Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Relative Pose Estimation through Affine Corrections of Monocular Depth Priors

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Estimating the unknown scale and shift of monocular depth maps jointly with camera motion turns depth priors into a reliable geometric cue that consistently improves relative pose estimation over keypoint-only baselines, in both…

desk verdict A solid, well-executed solver paper: explicit affine (scale+shift) correction of depth priors consistently helps relative pose, and the authors honestly flag where their global-affine model breaks down. read the letter →

arxiv 2501.05446 v3 pith:VX73DKX2 submitted 2025-01-09 cs.CV

Yifan Yu , Shaohui Liu , Rémi Pautrat , Marc Pollefeys , Viktor Larsson This is my paper · ORCID
classification cs.CV
keywords relativeposeestimationmonoculardepthpriorsaffinecorrectionminimalsolvershybridRANSACfocallengthdepth-guidedgeometryessentialmatrix
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

Monocular depth models usually predict depth only up to an unknown scale and shift, yet existing geometric pipelines feed those predictions to solvers that assume the ambiguity is just a single scale. This paper argues that ignoring the per-image shift is the main reason depth priors have not helped relative pose estimation, and it develops three solvers that jointly estimate the affine corrections $\beta_1, \beta_2$ and the scale ratio $\alpha$ together with the relative pose — for calibrated cameras, for a shared unknown focal length, and for two unknown focal lengths. The solvers are embedded in a hybrid RANSAC pipeline that also uses classic keypoint solvers, so the depth priors contribute when they are reliable and are safely ignored when they are not. Across indoor and outdoor benchmarks, the hybrid estimator beats keypoint-only baselines and prior scale-only depth-based methods, and the improvement holds across different feature matchers and depth models — including, surprisingly, models trained to predict metric depth.

What carries the argument

The load-bearing identity is the pairwise-distance constraint $||\delta^{(1)}_{jk}||^2 = ||\delta^{(2)}_{jk}||^2$ between corresponding lifted 3D points: since a rigid transform preserves distances, these equations eliminate rotation and translation and can be solved first for the depth scale $\alpha$, the shifts $\beta_1, \beta_2$, and the focal lengths. The paper re-parameterizes with $\gamma = \alpha^2$ and $\omega = 1/f^2$ to lower the polynomial degree and constructs Gröbner-basis solvers via syzygy-based elimination, yielding at most 4 solutions for the calibrated and two-focal cases and 8 for the shared-focal case. Rotation and translation are then recovered by an orthogonal Procrustes alignment of the affinely corrected 3D points, and the solvers run inside a hybrid LO-MSAC that alternates between depth-aware and classic minimal solvers and jointly minimizes reprojection-plus-Sampson cost.

What would settle it

Take image pairs with ground-truth depth and corrupt each depth prior with two different shifts applied to two spatial halves of the scene, sweeping the shift difference from zero up to roughly half the median depth. If the global-affine solver's pose error rises toward or above the point-based baseline as the regional shift difference grows, the single-global-correction model is the limiting factor — the regime the paper's own limitation section says occurs in practice.

Watch

Extended reading notes

Core claim

The paper's central object is the affine depth-correction model $\widehat{D}_1 = D_1 + \beta_1$, $\widehat{D}_2 = \alpha(D_2 + \beta_2)$, in which each monocular depth map is corrected by its own shift and the two maps are related by a single scale ratio before back-projection. Because rigid motion preserves distances, the squared distances between corresponding lifted 3D points must agree across the two views, and these constraints contain no rotation or translation; they form a polynomial system in the affine parameters and any unknown focal lengths. From that system the paper builds three solvers — a minimal calibrated 3-point solver, a shared-focal 4-point solver, and a two-focal 4-point solver — recovers rotation and translation by rigid alignment of the corrected 3D points, and wraps the whole thing in a hybrid locally optimized RANSAC that scores both depth-induced reprojection error and classic Sampson error. On ScanNet, MegaDepth, ETH3D, and a generated two-focal benchmark, the method consistently outperforms keypoint-only baselines and prior scale-only depth solvers, and the shift term proves beneficial even for metric depth priors.

Load-bearing premise

Each depth map is assumed to be correct up to a single global scale and a single global shift for the entire image, so every pairwise-distance constraint inherits that one-correction model and would break if the depth error varies across the scene.

Editorial extensions

If this is right

  • Modeling the shift, not just the scale, of monocular depth priors is what unlocks their value: the scale-only ablated version loses most of the gain, and the synthetic study shows shifts above 10% of median depth are common even with metric depth models.
  • The affine-corrected depth cue helps most where keypoint geometry is weakest: image pairs with limited covisibility show the largest pose-error reductions.
  • The hybrid design makes the gains portable — the same pipeline improves over point-based baselines with SuperPoint+LightGlue, RoMa, and MASt3R matches and with Omnidata, Marigold, Depth-Anything, and MoGe depths, and the gains grow as those models improve.
  • Uncalibrated settings gain the most: on the two-focal benchmark the median pose and focal errors drop to roughly half the 7-point fundamental matrix baseline, showing that depth priors also anchor focal length estimation.
  • The pipeline is fast enough for practical use, with median runtimes of 31ms (calibrated), 65ms (shared-focal), and 129ms (two-focal) per image pair on CPU.

