Pith. sign in

REVIEW 3 major objections 5 minor 48 references

An Evaluation of Feature Matchers for Fundamental Matrix Estimation

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read The classic SIFT+RANSAC pipeline can be upgraded modularly to estimate fundamental matrices more accurately, and a two-stage Coarse-to-Fine RANSAC does best on all tested datasets.

desk verdict Useful benchmark and practical matching systems, but the CF-RSC 'significantly outperforms' claim only holds on wide-baseline data. read the letter →

arxiv 1908.09474 v2 pith:46OJ6QSA submitted 2019-08-26 cs.CV

classification cs.CV
keywords fundamentalmatrixestimationfeaturematchingcorrespondencepruningRANSAClocalfeaturesevaluationbenchmarkepipolargeometryCoarse-to-Fine
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

This paper evaluates whether swapping modern components into the classic two-view matching pipeline—SIFT features, ratio test, RANSAC with the 8-point algorithm—improves fundamental matrix estimation. Using four large-scale datasets and resolution-normalized metrics, it reports that learned descriptors (HardNet++, RootSIFT-PCA), correspondence pruning (GMS, LPM), and the LMedS estimator each outperform their classic counterparts in the right settings. The authors then combine the winners into three practical matching systems that reach the accuracy of the expensive CODE matcher at far lower runtime. They also propose Coarse-to-Fine RANSAC, which uses graph-cut RANSAC only to prune outliers and LMedS to fit the model, and report that it achieves the highest recall on every dataset, well above the classic baseline.

What carries the argument

The evaluation is carried by three standardized metrics: the normalized symmetric geometric distance (NSGD), which divides the symmetric epipolar distance by the image diagonal so errors are comparable across resolutions; %Recall, the fraction of estimates with NSGD below 0.05; and %Inlier, the ratio of matches lying within a normalized epipolar threshold. The proposed method's engine is a two-stage estimator: graph-cut RANSAC first supplies a clean inlier set, then LMedS fits the fundamental matrix. The practical matching systems pair DoG or HesAffNet detectors with RootSIFT-PCA or HardNet++ descriptors, prune with GMS, and finish with LMedS fitting.

What would settle it

Select pairs for the same four datasets using a learned matcher's own inlier criterion rather than SIFT's, and obtain ground-truth cameras from known sensor poses or an independent reconstruction; if the recall ranking of HardNet++, GMS, LMedS, and CF-RSC changes substantially, the benchmark's SIFT-centric construction is the cause.

Watch

Extended reading notes

Core claim

The central finding is that the long-standing SIFT + ratio test + RANSAC/8-point recipe is not the ceiling for fundamental matrix estimation: each pipeline stage can be swapped for a newer alternative and yield measurable gains, and the gains compound. On the evaluation's normalized metric, HardNet++ and RootSIFT-PCA consistently beat SIFT descriptors, GMS and LPM raise inlier rates while keeping enough matches, and LMedS fits better than RANSAC when inlier rates are high. Most importantly, GC-RANSAC and USAC, despite their poor standalone fitting recall on short-baseline data, are excellent outlier pruners; feeding their cleaned correspondences into LMedS gives Coarse-to-Fine RANSAC, which tops all four datasets (for instance 90.7% recall on T&T versus 70.0% baseline and 60.9% on CPC versus 29.2%).

Load-bearing premise

The result rests on the benchmark's way of choosing which image pairs count as matchable and how ground-truth cameras are obtained: if SIFT-based pair selection and structure-from-motion reconstruction bias the test set toward SIFT-like features, the reported ranking of matchers and systems would be partly an artifact of the benchmark construction rather than intrinsic quality.

