Pith. sign in

REVIEW 4 major objections 4 minor 1 cited by

Learning Two-View Correspondences and Geometry Using Order-Aware Network

T0 review · 4 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read Order-Aware Network claims that learning a canonical-order clustering of sparse correspondences, followed by spatial filtering over the clusters, substantially improves two-view geometry estimation over prior learning-based outlier…

desk verdict Solid empirical paper with fair comparisons; the 'canonical order' mechanism is under-justified but the accuracy gains hold up. read the letter →

arxiv 1908.04964 v1 pith:VOGTFBXH submitted 2019-08-14 cs.CV cs.CGcs.LG

Jiahui Zhang , Dawei Sun , Zixin Luo , Anbang Yao , Lei Zhou , Tianwei Shen , Yurong Chen , Long Quan
show 1 more author
Hongen Liao
This is my paper · ORCID
classification cs.CVcs.CGcs.LG
keywords two-viewgeometryessentialmatrixoutlierrejectiondifferentiablepoolingcanonicalorderspatialcorrelationcorrespondencelearningrelativeposeestimation
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's central claim is that the performance of learning-based outlier rejection for two-view geometry is limited not only by per-correspondence feature processing but by the lack of stable relational structure among the correspondences. To fix this, the Order-Aware Network learns to pool the unordered set of putative correspondences into a fixed number of clusters via a permutation-invariant soft assignment, which yields the clusters in a canonical order. That order allows the network to apply spatially shared filters over the clusters to model global context, and an Order-Aware unpooling step returns the features to the original correspondence size. The authors report that the resulting model substantially improves both inlier/outlier classification and essential matrix regression on outdoor (YFCC100M) and indoor (SUN3D) datasets, with the strongest gains on unknown scenes without RANSAC post-processing.

What carries the argument

The engine of the method is the differentiable pooling (DiffPool) layer: it learns a soft assignment matrix $S_{\text{pool}} \in \mathbb{R}^{N \times M}$ that averages the $N$ input correspondences into $M$ clusters. Because both the assignment network and the averaging are permutation-equivariant, the pooled representation is unchanged by any reordering of the input, and the clusters therefore come out in a fixed canonical order. That order is what makes the second piece usable: an Order-Aware Filtering block applies a shared perceptron across the spatial (cluster) dimension to model global context, which would be meaningless if the clusters changed identity between inputs. The third piece, the Order-Aware DiffUnpool layer, learns the unpooling assignment from the fine-level features rather than from the pooled features, which keeps the upsampled features aligned with the original correspondences so per-correspondence predictions can be made.

What would settle it

Run the learned model on a test set but with a random permutation applied to the cluster indices after pooling and before the Order-Aware Filtering block; if accuracy does not drop, the ordering is not actually doing the work. Or compute, over many image pairs, the spatial distribution of the top-response correspondences for each cluster index and measure the overlap; large variance across pairs would indicate that the shared spatial filter is operating on semantically arbitrary groupings.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that sparse, unordered correspondences can be organized hierarchically by a learned differentiable pooling layer, and that the canonical order thereby induced enables a spatial correlation operation that captures global context better than PointNet-style context normalization. The network jointly predicts inlier probabilities and regresses the essential matrix through a differentiable weighted eight-point algorithm. In the paper's experiments, the full model (with a geometry loss and an iterative refinement stage) raises mAP@5° without RANSAC by 15.78 percentage points on unknown outdoor scenes and by 7.03 percentage points on unknown indoor scenes relative to the PointCN baseline, and it retains gains when RANSAC is applied as post-processing.

Load-bearing premise

The canonical order of clusters is fixed by the network's parameters, not by any guarantee that cluster i always corresponds to the same motion or image region; the paper's only support for this semantic consistency is a qualitative visualization of average cluster responses across a few pairs.

Editorial extensions

If this is right

  • On both outdoor and indoor unknown scenes, the full model improves mAP@5° without RANSAC by about 16 and 7 percentage points over the PointCN baseline, and the gain remains when RANSAC is applied as post-processing.
  • The ablation study isolates the contribution of each new component: the Order-Aware DiffUnpool alone gives a 5.23-point gain on unknown outdoor scenes, and the Order-Aware Filtering block adds another 3.77 points, so both are load-bearing.
  • The method works with both SIFT and learned (SuperPoint) keypoints, and with learned features it improves over RANSAC alone on both indoor and outdoor scenes.
  • Because the pooling reduces 2000 correspondences to 500 clusters, the architecture can accept variable-length inputs, which is useful for real images with more or fewer than 2000 keypoints.