Reading between the lines

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

  • A fitted shift value can be read as a per-image bias estimate for a depth model, which suggests a practical diagnostic: large fitted shifts on a 'metric' model flag systematic bias, and an affine fit against ground truth could serve as a model-quality measure for geometry tasks instead of raw metric error.
  • The same global-affine correction idea should transfer to any task that aligns per-view depth or shape priors, such as point-cloud registration, SLAM front-ends, or multi-view bundle adjustment over the affine parameters, because those pipelines face the identical scale-and-shift ambiguity.
  • The most natural extension is region-based affine correction — estimating different shifts for different surfaces, guided by segmentation — which would keep the pairwise-distance machinery unchanged while removing the paper's main limitation; the paper itself flags this direction.
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 paper proposes three RANSAC-based solvers for two-view relative pose estimation that jointly estimate the relative pose and per-image affine (scale and shift) corrections of monocular depth priors, covering calibrated, shared-focal, and two-focal camera settings. The solvers are integrated into a hybrid LO-MSAC pipeline that combines depth-aware solvers with classic point-based solvers and both reprojection and Sampson errors. Experiments on ScanNet, MegaDepth, ETH3D, and a generated 2D-3D-S benchmark show consistent improvements over keypoint-only baselines and prior depth-based baselines across multiple matchers and MDE models.

Significance. If the solver derivations are correct, the paper offers a practical and effective way to leverage off-the-shelf monocular depth priors for classical geometric estimation. The empirical study is unusually thorough: multiple datasets, multiple feature matchers, multiple MDE models, ablations of the hybrid components, and runtime measurements are reported, and code is released. The non-circularity burden is also met: pose accuracy is measured against ground-truth pose, and the depth priors come from independently trained models. The main weaknesses are that the solver derivations are asserted rather than demonstrated, and the acknowledged limitation of the global affine model is never stress-tested.

major comments (3)
  1. [Sec. 4.1] The three solver descriptions omit the actual polynomial systems and only assert solution counts ("at most 4", "8", "4") and template sizes (12x12, 36x36, 40x40). Since the solvers are the central contribution, the manuscript should provide the equations after the substitutions γ=α², ω=1/f², and ω1,ω2, or give a supplementary derivation. In particular, the shared-focal solver "only select[s] 4 out of the 6 possible equations" and the two-focal solver "requires using 5 out of the 6 equations," but which equations are selected is not stated. If the choice is arbitrary, this should be justified or shown to be invariant; otherwise readers cannot verify correctness or reproduce the solver.
  2. [Sec. D and Eq. (5)] The distance constraints in Eq. (5) are exact only under the global affine model in Eq. (4). Section D concedes that the estimated shift β can vary across regions and surfaces and that the same depth map can yield different β groupings for different image pairs. The paper never quantifies the residual spatially varying component after the global fit, nor does it test the solvers' sensitivity to this misspecification. Because Eq. (5) is the foundation of all three solvers, this is a load-bearing gap: a synthetic experiment with per-region shifts, or an analysis of the per-pixel residual after fitting the global affine model, is needed to establish the generality of the claimed improvements and to understand when the distance constraints become inconsistent.
  3. [Sec. 4.1 and Sec. C] The shared-focal and two-focal solvers are not minimal and drop one or two of the six pairwise distance constraints. The manuscript acknowledges this but does not analyze the consequences: the solved (α, β1, β2) may not satisfy the full set of constraints on the minimal sample, and the subsequent rigid alignment may therefore be biased by the choice of which equations are dropped. The authors state that the hybrid RANSAC pipeline can handle this, but no experiment reports how often minimal samples are inconsistent or whether the dropped-equation choice affects accuracy. At minimum, the paper should specify the equation selection rule and add a synthetic test of solver stability on noise-free data.
minor comments (5)
  1. [Sec. 3] The notation changes from Eq. (1), with a_i and b_i, to Eq. (4), with α, β1, β2; the relation β_i = b_i/a_i should be stated explicitly in the main text.
  2. [Sec. 4.1] After the reparameterization γ=α², the paper should state whether α is assumed positive or whether both signs are checked when recovering the scale.
  3. [Fig. 6] The axes in Fig. 6 are not labeled with units; the reader must infer that the horizontal axis is the shift normalized by the median depth. Please label the axes consistently with Fig. 5.
  4. [Tabs. 3-5] The rows labeled "Reference entry - DUSt3R" and "Reference entry - MASt3R" are not clearly integrated with the main comparison; it would be clearer to state in the caption that these are reference results, not baselines run under the same RANSAC protocol.
  5. [Sec. 4.4 and Supp. A] The hyperparameters τr, τs, and λs are described as tuned or fixed without a sensitivity analysis; a short table showing performance under reasonable variations of these thresholds would increase confidence in the conclusions.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: pose and affine corrections are jointly estimated from external depth priors and correspondences, and all claims are checked against ground-truth pose.

