 4938–4947

2020
[32]

Philipp Lindenberger, Paul-Edouard Sarlin, and Marc Pollefeys. 2023. LightGlue: Local Feature Matching at Light Speed. InProceedings of the IEEE/CVF Interna- tional Conference on Computer Vision. IEEE/CVF, Paris, France, 17581–17592

2023
[2023]

InProceedings of the IEEE International Conference on Robotics and Automation

GNM: A General Navigation Model to Drive Any Robot. InProceedings of the IEEE International Conference on Robotics and Automation. IEEE, London, United Kingdom, 7226–7233

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.

[1] [1]

Dhruv Shah, Ajay Sridhar, Arjun Bhorkar, Noriaki Hirose, and Sergey Levine

[2] [2]

Dhruv Shah, Ajay Sridhar, Nitish Dashora, Kyle Stachowicz, Kevin Black, No- riaki Hirose, and Sergey Levine. 2023. ViNT: A Foundation Model for Visual Navigation. InProceedings of the 7th Conference on Robot Learning (Proceedings of Machine Learning Research, Vol. 229). PMLR, Atlanta, GA, USA, 711–733

2023

[3] [3]

Ajay Sridhar, Dhruv Shah, Catherine Glossop, and Sergey Levine. 2024. NoMaD: Goal Masked Diffusion Policies for Navigation and Exploration. InProceedings of the IEEE International Conference on Robotics and Automation. IEEE, Yokohama, Japan, 63–70

2024

[4] [4]

Yijie Deng, Shuaihang Yuan, and Yi Fang. 2026. AnyImageNav: Any-View Ge- ometry for Precise Last-Meter Image-Goal Navigation. arXiv:2604.05351 [cs.RO]

arXiv 2026

[5] [5]

Hongtao Liang, Yihe Diao, Yuhang Wu, Fuhui Zhou, and Qihui Wu. 2026. Syn- ergetic Empowerment: Wireless Communications Meets Embodied Intelligence. IEEE Communications Magazine64 (2026), 1–8. doi:10.1109/MCOM.001. 2500569

doi:10.1109/mcom.001 2026

[6] [6]

Yaxuan Li, Jiarui Zeng, Shaofei Huang, and Zhedong Zheng. 2026. Last-Meter Precision Navigation for UA Vs: A Diffusion-Refined Aerial Visual Servoing Approach. arXiv:2607.04352 [cs.CV] https://arxiv.org/abs/2607.04352

arXiv 2026

[7] [7]

Fabian Deuser, Yaxuan Li, Tingyu Wang, Yujiao Shi, Anna Bößendörfer, Shaofei Huang, Xiao Pan, Zhedong Zheng, and Roger Zimmermann. 2026. The 4th Work- shop on UA Vs in Multimedia: Capturing the World from a New Perspective. In Proceedings of the 34th ACM International Conference on Multimedia Workshop. Association for Computing Machinery, New York, NY , U...

2026

[8] [8]

Shuzhe Wang, Vincent Leroy, Yohann Cabon, Boris Chidlovskii, and Jerome Revaud. 2024. DUSt3R: Geometric 3D Vision Made Easy. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE/CVF, Seattle, W A, USA, 20697–20709

2024