Reading between the lines

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

  • The paper's own evidence for the semantic consistency of the canonical cluster order is only qualitative (Fig. 7); a quantitative test, e.g., measuring the entropy of top-response image regions per cluster over many scenes, would tell whether the shared spatial filter is genuinely order-meaningful.
  • The same architectural recipe—permutation-invariant pooling to a canonical order, spatial filtering over the pooled units, then order-aware unpooling—could generalize to other unordered-set prediction tasks such as multi-view correspondences, point cloud registration, or even set-based graph classification.
  • The observation that training on more sequences (68 vs. 2) sharply improves generalization suggests that data scale, as much as architecture, accounts for part of the reported gains; a controlled study with matched training data would separate the two effects.
  • If cluster order consistency turns out to be fragile, an attention mechanism over clusters could replace the fixed spatial filter, potentially preserving the global-context benefit without requiring a stable cluster semantics.
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

4 major / 4 minor

Summary. The paper proposes Order-Aware Network (OANet) for outlier rejection and two-view geometry estimation from putative correspondences. The network is built on PointCN blocks and adds three operations: a Differentiable Pooling (DiffPool) layer that clusters unordered correspondences via a learned soft assignment matrix, an Order-Aware Differentiable Unpooling (DiffUnpool) layer that maps the coarsened features back to the original correspondence size, and an Order-Aware Filtering block containing a Spatial Correlation layer applied along the cluster dimension. The network predicts per-correspondence inlier probabilities and regresses the essential matrix through a differentiable weighted eight-point algorithm. Experiments on YFCC100M and SUN3D, with SIFT and SuperPoint features, report consistent improvements over PointCN, PointNet++, N3Net, and DFE, with ablation studies isolating each proposed component.

Significance. If the reported results hold, the hierarchical clustering view of sparse correspondences is a useful and nontrivial contribution to learning-based geometric matching. The paper has clear strengths: all baselines were retrained on the same data, the ablations in Table 2 isolate each proposed operation, the closed-form differentiable essential-matrix regression is a clean construction, and the method is evaluated on both indoor and outdoor datasets. The main weakness is that the paper's central mechanism, the 'canonical order' of clusters that justifies the Order-Aware Filtering block, is only qualitatively supported and not quantitatively verified. This, together with missing error bars and an internal numerical inconsistency in the headline claim, means the paper needs revision before its claims are fully supported.