full rationale

The derivation chain is self-contained. Eq. (4) postulates an affine model for the depth priors; Eq. (5) follows from the length-preserving property of rigid transforms and is not defined in terms of the pose output. The calibrated/shared-focal/two-focal solvers are derived from these pairwise-distance constraints and are solved with a general algebraic method [34]; the solution counts are facts about the polynomial systems, not fitted values. The affine parameters (α, β1, β2) and focal lengths are estimated jointly with R, t from the same test-time correspondences and depth priors, and are never fitted to the ground-truth pose used for evaluation; therefore the reported pose accuracy is an empirical measurement, not a consequence of construction. The ablation in Table 6 measures a real difference between including and omitting shift parameters, and the synthetic experiment in Fig. 5 varies an injected shift while keeping the evaluation metric fixed, so the 'shift helps' claim is not forced by the solver design. The paper's self-citations ([12], [33], [34]) are to generic RANSAC/hybrid-estimation tooling, a pose library, and a general minimal-solver reduction technique; none imports a problem-specific uniqueness theorem or an ansatz that already contains the conclusion. The limitation in Sec. D that β can vary across regions/surfaces and even across image pairs is a genuine model-misspecification caveat—it could affect robustness—but it is not circularity: it does not make Eq. (5) equivalent to an input, nor does it rename a fitted value as a prediction. The only tuned quantities are experimental thresholds and λs, which are hyperparameters rather than load-bearing derivation steps.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The method introduces no new physical entities; the affine corrections (α, β1, β2) are estimation parameters, not invented objects. The free parameters are experimental hyperparameters tuned per dataset. The main assumptions are standard rigid geometry and a global affine model of MDE depth errors, the latter explicitly acknowledged as an idealization by the authors.

free parameters (3)
  • RANSAC reprojection threshold τr = 8px (ScanNet, ETH3D), 16px (MegaDepth, 2D-3D-S)
    Set per dataset for best performance of method and baselines (Supp. A, Hyperparameters). Determines inlier sets for depth-induced reprojection error.
  • Sampson error threshold τs = 2px (ScanNet), 1px (other datasets)
    Set per dataset in Supp. A; controls which correspondences count as inliers for the epipolar error term.
  • Sampson error weight λs = 1.0
    Empirically fixed in all experiments; balances depth-induced reprojection error and Sampson error in Eq. (9)-(10). The paper notes this is tunable to account for depth prior reliability.
assumptions (5)
  • standard math Rigid scene: pairwise distances between corresponding 3D points are preserved under the relative pose, so Eq. (5) holds.
    Invoked in Sec. 4.1 to eliminate R and t and build polynomial systems in α, β1, β2, and focal lengths.
  • domain assumption Global affine ambiguity model: predicted depth is related to metric depth by bD1 = D1 + β1 and bD2 = α(D2 + β2) (Eq. 4).
    Central modeling assumption about MDE outputs; the paper's Section D acknowledges that shifts can vary across regions, so this is an idealization.
  • domain assumption Simplified intrinsics: uncalibrated cameras have K = diag(f, f, 1) with zero principal point and square pixels (Eq. 6).
    Used in the shared-focal and two-focal solvers; ignores principal point offsets and aspect ratio, which can bias focal estimates on real images.
  • standard math Polynomial systems have the claimed finite solution counts (≤4, 8, ≤4) generated via the syzygy-based method of Larsson et al. [34].
    The solver construction in Sec. 4.1 depends on this computational algebra result and on the reported template sizes; counts are asserted, not proved in the paper.
  • domain assumption A sufficient fraction of correspondences and depth priors are inliers for RANSAC/MSAC to find the correct hypothesis.
    The hybrid LO-MSAC pipeline (Sec. 4.4) relies on a workable inlier ratio for both point and depth data types; the method is not designed for near-zero inlier rates.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Relative Pose Estimation through Affine Corrections of Monocular Depth Priors." pith.science (2026). https://pith.science/paper/VX73DKX2

@misc{pith2026250105446,
  author       = {Pith},
  title        = {Pith review of: Relative Pose Estimation through Affine Corrections of Monocular Depth Priors},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VX73DKX2}},
  note         = {Machine review of arXiv:2501.05446}
}
read the original abstract

Monocular depth estimation (MDE) models have undergone significant advancements over recent years. Many MDE models aim to predict affine-invariant relative depth from monocular images, while recent developments in large-scale training and vision foundation models enable reasonable estimation of metric (absolute) depth. However, effectively leveraging these predictions for geometric vision tasks, in particular relative pose estimation, remains relatively under explored. While depths provide rich constraints for cross-view image alignment, the intrinsic noise and ambiguity from the monocular depth priors present practical challenges to improving upon classic keypoint-based solutions. In this paper, we develop three solvers for relative pose estimation that explicitly account for independent affine (scale and shift) ambiguities, covering both calibrated and uncalibrated conditions. We further propose a hybrid estimation pipeline that combines our proposed solvers with classic point-based solvers and epipolar constraints. We find that the affine correction modeling is beneficial to not only the relative depth priors but also, surprisingly, the "metric" ones. Results across multiple datasets demonstrate large improvements of our approach over classic keypoint-based baselines and PnP-based solutions, under both calibrated and uncalibrated setups. We also show that our method improves consistently with different feature matchers and MDE models, and can further benefit from very recent advances on both modules. Code is available at https://github.com/MarkYu98/madpose.

Figures

Figures reproduced from arXiv: 2501.05446 by the authors.