Editorial extensions

If this is right

  • A practitioner can choose components per scenario: DoG+HardNet++ with GMS and LMedS for general scenes, and HesAffNet+HardNet++ for wide-baseline scenes.
  • The two-stage idea of aggressive pruning followed by robust fitting can be dropped into any geometric estimation task that currently uses a RANSAC-family estimator.
  • The reported gains imply that benchmark scores of local features alone do not predict end-to-end geometry quality; the full matching-and-estimation pipeline must be evaluated.
  • Because CODE's accuracy is matched at several orders of magnitude lower cost, the proposed systems are practical for real-time or large-scale structure-from-motion and SLAM.
  • The evaluation protocol itself, with NSGD and %Recall, lets different datasets and image resolutions be compared on a single scale.

Reading between the lines

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

  • The benchmark constructs pairs using SIFT inliers and structure-from-motion geometry; if SIFT systematically misses pairs that learned features would match, the ranking could shift when pair selection is matcher-agnostic.
  • CF-RSC's separation of outlier removal from model fitting might generalize to essential matrix and homography estimation, where a similar coarse/fine behavior could be tested.
  • The normalized error metric makes cross-dataset comparison possible; reporting recall curves instead of a single threshold would show whether the ranking is threshold-dependent.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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 presents a comparative evaluation of local feature descriptors, correspondence pruning methods, and robust estimators for two-view fundamental matrix estimation. The evaluation is conducted on four datasets (TUM, KITTI, Tanks and Temples, and a Community Photo Collection) using a uniform pipeline with a SIFT baseline. The authors propose three matching systems that combine recent features with GMS pruning and LMedS fitting, and a Coarse-to-Fine RANSAC (CF-RSC) that uses GC-RANSAC for outlier pruning followed by LMedS model fitting. They report that CF-RSC significantly outperforms alternative estimators, and that the proposed matching systems perform on par with the much slower CODE system.

Significance. The paper addresses a timely and practical question: whether recent advances in local features, correspondence pruning, and robust estimation translate into improvements in a standard matching pipeline. The scale and breadth of the evaluation, the public release of the evaluation pipeline, and the careful definition of normalized metrics are genuine strengths. If the results are robust, the paper offers a useful practical guide and a simple, effective estimator. However, the headline claim of significant improvement for CF-RSC is currently under-supported, and the benchmark construction for two of the four datasets has a potential SIFT-related circularity. These issues need to be resolved before the paper's conclusions can be fully accepted.

major comments (3)
  1. [Section 5.3, Table 5] The statement that CF-RSC 'significantly outperforms other alternatives' is not justified by the reported numbers on the short-baseline datasets. On TUM, CF-RSC achieves 69.30% recall versus 69.20% for LMedS (a 0.1 percentage point difference); on KITTI the difference is 92.30% versus 91.80% (0.5 percentage points). With 1000 randomly selected test pairs per dataset, the approximate binomial standard error at these recall levels is about 1.5 and 0.9 percentage points, respectively, so both differences are within sampling error. No confidence intervals, variance over repeated runs, or paired significance tests are reported anywhere in the paper. The claim should be restricted to the wide-baseline datasets (T&T and CPC) where the gains are 7–16 percentage points, or it should be accompanied by a proper statistical analysis.
  2. [Section 4 (Image Pairs Construction and Ground Truth)] For the T&T and CPC datasets, the test pairs are selected based on the number of SIFT inliers (more than 20), and the ground-truth fundamental matrices are derived from COLMAP reconstructions, which themselves rely on SIFT matching. This introduces a potential circularity that can systematically favor SIFT-based features and methods operating on SIFT correspondences, and it may inflate the apparent improvements of learned descriptors and pruning methods on these datasets. The authors should explicitly acknowledge this limitation and, ideally, validate the main conclusions on a subset of pairs that are selected independently of SIFT or on data with sensor ground truth. The TUM and KITTI results provide some independent support, but the large differences on T&T and CPC, which drive much of the paper's practical advice, are affected by this concern.
  3. [General (Tables 2, 4, 5)] The paper does not report any measure of uncertainty for the %Recall and %Inlier numbers. Given that many comparisons in Table 2 are within 1–2 percentage points, the authors should either provide error bars, confidence intervals, or significance tests, or explicitly caution the reader against interpreting small differences as meaningful. This is particularly important for the feature comparison, where the ranking of methods on TUM and KITTI could change with a different random split.
minor comments (5)
  1. [Table 3] The runtime table mixes measurements from two different machines (L and W), so runtimes of methods evaluated on different machines are not directly comparable. Please state this explicitly in the caption or, better, report runtimes of all methods on a single machine for the headline comparisons.
  2. [Section 3.1] The choice of the %Recall threshold (0.05) and the inlier threshold parameter α (0.003) is arbitrary. The authors note that recall curves can be used; providing recall curves for the key comparisons (e.g., CF-RSC vs. LMedS) would strengthen the evaluation.
  3. [Tables 2 and 5] The abbreviation 'GC-RSC' is used in the tables but the method is introduced as GC-RANSAC in the text. Define the abbreviation in the table captions or use 'GC-RANSAC' throughout for clarity.
  4. [Section 1 (Introduction)] The claim that learned descriptors that perform better on standard benchmarks do not necessarily improve matching quality cites Balntas et al. [4]; consider also citing the HPatches benchmark [5] here, as it is already in the bibliography and directly supports this point.
  5. [References and Section 2] There is a typo in the reference [22] title: 'calibrarion' should be 'calibration'. Also, 'Armanguèet al.' in Section 2 should be 'Armangué et al.' with proper accent and spacing.