major comments (4)
  1. [3.2 (Eqs. 10-11) and 3.4] The 'canonical order' claim is not established by the permutation-invariance proof. Equations (10)-(11) show that for a fixed input, permuting the input rows permutes the assignment matrix and leaves the pooled output unchanged. This does not imply that cluster index i contains correspondences with the same semantic meaning (e.g., the same image region or motion) across different image pairs. The Spatial Correlation layer in Section 3.4 applies a shared filter along the cluster dimension, so its benefit relies on cross-instance consistency of the cluster ordering. The only evidence offered is the qualitative statement in Section 4.6 that motions are 'roughly consistent' (Fig. 7), with no quantitative measure. This is load-bearing because the method's name and main architectural novelty depend on order-awareness. Please add a quantitative analysis, for example by measuring cluster-assignment consistency across image pairs, or an experiment that randomly permutes the cluster order at inference time and reports the change in mAP, or soften the claim and reframe the Spatial Correlation layer as operating on a deterministic but not necessarily semantically consistent ordering.
  2. [4.5] The sentence 'showing improvements of 15.78% and 7.03% over PointCN on both outdoor and indoor unknown scenes without RANSAC' does not match Table 3. The values 15.78 and 7.03 are the absolute percentage-point differences in the with-RANSAC columns for unknown scenes (Ours++ 39.33 vs. PointCN 23.55 outdoors; Ours++ 16.39 vs. PointCN 9.36 indoors). The without-RANSAC differences are 4.20 and 1.52 percentage points, respectively. Please correct the sentence to refer to the correct columns and specify whether the reported improvements are absolute percentage points or relative percentages.
  3. [4.4 and 4.5] No error bars or significance tests are reported for any of the main comparisons. Table 2 and Table 3 appear to report single runs, and Section 4.5 states that N3Net was run three times with the best result used. Since several of the claimed gains are small in absolute terms (e.g., 4.20 percentage points over PointCN without RANSAC on outdoor unknown scenes), the improvements could be within run-to-run variability. Please report the mean and standard deviation over at least three seeds for the main methods, or otherwise justify that the reported differences are not noise.
  4. [4.2 and Table 1] The paper changes the RANSAC inlier threshold in findEssentialMat from 0.01 to 0.001 after observing that it improves results. Please state explicitly whether this threshold is used for all methods in Table 3 and how it was selected. If it was chosen by looking at test-set performance, the comparisons are not a purely fair evaluation; if it was chosen on a validation split and applied uniformly, that should be stated. This is important because the with-RANSAC numbers are a central part of the comparison table.
minor comments (4)
  1. [Eqs. (14)-(15) and Fig. 3b] The notation in the Order-Aware DiffUnpool layer is inconsistent. The text says S_unpool is in R^{N x M} and that softmax is applied along the column dimension, but the interpretation of Eq. (15) as a weighted average of clusters for each node requires row normalization of the N x M matrix. Please align the matrix dimensions, the softmax direction, and the accompanying explanation.
  2. [Table 1] The caption says 'Results with/without RANSAC under error thresholds of 5°, 10° and 20°' but the table structure shows mAP5/10/20 as separate columns, with the with/without RANSAC values inside each cell. Clarify the layout and explicitly label which value corresponds to RANSAC post-processing.
  3. [3.1] The weakly supervised labels s in Eq. (5) are generated using the same epipolar-distance threshold (10^-4) that is used to evaluate the essential-matrix regression. This is a standard pseudo-labeling setup and is not circular because the labels target the classification term only, but the paper should state this explicitly to avoid misunderstanding.
  4. [4.3 and 3.3] The paper says the input typically has N=2000 correspondences but also notes that the method handles variable sizes. Please clarify how inputs of different sizes are batched and whether the DiffPool layer always pools to a fixed number of clusters in that case.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's empirical comparisons are self-contained and re-trained under identical settings; the shared epipolar-threshold label/evaluation convention is standard practice and does not force the central pose-estimation result.

full rationale

Order-Aware Network is an empirical architecture paper; there is no derivation chain in which a predicted quantity is defined from, or fitted to, the quantity it claims to predict. The only formal derivation is the permutation-invariance proof of the DiffPool layer (Eqs. 9-11), which follows directly from the permutation-equivariance of hpool and softmax and does not assume its conclusion. The subsequent claim that pooled clusters appear in a 'canonical order' is a definitional consequence for a fixed input with respect to the learned parameters; the additional claim that the same cluster index is semantically consistent across image pairs is supported only by the qualitative Fig. 7, which is a correctness/robustness concern rather than a circular one. Self-citations ([18,19,43,45]) appear only in related work or as background and are not load-bearing; DiffPool is cited to external work [42], and all baselines ([21,26,30]) are either external or re-implemented and re-trained under the same data and protocol. The training labels for the inlier classifier use a 10^-4 symmetric-epipolar-distance threshold, and Tab. 4 reports precision/recall at the same threshold; while this makes the classification metrics somewhat self-referential to the label generator, all compared methods use the same labels, and the paper's central mAP/pose metric is based on angular error against ground-truth camera poses, which is independent. No fitted parameter is renamed as a prediction, and no load-bearing conclusion reduces to a self-citation. Score 0.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

The empirical gains rest on hand-set hyperparameters (cluster count, label threshold) and assumptions about the DiffPool assignment behavior. No new physical entities are introduced.