Figure 1
Figure 1. Our method jointly estimates affine corrections of [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Pipeline overview: Our method takes a pair of images as input, runs off-the-shelf feature matching and monocular depth estimation, then jointly estimates the relative pose, scale and shift parameters of the two depth maps, and optionally the focal lengths. Our method incorporates monocular depth priors in all stages (in green) of hybrid LO-MSAC [12, 35], including 3 new depth-aware solvers, while still being able to… view at source ↗
Figure 3
Figure 3. Pose error AUCs on sampled indoor ETH3D [ [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Visualization on ETH3D [59]. Left: back-projected GT depth with pose found by point-based method (translation rescaled to match GT); Middle: back-projected depth priors from Marigold [32] aligned using the scale, shifts, pose, and focal length from our method; Right: G…
Figure 5
Figure 5. Figure 5: Rotation and translation error by adding shift values to [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Distribution (probabilistic density) of GT shift values [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Additional visualizations on ETH3D [59] with shared-focal setting. Left: back-projected GT depth with relative pose found by PoseLib-6pt [33] and translation rescaled to match scale with GT translation; Middle: back-projected depth priors from Marigold [32] aligned usi…
Figure 8
Figure 8. Figure 8: Additional visualizations on Stanford 2D-3D-S [ [PITH_FULL_IMAGE:figures/full_fig_p015_8.png]
Figure 9
Figure 9. Figure 9: Visualization of aligned point clouds on image pairs from ScanNet-1500[ [PITH_FULL_IMAGE:figures/full_fig_p016_9.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. RePoseD: Efficient Relative Pose Estimation With Known Depth Information

    cs.CV 2025-01 conditional novelty 6.0 of 10

    New efficient minimal solvers estimate relative camera pose jointly with unknown depth scale and shift, improving speed and often accuracy over prior depth-aware solvers.

Reference graph

Works this paper leans on

84 extracted references · 69 canonical work pages · cited by 1 Pith paper

  1. [4]

    Relative pose solvers us- ing monocular depth

    Daniel Barath and Chris Sweeney. Relative pose solvers us- ing monocular depth. In ICPR, 2022. 1, 2, 5, 6, 13

    2022

  2. [17]

    Fundamental matrix esti- mation using relative depths

    Yaqing Ding, V ´aclav V´avra, Snehal Bhayani, Qianliang Wu, Jian Yang, and Zuzana Kukelova. Fundamental matrix esti- mation using relative depths. In ECCV, 2024. 1, 2, 5, 6, 7, 12

    work page 2024

  3. [1]

    Ceres Solver, 2023

    Sameer Agarwal, Keir Mierle, and The Ceres Solver Team. Ceres Solver, 2023. 6

    2023

  4. [2]

    Joint 2d-3d-semantic data for indoor scene understanding

    Iro Armeni, Sasha Sax, Amir R Zamir, and Silvio Savarese. Joint 2d-3d-semantic data for indoor scene understanding. arXiv:1702.01105, 2017. 5, 7, 12, 13, 15

    arXiv 2017

  5. [3]

    Graph-cut RANSAC

    Daniel Barath and Jiri Matas. Graph-cut RANSAC. In CVPR, 2018. 5, 12, 13

    2018

  6. [5]

    Stereoglue: Robust estimation with single-point solvers

    Daniel Barath, Dmytro Mishkin, Luca Cavalli, Paul-Edouard Sarlin, Petr Hruby, and Marc Pollefeys. Stereoglue: Robust estimation with single-point solvers. In ECCV, 2024. 2

    2024

  7. [6]

    Two-view geometry scoring without correspondences

    Axel Barroso-Laguna, Eric Brachmann, Victor Adrian Prisacariu, Gabriel J Brostow, and Daniyar Turmukhambe- tov. Two-view geometry scoring without correspondences. In CVPR, 2023. 2

    work page 2023

  8. [7]

    Adabins: Depth estimation using adaptive bins

    Shariq Farooq Bhat, Ibraheem Alhashim, and Peter Wonka. Adabins: Depth estimation using adaptive bins. In CVPR,

    work page

Show all 84 references
  1. [8]

    Zoedepth: Zero-shot transfer by com- bining relative and metric depth

    Shariq Farooq Bhat, Reiner Birkl, Diana Wofk, Peter Wonka, and Matthias M ¨uller. Zoedepth: Zero-shot transfer by com- bining relative and metric depth. arXiv:2302.12288, 2023. 1, 2

    2023 arXiv

  2. [9]

    Midas v3.1 – a model zoo for robust monocular relative depth estimation

    Reiner Birkl, Diana Wofk, and Matthias M ¨uller. Midas v3.1 – a model zoo for robust monocular relative depth estimation. arXiv:2307.14460, 2023. 1, 3

    2023 arXiv

  3. [10]

    Depth pro: Sharp monocular metric depth in less than a second

    Aleksei Bochkovskii, Ama ¨el Delaunoy, Hugo Germain, Marcel Santos, Yichao Zhou, Stephan R Richter, and Vladlen Koltun. Depth pro: Sharp monocular metric depth in less than a second. arXiv:2410.02073, 2024. 2

    2024 arXiv

  4. [11]

    From projective to euclidean space un- der any practical situation, a criticism of self-calibration

    Sylvain Bougnoux. From projective to euclidean space un- der any practical situation, a criticism of self-calibration. In ICCV, 1998. 5

    1998

  5. [12]