Circularity Check

0 steps flagged · score 1.0 of 10

No circular derivation; SIFT-based benchmark construction is a selection-bias caveat, not a circular reduction.

full rationale

This is an empirical evaluation paper rather than a derivation chain: it replaces components of a classic matching pipeline and measures %Recall on four datasets, with no parameter fitted to the reported metric and no uniqueness theorem imported from self-citation. The proposed matching systems are explicit compositions of published components (RootSIFT-PCA/HardNet++/HesAffNet + ratio test + GMS + LMedS), and CF-RSC is explicitly GC-RANSAC pruning followed by LMedS fitting; neither renames a known result nor defines its input in terms of its output. TUM and KITTI use sensor-based ground truth, giving independent support to the main comparisons. The only self-reference is that wide-baseline T&T/CPC test pairs are selected by SIFT inlier counts and their COLMAP ground truth is SIFT-based SfM, which can bias the ranking of matchers on those two datasets; however, this is a benchmark-selection caveat, not a case where an output equals an input by construction. The 'significantly outperforms' claim also lacks significance tests (e.g., TUM 69.30 vs 69.20), but statistical overclaim is a correctness risk, not circularity. Overall, the paper's central claims have independent content and no circular step is exhibited.

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

The central evaluation depends mainly on thresholds and on the assumption that the benchmark construction does not systematically favor the baseline matcher. No new physical entities are introduced; NSGD and CF-RSC are protocol and method choices.

free parameters (4)
  • Inlier-rate threshold alpha = 0.003
    Used to decide whether a match is an inlier in %Inlier and %Inlier-m; the value is chosen by hand and affects matching-quality numbers.
  • %Recall threshold = 0.05 (normalized SGD)
    FMs with NSGD below this value count as accurate; the paper asserts results are insensitive but shows no threshold sweep.
  • Lowe's ratio test threshold = 0.8
    Used in the baseline and proposed systems to prune correspondences before estimation.
  • Maximum iterations for Matlab RANSAC/LMedS/MSAC = 2000
    An implementation limit that can affect robust fitting quality and therefore the comparison with USAC and GC-RANSAC.
assumptions (4)
  • domain assumption COLMAP-based camera parameters for T&T and CPC provide accurate ground-truth projection matrices.
    Stated in Sec 4 (Ground Truth) with average reprojection error below one pixel; used to compute F, but relies on feature matching similar to the methods under evaluation.
  • domain assumption Image pairs selected by requiring more than 20 SIFT inliers are representative of typical matching tasks.
    Stated in Sec 4 (Image Pairs Construction); this restricts the test set to pairs SIFT can already match, potentially hiding SIFT failures.
  • standard math Zhang's symmetric geometric distance is a valid measure of fundamental matrix error.
    Sec 3.1 adopts SGD from [48] and normalizes it by image diagonal; the metric's validity is taken from prior work.
  • domain assumption Reference implementations of RANSAC, LMedS, and MSAC (Matlab) are comparable to the authors' implementations of USAC and GC-RANSAC.
    Sec 5.1 states Matlab functions with a 2000-iteration cap; differences in implementation quality or stopping criteria can change rankings.