free parameters (4)
  • Number of clusters M = 500
    After DiffPool, the number of nodes is reduced to a fixed 500, which the paper says 'gives best performance' (Section 4.3). This is a hand-chosen architectural hyperparameter.
  • Inlier label threshold = 1e-4
    Weakly supervised inlier/outlier labels are generated using a symmetric epipolar distance threshold of 1e-4 (Section 3.1). Changing this threshold would change the training labels and thus the learned model.
  • RANSAC inlier threshold = 0.001
    The paper changes OpenCV's findEssentialMat inlier threshold from 0.01 to 0.001 and reports that this improves RANSAC results (Section 4.2). This is a tuned evaluation hyperparameter.
  • Geometry loss clamp = 0.1
    Clamping the geometry losses to 0.1 'works best in our case' (Section 4.4). A hand-set value for the loss.
assumptions (3)
  • standard math The weighted eight-point algorithm's eigenvector solution for the essential matrix is differentiable with respect to correspondence weights.
    The paper relies on this to train the essential matrix regression term end-to-end (Section A.1, citing [11,21]).
  • domain assumption Permutation invariance of DiffPool implies a canonical order of clusters.
    The paper argues clusters are in a canonical, input-order-independent arrangement (Section 3.2). The property proven is invariance of the pooled features, not cross-input semantic consistency of cluster indices; the latter is an assumption used by the Spatial Correlation layer.
  • ad hoc to paper Softmax assignments can be used as soft clustering without additional cluster-regularization losses.
    Original DiffPool typically adds auxiliary link-prediction and entropy losses to train the assignment matrix (Ying et al., 2018). This paper uses a plain softmax assignment without such losses, an implicit modeling choice not discussed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning Two-View Correspondences and Geometry Using Order-Aware Network." pith.science (2026). https://pith.science/paper/VOGTFBXH

@misc{pith2026190804964,
  author       = {Pith},
  title        = {Pith review of: Learning Two-View Correspondences and Geometry Using Order-Aware Network},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VOGTFBXH}},
  note         = {Machine review of arXiv:1908.04964}
}
read the original abstract

Establishing correspondences between two images requires both local and global spatial context. Given putative correspondences of feature points in two views, in this paper, we propose Order-Aware Network, which infers the probabilities of correspondences being inliers and regresses the relative pose encoded by the essential matrix. Specifically, this proposed network is built hierarchically and comprises three novel operations. First, to capture the local context of sparse correspondences, the network clusters unordered input correspondences by learning a soft assignment matrix. These clusters are in a canonical order and invariant to input permutations. Next, the clusters are spatially correlated to form the global context of correspondences. After that, the context-encoded clusters are recovered back to the original size through a proposed upsampling operator. We intensively experiment on both outdoor and indoor datasets. The accuracy of the two-view geometry and correspondences are significantly improved over the state-of-the-arts. Code will be available at https://github.com/zjhthu/OANet.git.

Figures

Figures reproduced from arXiv: 1908.04964 by the authors.

Figure 1
Figure 1. The Order-Aware Network to learn two-view correspon [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. Designs of Differentiable Unpooling layer. (a) Plain Dif [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Order-Aware Filtering block. We insert the Spatial Cor [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figures from the paper (3 more)
Figure 5
Figure 5. Figure 5: Matching results using RANSAC (top), PointCN [ [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: DiffUnpool layer visualization. Top 15 responses in dif [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: DiffUnpool layer visualization. Top 20 responses in the [PITH_FULL_IMAGE:figures/full_fig_p008_7.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. TurboReg: TurboClique for Robust and Efficient Point Cloud Registration

    cs.CV 2025-07 conditional novelty 6.0 of 10

    TurboReg estimates 3D transformations from fixed-size three-match cliques in a tightly constrained compatibility graph, found by a linear-time pivot-guided search, matching or beating maximal clique methods at far hig...