    Hybrid Camera Pose Estimation

    Federico Camposeco, Andrea Cohen, Marc Pollefeys, and Torsten Sattler. Hybrid Camera Pose Estimation. In CVPR,

  6. [13]

    Locally opti- mized ransac

    Ond ˇrej Chum, Ji ˇr´ı Matas, and Josef Kittler. Locally opti- mized ransac. In Pattern Recognition: 25th DAGM Sympo- sium, Magdeburg, Germany, September 10-12, 2003. Pro- ceedings 25, 2003. 2

    2003

  7. [14]

    Deepfactors: Real-time probabilistic dense monocular slam

    Jan Czarnowski, Tristan Laidlow, Ronald Clark, and An- drew J Davison. Deepfactors: Real-time probabilistic dense monocular slam. IEEE Robotics and Automation Letters , 5 (2):721–728, 2020. 1, 2

    2020

  8. [15]

    Chang, Manolis Savva, Maciej Hal- ber, Thomas Funkhouser, and Matthias Nießner

    Angela Dai, Angel X. Chang, Manolis Savva, Maciej Hal- ber, Thomas Funkhouser, and Matthias Nießner. Scannet: Richly-annotated 3d reconstructions of indoor scenes. In CVPR, 2017. 5, 6, 8, 12, 13, 16

    2017

  9. [16]

    Superpoint: Self-supervised interest point detection and description

    Daniel DeTone, Tomasz Malisiewicz, and Andrew Rabi- novich. Superpoint: Self-supervised interest point detection and description. In CVPR Deep Learning for Visual SLAM Workshop, 2018. 2, 3, 5

    2018

  10. [18]

    D2-net: A trainable cnn for joint detection and description of local features

    Mihai Dusmanu, Ignacio Rocco, Tomas Pajdla, Marc Polle- feys, Josef Sivic, Akihiko Torii, and Torsten Sattler. D2-net: A trainable cnn for joint detection and description of local features. arXiv:1905.03561, 2019. 2

    1905 arXiv

  11. [19]

    Dkm: Dense kernelized feature matching for geometry estimation

    Johan Edstedt, Ioannis Athanasiadis, M ˚arten Wadenb ¨ack, and Michael Felsberg. Dkm: Dense kernelized feature matching for geometry estimation. In CVPR, 2023. 2

    2023

  12. [20]

    Roma: Robust dense feature matching

    Johan Edstedt, Qiyu Sun, Georg B ¨okman, M ˚arten Wadenb¨ack, and Michael Felsberg. Roma: Robust dense feature matching. In CVPR, 2024. 2, 3, 6

    2024

  13. [21]

    Omnidata: A scalable pipeline for making multi-task mid-level vision datasets from 3d scans

    Ainaz Eftekhar, Alexander Sax, Jitendra Malik, and Amir Zamir. Omnidata: A scalable pipeline for making multi-task mid-level vision datasets from 3d scans. In CVPR, 2021. 3

    2021

  14. [22]

    Relative pose from deep learned depth and a single affine correspondence

    Ivan Eichhardt and Daniel Barath. Relative pose from deep learned depth and a single affine correspondence. In ECCV,

  15. [23]

    Depth map prediction from a single image using a multi-scale deep net- work

    David Eigen, Christian Puhrsch, and Rob Fergus. Depth map prediction from a single image using a multi-scale deep net- work. In NeurIPS, 2014. 2

    2014

  16. [24]

    Fischler and Robert C

    Martin A. Fischler and Robert C. Bolles. Random sample consensus: A paradigm for model fitting with applications to image analysis and automated cartography. Commun. ACM,

  17. [25]

    Deep ordinal regression net- work for monocular depth estimation

    Huan Fu, Mingming Gong, Chaohui Wang, Kayhan Bat- manghelich, and Dacheng Tao. Deep ordinal regression net- work for monocular depth estimation. In CVPR, 2018. 2

    2018

  18. [26]

    Geowiz- ard: Unleashing the diffusion priors for 3d geometry estima- tion from a single image

    Xiao Fu, Wei Yin, Mu Hu, Kaixuan Wang, Yuexin Ma, Ping Tan, Shaojie Shen, Dahua Lin, and Xiaoxiao Long. Geowiz- ard: Unleashing the diffusion priors for 3d geometry estima- tion from a single image. In ECCV, 2024. 2

    2024

  19. [27]

    Digging into self-supervised monocular depth estimation

    Cl ´ement Godard, Oisin Mac Aodha, Michael Firman, and Gabriel J Brostow. Digging into self-supervised monocular depth estimation. In CVPR, 2019. 2

    2019

  20. [28]

    3d packing for self-supervised monocular depth estimation

    Vitor Guizilini, Rares Ambrus, Sudeep Pillai, Allan Raven- tos, and Adrien Gaidon. 3d packing for self-supervised monocular depth estimation. In CVPR, 2020. 1

    2020

  21. [29]

    Multiple View Ge- ometry in Computer Vision

    Richard Hartley and Andrew Zisserman. Multiple View Ge- ometry in Computer Vision . Cambridge University Press,

  22. [30]

    py- bind11 – seamless operability between c++11 and python,

    Wenzel Jakob, Jason Rhinelander, and Dean Moldovan. py- bind11 – seamless operability between c++11 and python,