how reviews work

0 comments
Cite this review

Pith. "Pith review of An Evaluation of Feature Matchers for Fundamental Matrix Estimation." pith.science (2026). https://pith.science/paper/46OJ6QSA

@misc{pith2026190809474,
  author       = {Pith},
  title        = {Pith review of: An Evaluation of Feature Matchers for Fundamental Matrix Estimation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/46OJ6QSA}},
  note         = {Machine review of arXiv:1908.09474}
}
read the original abstract

Matching two images while estimating their relative geometry is a key step in many computer vision applications. For decades, a well-established pipeline, consisting of SIFT, RANSAC, and 8-point algorithm, has been used for this task. Recently, many new approaches were proposed and shown to outperform previous alternatives on standard benchmarks, including the learned features, correspondence pruning algorithms, and robust estimators. However, whether it is beneficial to incorporate them into the classic pipeline is less-investigated. To this end, we are interested in i) evaluating the performance of these recent algorithms in the context of image matching and epipolar geometry estimation, and ii) leveraging them to design more practical registration systems. The experiments are conducted in four large-scale datasets using strictly defined evaluation metrics, and the promising results provide insight into which algorithms suit which scenarios. According to this, we propose three high-quality matching systems and a Coarse-to-Fine RANSAC estimator. They show remarkable performances and have potentials to a large part of computer vision tasks. To facilitate future research, the full evaluation pipeline and the proposed methods are made publicly available.

Figures

Figures reproduced from arXiv: 1908.09474 by the authors.

Figure 1
Figure 1. Outputs of the two-view matching and geometry estimation pipeline. (a) shows [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 47 canonical work pages

  1. [1]

    Building Rome in a day

    Sameer Agarwal, Yasutaka Furukawa, Noah Snavely, Ian Simon, Brian Curless, Steven M Seitz, and Richard Szeliski. Building Rome in a day. Communications of the ACM, 54(10):105–112, 2011

  2. [2]

    Three things everyone should know to improve object retrieval

    Relja Arandjelovic and Andrew Zisserman. Three things everyone should know to improve object retrieval. In IEEE Conference on Computer Vision and Pattern Recog- nition (CVPR), pages 2911–2918. IEEE, 2012

  3. [3]

    Overall view regarding fundamental matrix esti- mation

    Xavier Armangué and Joaquim Salvi. Overall view regarding fundamental matrix esti- mation. Image and Vision Computing, 21(2):205–220, 2003

  4. [4]

    Learning local feature descriptors with triplets and shallow convolutional neural networks

    Vassileios Balntas, Edgar Riba, Daniel Ponsa, and Krystian Mikolajczyk. Learning local feature descriptors with triplets and shallow convolutional neural networks. In British Machine Vision Conference (BMVC), page 3, 2016

  5. [5]

    HPatches: A benchmark and evaluation of handcrafted and learned local descriptors

    Vassileios Balntas, Karel Lenc, Andrea Vedaldi, and Krystian Mikolajczyk. HPatches: A benchmark and evaluation of handcrafted and learned local descriptors. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , pages 5173–5182. IEEE, 2017

  6. [6]

    Graph-Cut RANSAC

    Daniel Barath and Jiri Matas. Graph-Cut RANSAC. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018

  7. [7]

    GMS: Grid-based motion statistics for fast, ultra-robust feature correspondence

    JiaWang Bian, Wen-Yan Lin, Yasuyuki Matsushita, Sai-Kit Yeung, Tan Dat Nguyen, and Ming-Ming Cheng. GMS: Grid-based motion statistics for fast, ultra-robust feature correspondence. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 4181–4190. IEEE, 2017