Reference graph

Works this paper leans on

47 extracted references · 45 canonical work pages · cited by 1 Pith paper

  1. [1]

    Gms: Grid-based motion statistics for fast, ultra-robust feature cor- respondence

    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 cor- respondence. In Computer Vision and Pattern Recognition (CVPR), 2017. 1, 2

    work page 2017

  2. [2]

    Dsac differentiable ransac for camera localization

    Eric Brachmann, Alexander Krull, Sebastian Nowozin, Jamie Shotton, Frank Michel, Stefan Gumhold, and Carsten Rother. Dsac differentiable ransac for camera localization. In Computer Vision and Pattern Recognition (CVPR), 2017. 2

    work page 2017

  3. [3]

    Self-Improving Visual Odometry

    Daniel DeTone, Tomasz Malisiewicz, and Andrew Rabi- novich. Self-improving visual odometry. arXiv preprint arXiv:1812.03245, 2018. 2

    work page Pith review arXiv 2018

  4. [4]

    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 Computer Vision and Pattern Recognition Workshops (CVPRW), 2018. 1, 2, 3, 8

    work page 2018

  5. [5]

    Splinecnn: Fast geometric deep learning with continuous b-spline kernels

    Matthias Fey, Jan Eric Lenssen, Frank Weichert, and Hein- rich M ¨uller. Splinecnn: Fast geometric deep learning with continuous b-spline kernels. In Computer Vision and Pattern Recognition (CVPR), 2018. 3

    work page 2018

  6. [6]

    Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography.Communications of the ACM, 1981

    Martin A Fischler and Robert C Bolles. Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography.Communications of the ACM, 1981. 2

    work page 1981

  7. [7]

    3d semantic segmentation with submanifold sparse convolutional networks

    Benjamin Graham, Martin Engelcke, and Laurens van der Maaten. 3d semantic segmentation with submanifold sparse convolutional networks. In Computer Vision and Pattern Recognition (CVPR), 2018. 3

    work page 2018

  8. [8]

    Inductive representation learning on large graphs

    Will Hamilton, Zhitao Ying, and Jure Leskovec. Inductive representation learning on large graphs. In Advances in Neu- ral Information Processing Systems (NIPS), 2017. 3

    work page 2017

Show all 47 references
  1. [9]

    Multiple view ge- ometry in computer vision

    Richard Hartley and Andrew Zisserman. Multiple view ge- ometry in computer vision . Cambridge university press,

  2. [10]

    Reconstructing 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. Reconstructing the world* in six days*(as captured by the yahoo 100 million image dataset). In Computer Vision and Pattern Recognition (CVPR), 2015. 6

    2015

  3. [11]

    Matrix backpropagation for deep networks with structured layers

    Catalin Ionescu, Orestis Vantzos, and Cristian Sminchisescu. Matrix backpropagation for deep networks with structured layers. In International Conference on Computer Vision (ICCV), 2015. 11

    2015

  4. [12]

    Semi-supervised classifi- cation with graph convolutional networks

    Thomas N Kipf and Max Welling. Semi-supervised classifi- cation with graph convolutional networks. 2017. 3

    2017

  5. [13]

    Escape from cells: Deep kd-networks for the recognition of 3d point cloud mod- els

    Roman Klokov and Victor Lempitsky. Escape from cells: Deep kd-networks for the recognition of 3d point cloud mod- els. In International Conference on Computer Vision (ICCV),

  6. [14]

    Undeepvo: Monocular visual odometry through unsuper- vised deep learning

    Ruihao Li, Sen Wang, Zhiqiang Long, and Dongbing Gu. Undeepvo: Monocular visual odometry through unsuper- vised deep learning. InInternational Conference on Robotics and Automation (ICRA), 2018. 2

    2018

  7. [15]

    Bilateral func- tions for global motion modeling

    Wen-Yan Daniel Lin, Ming-Ming Cheng, Jiangbo Lu, Hong- sheng Yang, Minh N Do, and Philip Torr. Bilateral func- tions for global motion modeling. In European Conference on Computer Vision (ECCV), 2014. 1, 2, 3

    2014

  8. [16]

    A computer algorithm for reconstructing a scene from two projections

    H Christopher Longuet-Higgins. A computer algorithm for reconstructing a scene from two projections. Nature, 1981. 3

    1981

  9. [17]

    Distinctive image features from scale- invariant keypoints

    David G Lowe. Distinctive image features from scale- invariant keypoints. International Journal of Computer Vi- sion (IJCV), 2004. 2, 3

    2004

  10. [18]

    Contextdesc: Lo- cal descriptor augmentation with cross-modality context

    Zixin Luo, Tianwei Shen, Lei Zhou, Jiahui Zhang, Yao Yao, Shiwei Li, Tian Fang, and Long Quan. Contextdesc: Lo- cal descriptor augmentation with cross-modality context. In Computer Vision and Pattern Recognition (CVPR), 2019. 2