  23. [31]

    3d common corruptions and data augmentation

    O ˘guzhan Fatih Kar, Teresa Yeo, Andrei Atanov, and Amir Zamir. 3d common corruptions and data augmentation. In CVPR, 2022. 1, 6, 12

    2022

  24. [32]

    Repurpos- ing diffusion-based image generators for monocular depth estimation

    Bingxin Ke, Anton Obukhov, Shengyu Huang, Nando Met- zger, Rodrigo Caye Daudt, and Konrad Schindler. Repurpos- ing diffusion-based image generators for monocular depth estimation. In CVPR, 2024. 1, 2, 3, 6, 7, 8, 12, 15

    2024

  25. [33]

    PoseLib - Minimal Solvers for Camera Pose Estimation, 2020

    Viktor Larsson and contributors. PoseLib - Minimal Solvers for Camera Pose Estimation, 2020. 2, 5, 12, 15

    2020

  26. [34]

    Effi- cient solvers for minimal problems by syzygy-based reduc- tion

    Viktor Larsson, Kalle Astrom, and Magnus Oskarsson. Effi- cient solvers for minimal problems by syzygy-based reduc- tion. In CVPR, 2017. 4, 13

    2017

  27. [35]

    Fixing the Lo- cally Optimized RANSAC

    Karel Lebeda, Jiri Matas, and Ondrej Chum. Fixing the Lo- cally Optimized RANSAC. In BMVC, 2012. 3, 5 9

    2012

  28. [36]

    Fixing the locally optimized ransac–full experimental evaluation

    Karel Lebeda, Jirı Matas, and Ondrej Chum. Fixing the locally optimized ransac–full experimental evaluation. In BMVC, 2012. 2

    2012

  29. [37]

    Ground- ing image matching in 3d with mast3r

    Vincent Leroy, Yohann Cabon, and J´erˆome Revaud. Ground- ing image matching in 3d with mast3r. In ECCV, 2024. 2, 5, 6, 7

    2024

  30. [38]

    Megadepth: Learning single- view depth prediction from internet photos

    Zhengqi Li and Noah Snavely. Megadepth: Learning single- view depth prediction from internet photos. In CVPR, 2018. 2, 5, 6, 7, 8, 12

    2018

  31. [39]

    Binsformer: Revisiting adaptive bins for monocular depth estimation

    Zhenyu Li, Xuyang Wang, Xianming Liu, and Junjun Jiang. Binsformer: Revisiting adaptive bins for monocular depth estimation. IEEE Transactions on Image Processing, 2024. 2

    2024

  32. [40]

    LightGlue: Local Feature Matching at Light Speed

    Philipp Lindenberger, Paul-Edouard Sarlin, and Marc Polle- feys. LightGlue: Local Feature Matching at Light Speed. In ICCV, 2023. 2, 3, 5

    2023

  33. [41]

    Depth-guided sparse structure-from-motion for movies and tv shows

    Sheng Liu, Xiaohan Nie, and Raffay Hamid. Depth-guided sparse structure-from-motion for movies and tv shows. In CVPR, 2022. 1, 2, 5

    2022

  34. [42]

    Deeprelativefusion: Dense monocular slam using single-image relative depth prediction

    Shing Yan Loo, Syamsiah Mashohor, Sai Hong Tang, and Hong Zhang. Deeprelativefusion: Dense monocular slam using single-image relative depth prediction. In IROS, 2021. 2

    2021

  35. [43]

    Distinctive image features from scale- invariant keypoints

    David G Lowe. Distinctive image features from scale- invariant keypoints. IJCV, 60:91–110, 2004. 2

    2004

  36. [44]

    Consistent video depth estimation.ACM TOG, 39(4):71–1, 2020

    Xuan Luo, Jia-Bin Huang, Richard Szeliski, Kevin Matzen, and Johannes Kopf. Consistent video depth estimation.ACM TOG, 39(4):71–1, 2020. 2

    2020

  37. [45]

    Faugeras

    Quang-Tuan Luong and Olivier D. Faugeras. The fundamen- tal matrix: Theory, algorithms, and stability analysis. IJCV,

  38. [46]

    An efficient solution to the five-point relative pose problem

    David Nist ´er. An efficient solution to the five-point relative pose problem. In CVPR, 2003. 2, 3, 5

    2003

  39. [47]

    Ecodepth: Effective conditioning of diffusion models for monocular depth estimation

    Suraj Patni, Aradhye Agarwal, and Chetan Arora. Ecodepth: Effective conditioning of diffusion models for monocular depth estimation. In CVPR, 2024. 2

    2024

  40. [48]

    Unidepth: Universal monocular metric depth estimation

    Luigi Piccinelli, Yung-Hsu Yang, Christos Sakaridis, Mattia Segu, Siyuan Li, Luc Van Gool, and Fisher Yu. Unidepth: Universal monocular metric depth estimation. In CVPR,

  41. [49]

    Remode: Probabilistic, monocular dense reconstruction in real time

    Matia Pizzoli, Christian Forster, and Davide Scaramuzza. Remode: Probabilistic, monocular dense reconstruction in real time. In ICRA, 2014. 2

    2014

  42. [50]