  8. [8]

    Discriminative learning of local im- age descriptors

    Matthew Brown, Gang Hua, and Simon Winder. Discriminative learning of local im- age descriptors. IEEE Transactions on Pattern Recognition and Machine Intelligence (PAMI), pages 43–57, 2011

Show all 48 references
  1. [9]

    Kernel local descriptors with implicit rotation matching

    Andrei Bursuc, Giorgos Tolias, and Hervé Jégou. Kernel local descriptors with implicit rotation matching. In International Conference on Multimedia Retrieval , pages 595–

  2. [10]

    Performance evaluation of RANSAC family

    Sunglok Choi, Taemin Kim, and Wonpil Yu. Performance evaluation of RANSAC family. International Journal on Computer Vision (IJCV), 24(3):271–300, 1997

  3. [11]

    MonoSLAM: Real-time single camera slam

    Andrew J Davison, Ian D Reid, Nicholas D Molton, and Olivier Stasse. MonoSLAM: Real-time single camera slam. IEEE Transactions on Pattern Recognition and Machine Intelligence (PAMI), 29(6):1052–1067, 2007

  4. [12]

    Domain-size pooling in local descriptors: DSP- SIFT

    Jingming Dong and Stefano Soatto. Domain-size pooling in local descriptors: DSP- SIFT. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 5097–5106, 2015

  5. [13]

    Fundamental matrix estimation: A study of error criteria.Pattern Recognition Letters, 32(2):383–391, 2011

    Mohammed E Fathy, Ashraf S Hussein, and Mohammed F Tolba. Fundamental matrix estimation: A study of error criteria.Pattern Recognition Letters, 32(2):383–391, 2011. 12 BIAN ET AL.: AN EV ALUA TION OF FEA TURE MA TCHERS

  6. [14]

    Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography

    Martin A Fischler and Robert C Bolles. Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography. Commu- nications of the ACM, 24(6):381–395, 1981

  7. [15]

    SVO: Fast semi-direct monocular visual odometry

    Christian Forster, Matia Pizzoli, and Davide Scaramuzza. SVO: Fast semi-direct monocular visual odometry. In International Conference on Robotics and Automation (ICRA), pages 15–22. IEEE, 2014

  8. [16]

    Are we ready for autonomous driv- ing? the KITTI vision benchmark suite

    Andreas Geiger, Philip Lenz, and Raquel Urtasun. Are we ready for autonomous driv- ing? the KITTI vision benchmark suite. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 3354–3361. IEEE, 2012

  9. [17]

    Multiple view geometry in computer vision

    Richard Hartley and Andrew Zisserman. Multiple view geometry in computer vision . Cambridge university press, 2003

  10. [18]

    In defense of the eight-point algorithm

    Richard I Hartley. In defense of the eight-point algorithm. IEEE Transactions on Pattern Recognition and Machine Intelligence (PAMI), 19(6):580–593, 1997

  11. [19]

    Comparative evaluation of bi- nary features

    Jared Heinly, Enrique Dunn, and Jan-Michael Frahm. Comparative evaluation of bi- nary features. In European Conference on Computer Vision (ECCV), pages 759–773. Springer, 2012

  12. [20]

    Re- constructing the world* in six days*(as captured by the yahoo 100 million image dataset)

    Jared Heinly, Johannes L Schonberger, Enrique Dunn, and Jan-Michael Frahm. Re- constructing the world* in six days*(as captured by the yahoo 100 million image dataset). In IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , pages 3287–3295, 2015

  13. [21]

    Tanks and Tem- ples: Benchmarking large-scale scene reconstruction

    Arno Knapitsch, Jaesik Park, Qian-Yi Zhou, and Vladlen Koltun. Tanks and Tem- ples: Benchmarking large-scale scene reconstruction. ACM Transactions on Graphics (TOG), 36(4):78, 2017

  14. [22]

    An evaluation of the performance of RANSAC algorithms for stereo camera calibrarion