    2019

  11. [19]

    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 Eu- ropean Conference on Computer Vision (ECCV), 2018. 2

    2018

  12. [20]

    Geometric deep learning on graphs and manifolds using mixture model cnns

    Federico Monti, Davide Boscaini, Jonathan Masci, Emanuele Rodola, Jan Svoboda, and Michael M Bronstein. Geometric deep learning on graphs and manifolds using mixture model cnns. In Computer Vision and Pattern Recognition (CVPR), 2017. 3

    2017

  13. [21]

    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 Computer Vision and Pattern Recogni- tion (CVPR), 2018. 1, 2, 3, 4, 6, 7, 8, 11

    2018

  14. [22]

    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, 2015. 1

    2015

  15. [23]

    Learning convolutional neural networks for graphs

    Mathias Niepert, Mohamed Ahmed, and Konstantin Kutzkov. Learning convolutional neural networks for graphs. In International Conference on Machine Learning (ICML) ,

  16. [24]

    Lf-net: learning local features from images

    Yuki Ono, Eduard Trulls, Pascal Fua, and Kwang Moo Yi. Lf-net: learning local features from images. In Advances in Neural Information Processing Systems (NIPS), 2018. 2

    2018

  17. [25]

    Automatic differentiation in pytorch

    Adam Paszke, Sam Gross, Soumith Chintala, Gregory Chanan, Edward Yang, Zachary DeVito, Zeming Lin, Al- ban Desmaison, Luca Antiga, and Adam Lerer. Automatic differentiation in pytorch. In NIPS-W, 2017. 6

    2017

  18. [26]

    Neural nearest neighbors net- works

    Tobias Pl ¨otz and Stefan Roth. Neural nearest neighbors net- works. In Advances in Neural Information Processing Sys- tems (NIPS), 2018. 2, 7, 8

    2018

  19. [27]

    Pointnet: Deep learning on point sets for 3d classification and segmentation

    Charles R Qi, Hao Su, Kaichun Mo, and Leonidas J Guibas. Pointnet: Deep learning on point sets for 3d classification and segmentation. In Computer Vision and Pattern Recogni- tion (CVPR), 2017. 1, 3, 7, 8

    2017

  20. [28]

    Pointnet++: Deep hierarchical feature learning on point sets in a metric space

    Charles Ruizhongtai Qi, Li Yi, Hao Su, and Leonidas J Guibas. Pointnet++: Deep hierarchical feature learning on point sets in a metric space. In Advances in Neural Informa- tion Processing Systems (NIPS), 2017. 1

    2017

  21. [29]

    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 Trans. Pattern Anal. Mach. Intell. (PAMI), 2013. 2

    2013

  22. [30]

    Deep fundamental matrix estimation

    Ren ´e Ranftl and Vladlen Koltun. Deep fundamental matrix estimation. In European Conference on Computer Vision (ECCV), 2018. 1, 2, 3, 6, 7, 8

    2018

  23. [31]

    Convo- lutional neural network architecture for geometric matching

    Ignacio Rocco, Relja Arandjelovic, and Josef Sivic. Convo- lutional neural network architecture for geometric matching. In Computer Vision and Pattern Recognition (CVPR), 2017. 2

    2017

  24. [32]

    U- net: Convolutional networks for biomedical image segmen- tation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U- net: Convolutional networks for biomedical image segmen- tation. In International Conference on Medical image com- puting and computer-assisted intervention, 2015. 7

    2015

  25. [33]

    Orb: An efficient alternative to sift or surf

    Ethan Rublee, Vincent Rabaud, Kurt Konolige, and Gary Bradski. Orb: An efficient alternative to sift or surf. In In- ternational Conference on Computer Vision (ICCV) , 2011. 3

    2011

  26. [34]

    Structure- from-motion revisited

    Johannes L Schonberger and Jan-Michael Frahm. Structure- from-motion revisited. In Computer Vision and Pattern Recognition (CVPR), 2016. 1

    2016

  27. [35]