    Deep fundamental matrix estimation

    Ren ´e Ranftl and Vladlen Koltun. Deep fundamental matrix estimation. In ECCV, 2018. 2

    2018

  43. [51]

    Towards robust monocular depth estimation: Mixing datasets for zero-shot cross-dataset transfer

    Ren ´e Ranftl, Katrin Lasinger, David Hafner, Konrad Schindler, and Vladlen Koltun. Towards robust monocular depth estimation: Mixing datasets for zero-shot cross-dataset transfer. IEEE TPAMI, 44(3), 2022. 2, 3

    2022

  44. [52]

    Monocular depth esti- mation using neural regression forest

    Anirban Roy and Sinisa Todorovic. Monocular depth esti- mation using neural regression forest. In CVPR, 2016. 2

    2016

  45. [53]

    Paul D. Sampson. Fitting conic sections to ”very scattered” data: An iterative refinement of the bookstein algorithm. Comput. Graph. Image Process., 1982. 2, 5

    1982

  46. [54]

    SuperGlue: Learning feature matching with graph neural networks

    Paul-Edouard Sarlin, Daniel DeTone, Tomasz Malisiewicz, and Andrew Rabinovich. SuperGlue: Learning feature matching with graph neural networks. In CVPR, 2020. 2

    2020

  47. [55]

    RansacLib - A Template-based *SAC Implementation, 2019

    Torsten Sattler et al. RansacLib - A Template-based *SAC Implementation, 2019. 6

    2019

  48. [56]

    Learning depth from single monocular images

    Ashutosh Saxena, Sung Chung, and Andrew Ng. Learning depth from single monocular images. In NeurIPS, 2005. 2

    2005

  49. [57]

    Make3d: Learning 3d scene structure from a single still image

    Ashutosh Saxena, Min Sun, and Andrew Y Ng. Make3d: Learning 3d scene structure from a single still image. IEEE TPAMI, 31(5):824–840, 2008. 2

    2008

  50. [58]

    The surprising effectiveness of diffusion models for optical flow and monocular depth estimation

    Saurabh Saxena, Charles Herrmann, Junhwa Hur, Abhishek Kar, Mohammad Norouzi, Deqing Sun, and David J Fleet. The surprising effectiveness of diffusion models for optical flow and monocular depth estimation. In NeurIPS, 2024. 2

    2024

  51. [59]

    Sch¨onberger, Silvano Galliani, Torsten Sattler, Konrad Schindler, Marc Pollefeys, and An- dreas Geiger

    Thomas Sch ¨ops, Johannes L. Sch¨onberger, Silvano Galliani, Torsten Sattler, Konrad Schindler, Marc Pollefeys, and An- dreas Geiger. A multi-view stereo benchmark with high- resolution images and multi-camera videos. In CVPR, 2017. 5, 7, 12, 13, 15

    2017

  52. [60]

    A generalized solution of the orthog- onal procrustes problem

    Peter H Sch ¨onemann. A generalized solution of the orthog- onal procrustes problem. In Psychometrika, 1966. 4

    1966

  53. [61]

    Animating street view

    Mengyi Shan, Brian Curless, Ira Kemelmacher-Shlizerman, and Steve Seitz. Animating street view. In SIGGRAPH Asia 2023 Conference Papers, pages 1–12, 2023. 1

    2023

  54. [62]

    D ¨arf: boosting radiance fields from sparse inputs with monocular depth adaptation

    Jiuhn Song, Seonghoon Park, Honggyu An, Seokju Cho, Min-Seop Kwak, Sungjin Cho, and Seungryong Kim. D ¨arf: boosting radiance fields from sparse inputs with monocular depth adaptation. In NeurIPS, 2023. 2

    2023

  55. [63]

    A minimal solution for relative pose with un- known focal length

    Henrik Stew ´enius, David Nist´er, Fredrik Kahl, and Frederik Schaffalitzky. A minimal solution for relative pose with un- known focal length. Image and Vision Computing, 2008. 2, 3, 5

    2008

  56. [64]

    Loftr: Detector-free local feature matching with transformers

    Jiaming Sun, Zehong Shen, Yuang Wang, Hujun Bao, and Xiaowei Zhou. Loftr: Detector-free local feature matching with transformers. In CVPR, 2021. 2

    2021

  57. [65]

    Cnn-slam: Real-time dense monocular slam with learned depth prediction

    Keisuke Tateno, Federico Tombari, Iro Laina, and Nassir Navab. Cnn-slam: Real-time dense monocular slam with learned depth prediction. In CVPR, pages 6243–6252, 2017. 2

    2017

  58. [66]

    Pdc-net+: Enhanced probabilistic dense cor- respondence network

    Prune Truong, Martin Danelljan, Radu Timofte, and Luc Van Gool. Pdc-net+: Enhanced probabilistic dense cor- respondence network. IEEE TPAMI, 45(8):10247–10266,

  59. [67]

    Disk: Learning local features with policy gradient

    Michał Tyszkiewicz, Pascal Fua, and Eduard Trulls. Disk: Learning local features with policy gradient. In NeurIPS,

  60. [68]

    Matchformer: Interleaving attention in transformers for feature matching

    Qing Wang, Jiaming Zhang, Kailun Yang, Kunyu Peng, and Rainer Stiefelhagen. Matchformer: Interleaving attention in transformers for feature matching. In ACCV, 2022. 2