    AJ Lacey, N Pinitkarn, and Neil A Thacker. An evaluation of the performance of RANSAC algorithms for stereo camera calibrarion. In British Machine Vision Confer- ence (BMVC), pages 1–10, 2000

  15. [23]

    CODE: Coherence based decision boundaries for feature cor- respondence

    Wen-Yan Lin, Fan Wang, Ming-Ming Cheng, Sai-Kit Yeung, Philip HS Torr, Minh N Do, and Jiangbo Lu. CODE: Coherence based decision boundaries for feature cor- respondence. IEEE Transactions on Pattern Recognition and Machine Intelligence (PAMI), 2017

  16. [24]

    Distinctive image features from scale-invariant keypoints.International Journal on Computer Vision (IJCV), 60(2):91–110, 2004

    David G Lowe. Distinctive image features from scale-invariant keypoints.International Journal on Computer Vision (IJCV), 60(2):91–110, 2004

  17. [25]

    GeoDesc: Learning local descriptors by integrating geometry constraints

    Zixin Luo, Tianwei Shen, Lei Zhou, Siyu Zhu, Runze Zhang, Yao Yao, Tian Fang, and Long Quan. GeoDesc: Learning local descriptors by integrating geometry constraints. In European Conference on Computer Vision (ECCV), 2018

  18. [26]

    Locality preserving matching

    Jiayi Ma, Ji Zhao, Junjun Jiang, Huabing Zhou, and Xiaojie Guo. Locality preserving matching. International Journal on Computer Vision (IJCV), 2018

  19. [27]

    A performance evaluation of local descrip- tors

    Krystian Mikolajczyk and Cordelia Schmid. A performance evaluation of local descrip- tors. IEEE Transactions on Pattern Recognition and Machine Intelligence (PAMI), 27 (10):1615–1630, 2005. BIAN ET AL.: AN EV ALUA TION OF FEA TURE MA TCHERS 13

  20. [28]

    A comparison of affine region detectors

    Krystian Mikolajczyk, Tinne Tuytelaars, Cordelia Schmid, Andrew Zisserman, Jiri Matas, Frederik Schaffalitzky, Timor Kadir, and Luc Van Gool. A comparison of affine region detectors. International Journal on Computer Vision (IJCV) , 65(1-2):43–72, 2005

  21. [29]

    Working hard to know your neighbor’s margins: Local descriptor learning loss

    Anastasiia Mishchuk, Dmytro Mishkin, Filip Radenovic, and Jiri Matas. Working hard to know your neighbor’s margins: Local descriptor learning loss. InNeural Information Processing Systems (NIPS), pages 4826–4837, 2017

  22. [30]

    Repeatability is not enough: Learn- ing affine regions via discriminability

    Dmytro Mishkin, Filip Radenovic, and Jiri Matas. Repeatability is not enough: Learn- ing affine regions via discriminability. In European Conference on Computer Vision (ECCV). Springer, 2018

  23. [31]

    ORB-SLAM: a versatile and accurate monocular slam system

    Raul Mur-Artal, Jose Maria Martinez Montiel, and Juan D Tardos. ORB-SLAM: a versatile and accurate monocular slam system. IEEE Transactions on Robotics (TRO), 31(5):1147–1163, 2015

  24. [32]

    From dusk till dawn: Modeling in the dark

    Filip Radenovic, Johannes L Schönberger, Dinghuang Ji, Jan-Michael Frahm, Ondrej Chum, and Jiri Matas. From dusk till dawn: Modeling in the dark. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 5488–5496, 2016

  25. [33]

    A comparative analysis of RANSAC techniques leading to adaptive real-time random sample consensus

    Rahul Raguram, Jan-Michael Frahm, and Marc Pollefeys. A comparative analysis of RANSAC techniques leading to adaptive real-time random sample consensus. In European Conference on Computer Vision (ECCV), pages 500–513. Springer, 2008

  26. [34]