    Yfcc100m: the new data in multimedia research

    Bart Thomee, David A Shamma, Gerald Friedland, Ben- jamin Elizalde, Karl Ni, Douglas Poland, Damian Borth, and Li-Jia Li. Yfcc100m: the new data in multimedia research. Communications of the ACM, 2016. 5, 6

    2016

  28. [36]

    Lempitsky

    Dmitry Ulyanov, Andrea Vedaldi, and Victor S. Lempitsky. Instance normalization: The missing ingredient for fast styl- ization. CoRR, 2016. 1

    2016

  29. [37]

    Demon: Depth and motion network for learning monocular stereo

    Benjamin Ummenhofer, Huizhong Zhou, Jonas Uhrig, Niko- laus Mayer, Eddy Ilg, Alexey Dosovitskiy, and Thomas Brox. Demon: Depth and motion network for learning monocular stereo. In Computer Vision and Pattern Recog- nition (CVPR), 2017. 2, 6

    2017

  30. [38]

    Visualsfm: A visual structure from motion system

    Changchang Wu et al. Visualsfm: A visual structure from motion system. 2011. 1

    2011

  31. [39]

    Sun3d: A database of big spaces reconstructed using sfm and object labels

    Jianxiong Xiao, Andrew Owens, and Antonio Torralba. Sun3d: A database of big spaces reconstructed using sfm and object labels. In Computer Vision and Pattern Recognition (CVPR), 2013. 5, 6

    2013

  32. [40]

    Spidercnn: Deep learning on point sets with parameterized convolutional filters

    Yifan Xu, Tianqi Fan, Mingye Xu, Long Zeng, and Yu Qiao. Spidercnn: Deep learning on point sets with parameterized convolutional filters. In European Conference on Computer Vision (ECCV), 2018. 3

    2018

  33. [41]

    Lift: Learned invariant feature transform

    Kwang Moo Yi, Eduard Trulls, Vincent Lepetit, and Pascal Fua. Lift: Learned invariant feature transform. In European Conference on Computer Vision (ECCV), 2016. 1, 2, 3

    2016

  34. [42]

    Hierarchical graph rep- resentation learning with differentiable pooling

    Zhitao Ying, Jiaxuan You, Christopher Morris, Xiang Ren, Will Hamilton, and Jure Leskovec. Hierarchical graph rep- resentation learning with differentiable pooling. InAdvances in Neural Information Processing Systems (NIPS) , 2018. 2, 3, 4

    2018

  35. [43]

    Efficient semantic scene comple- tion network with spatial group convolution

    Jiahui Zhang, Hao Zhao, Anbang Yao, Yurong Chen, Li Zhang, and Hongen Liao. Efficient semantic scene comple- tion network with spatial group convolution. In European Conference on Computer Vision (ECCV), 2018. 3

    2018

  36. [44]

    An end-to-end deep learning architecture for graph classification

    Muhan Zhang, Zhicheng Cui, Marion Neumann, and Yixin Chen. An end-to-end deep learning architecture for graph classification. In Thirty-Second AAAI Conference on Artifi- cial Intelligence, 2018. 2

    2018

  37. [45]

    Learn- ing and matching multi-view descriptors for registration of point clouds

    Lei Zhou, Siyu Zhu, Zixin Luo, Tianwei Shen, Runze Zhang, Mingmin Zhen, Tian Fang, and Long Quan. Learn- ing and matching multi-view descriptors for registration of point clouds. In European Conference on Computer Vision (ECCV), 2018. 1, 2, 3

    2018

  38. [46]

    Unsupervised learning of depth and ego-motion from video

    Tinghui Zhou, Matthew Brown, Noah Snavely, and David G Lowe. Unsupervised learning of depth and ego-motion from video. In Computer Vision and Pattern Recognition (CVPR),

  39. [2017]

    Supplementary appendix A.1 Weighted Eight-Point Algorithm Here we provide a detailed description of the weighted eight-point algorithm [21]

    2 A. Supplementary appendix A.1 Weighted Eight-Point Algorithm Here we provide a detailed description of the weighted eight-point algorithm [21]. Given N correspondences ci = ( xi 1,yi 1,xi 2,yi 2), 1 ≤ i ≤ N, we can construct a matrix X ∈ RN ×9, where each row has the form of...

Pith tools

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