    2022

  61. [69]

    Moge: Unlocking accurate monocular geometry estimation for open-domain images with optimal training supervision

    Ruicheng Wang, Sicheng Xu, Cassie Dai, Jianfeng Xiang, Yu Deng, Xin Tong, and Jiaolong Yang. Moge: Unlocking accurate monocular geometry estimation for open-domain images with optimal training supervision. In CVPR, 2025. 1, 2, 6, 7, 12

    2025

  62. [70]

    Dust3r: Geometric 3d vi- sion made easy

    Shuzhe Wang, Vincent Leroy, Yohann Cabon, Boris Chidlovskii, and Jerome Revaud. Dust3r: Geometric 3d vi- sion made easy. In CVPR, 2024. 2, 5, 6

    2024

  63. [71]

    Generalized differentiable ransac

    Tong Wei, Yash Patel, Alexander Shekhovtsov, Jiri Matas, and Daniel Barath. Generalized differentiable ransac. In ICCV, 2023. 2 10

    2023

  64. [72]

    Nerfingmvs: Guided optimization of neural radiance fields for indoor multi-view stereo

    Yi Wei, Shaohui Liu, Yongming Rao, Wang Zhao, Jiwen Lu, and Jie Zhou. Nerfingmvs: Guided optimization of neural radiance fields for indoor multi-view stereo. In ICCV, 2021. 2

    2021

  65. [73]

    Depthsplat: Connecting gaussian splatting and depth

    Haofei Xu, Songyou Peng, Fangjinhua Wang, Hermann Blum, Daniel Barath, Andreas Geiger, and Marc Pollefeys. Depthsplat: Connecting gaussian splatting and depth. In CVPR, 2025. 2

    2025

  66. [74]

    Depth anything: Unleashing the power of large-scale unlabeled data

    Lihe Yang, Bingyi Kang, Zilong Huang, Xiaogang Xu, Jiashi Feng, and Hengshuang Zhao. Depth anything: Unleashing the power of large-scale unlabeled data. In CVPR, 2024. 1, 2, 6, 8, 12

    2024

  67. [75]

    Depth any- thing v2

    Lihe Yang, Bingyi Kang, Zilong Huang, Zhen Zhao, Xiao- gang Xu, Jiashi Feng, and Hengshuang Zhao. Depth any- thing v2. In NeurIPS, 2024. 1, 2, 3, 6, 7, 12, 13

    2024

  68. [76]

    Metric3d: Towards zero-shot metric 3d prediction from a single image

    Wei Yin, Chi Zhang, Hao Chen, Zhipeng Cai, Gang Yu, Kaixuan Wang, Xiaozhi Chen, and Chunhua Shen. Metric3d: Towards zero-shot metric 3d prediction from a single image. In ICCV, 2023. 1, 2

    2023

  69. [77]

    Monosdf: Exploring monocular geometric cues for neural implicit surface reconstruction

    Zehao Yu, Songyou Peng, Michael Niemeyer, Torsten Sat- tler, and Andreas Geiger. Monosdf: Exploring monocular geometric cues for neural implicit surface reconstruction. In NeurIPS, 2022. 2

    2022

  70. [78]

    Wordepth: Vari- ational language prior for monocular depth estimation

    Ziyao Zeng, Daniel Wang, Fengyu Yang, Hyoungseob Park, Stefano Soatto, Dong Lao, and Alex Wong. Wordepth: Vari- ational language prior for monocular depth estimation. In CVPR, 2024. 2

    2024

  71. [79]

    fcop: Focal length estimation from category-level object priors

    Xinyue Zhang, Jiaqi Yang, Xiangting Meng, Abdelrahman Mohamed, and Laurent Kneip. fcop: Focal length estimation from category-level object priors. arXiv:2409.19641, 2024. 2

    2024 arXiv

  72. [80]

    To- wards better generalization: Joint depth-pose learning with- out posenet

    Wang Zhao, Shaohui Liu, Yezhi Shu, and Yong-Jin Liu. To- wards better generalization: Joint depth-pose learning with- out posenet. In CVPR, 2020. 2

    2020

  73. [81]

    Aliked: A lighter keypoint and descriptor extraction network via deformable transformation

    Xiaoming Zhao, Xingming Wu, Weihai Chen, Peter CY Chen, Qingsong Xu, and Zhengguo Li. Aliked: A lighter keypoint and descriptor extraction network via deformable transformation. IEEE Transactions on Instrumentation and Measurement, 72:1–16, 2023. 2

    2023

  74. [82]

    To learn or not to learn: Visual localization from essential matrices

    Qunjie Zhou, Torsten Sattler, Marc Pollefeys, and Laura Leal-Taixe. To learn or not to learn: Visual localization from essential matrices. In ICRA, 2020. 2 11 Relative Pose Estimation through Affine Corrections of Monocular Depth Priors Supplementary Material A. Additional Det...

    2020

  75. [84]

    depth priors

    giving the best results. The accuracy can further benefit from developments on more accurate MDE models. We include here a comparison of using different monoc- ular depth estimation (MDE) models with our method across three tasks and three datasets in Tab. 8. Our method can im...

  76. [2017]

    https://github.com/pybind/pybind11. 6

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.