    USAC: a universal framework for random sample consensus

    Rahul Raguram, Ondrej Chum, Marc Pollefeys, Jiri Matas, and Jan-Michael Frahm. USAC: a universal framework for random sample consensus. IEEE Transactions on Pattern Recognition and Machine Intelligence (PAMI), 35(8):2022–2038, 2013

  27. [35]

    Deep fundamental matrix estimation

    René Ranftl and Vladlen Koltun. Deep fundamental matrix estimation. In European Conference on Computer Vision (ECCV), pages 284–299, 2018

  28. [36]

    Robust regression and outlier detection , volume 589

    Peter J Rousseeuw and Annick M Leroy. Robust regression and outlier detection , volume 589. John wiley & sons, 1987

  29. [37]

    Structure-from-motion revisited

    Johannes L Schönberger and Jan-Michael Frahm. Structure-from-motion revisited. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , pages 4104– 4113, 2016

  30. [38]

    From single image query to detailed 3d reconstruction

    Johannes L Schönberger, Filip Radenovic, Ondrej Chum, and Jan-Michael Frahm. From single image query to detailed 3d reconstruction. In IEEE Conference on Com- puter Vision and Pattern Recognition (CVPR), pages 5126–5134, 2015

  31. [39]

    Com- parative evaluation of hand-crafted and learned local features

    Johannes L Schönberger, Hans Hardmeier, Torsten Sattler, and Marc Pollefeys. Com- parative evaluation of hand-crafted and learned local features. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 6959–6968. IEEE, 2017

  32. [40]

    Sturm, N

    J. Sturm, N. Engelhard, F. Endres, W. Burgard, and D. Cremers. A benchmark for the evaluation of RGB-D SLAM systems. In IEEE International Conference on Intelligent Robots and Systems (IROS), Oct. 2012. 14 BIAN ET AL.: AN EV ALUA TION OF FEA TURE MA TCHERS

  33. [41]

    L2-Net: Deep learning of discriminative patch descriptor in euclidean space

    Yurun Tian, Bin Fan, Fuchao Wu, et al. L2-Net: Deep learning of discriminative patch descriptor in euclidean space. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), volume 1, page 6, 2017

  34. [42]

    The development and comparison of robust methods for estimating the fundamental matrix

    Philip HS Torr and David W Murray. The development and comparison of robust methods for estimating the fundamental matrix. International Journal on Computer Vision (IJCV), 24(3):271–300, 1997

  35. [43]

    MLESAC: A new robust estimator with ap- plication to estimating image geometry

    Philip HS Torr and Andrew Zisserman. MLESAC: A new robust estimator with ap- plication to estimating image geometry. Computer Vision and Image Understanding (CVIU), 78(1):138–156, 2000

  36. [44]

    Performance characterization of fundamental matrix estimation under image degradation

    Philip HS Torr and A Zissermann. Performance characterization of fundamental matrix estimation under image degradation. Machine Vision and Applications , 9(5-6):321– 333, 1997

  37. [45]

    VLFeat: An open and portable library of com- puter vision algorithms

    Andrea Vedaldi and Brian Fulkerson. VLFeat: An open and portable library of com- puter vision algorithms. In ACM International Conference on Multimedia (ACM MM), pages 1469–1472. ACM, 2010

  38. [46]

    Robust global translations with 1DSFM

    Kyle Wilson and Noah Snavely. Robust global translations with 1DSFM. In European Conference on Computer Vision (ECCV), pages 61–75. Springer, 2014

  39. [47]

    Learning to find good correspondences

    Kwang Moo Yi, Eduard Trulls, Yuki Ono, Vincent Lepetit, Mathieu Salzmann, and Pascal Fua. Learning to find good correspondences. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018

  40. [48]

    Determining the epipolar geometry and its uncertainty: A review

    Zhengyou Zhang. Determining the epipolar geometry and its uncertainty: A review. International Journal on Computer Vision (IJCV) , 27(2):161–195, Mar 1998. ISSN 1573-1405. doi: 10.1023/A:1007941100561

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.