Pith. sign in

REVIEW 3 major objections 6 minor 76 references

MATCHA:Towards Matching Anything

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read MATCHA claims that a single dense feature descriptor, built by fusing diffusion-model and DINOv2 features under correspondence supervision, reaches state-of-the-art accuracy on geometric, semantic, and temporal matching at once.

desk verdict Temporal protocol under-specification makes the 'single feature SOTA across all three' claim premature, but the fusion idea and held-out gains are solid enough to warrant serious review. read the letter →

arxiv 2501.14945 v1 pith:ELYAVGF3 submitted 2025-01-24 cs.CV

Fei Xue , Sven Elflein , Laura Leal-Taixé , Qunjie Zhou This is my paper · ORCID
classification cs.CV
keywords unifiedfeaturecorrespondencegeometricmatchingsemantictemporaldiffusionmodelDINOv2pointtracking
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

MATCHA tries to establish that one dense feature descriptor can serve all three classic correspondence problems—geometric (same 3D point across views), semantic (same object part across different instances), and temporal (same point across video frames)—and can do so better than task-specific features. Starting from the diffusion-model features of DIFT and the object-level features of DINOv2, MATCHA learns an attention-based fusion that lets geometric and semantic descriptors exchange information under correspondence supervision, then concatenates the enhanced descriptors with DINOv2 features into a single output feature $F_m$. On HPatches and pose-estimation benchmarks the fused feature improves on DIFT; on SPair-71k, PF-Pascal, and PF-Willow it closes the gap to fully supervised semantic matchers; and on TAP-Vid it improves zero-shot temporal matching. If the claim holds, a single descriptor could replace per-task feature extractors in matching pipelines.

What carries the argument

The load-bearing mechanism is a transformer fusion module of $k$ self- and cross-attention blocks acting on the DIFT semantic feature $F_h$ and geometric feature $F_l$, supervised separately by a dual-softmax geometric loss and a contrastive-plus-flow semantic loss. The fused outputs $F_s$ and $F_g$ are then concatenated with the frozen DINOv2 descriptor $F_d$ (with channel strides) to form $F_m$, so the fusion stage is the only place where correspondence supervision enters. The design works by having semantic and geometric branches borrow information from each other during training, which the ablations claim makes the two descriptors more cooperative at the later concatenation step.

What would settle it

A concrete falsifier would be a benchmark suite with three tasks where task-specific features (e.g., DISK for geometric, SD4Match for semantic, DINOv2 for temporal) each beat MATCHA on their respective task and MATCHA does not beat them on any; the claim of a single feature reaching state of the art would then be false.

Watch

Extended reading notes

Core claim

The paper's central claim is that a single feature descriptor $F_m$, produced by supervised attention-based fusion of DIFT geometric and semantic features followed by channel-wise concatenation with DINOv2 features, achieves state-of-the-art results across geometric, semantic, and temporal matching—claimed as the first single feature to do so. The authors supervise only the intermediate semantic and geometric descriptors $F_s$ and $F_g$ with geometric and semantic losses, and leave $F_m$ itself unsupervised; the final descriptor is the concatenation of $F_g$, $F_s$, and $F_d$ (with channel strides). In their comparisons, MATCHA outperforms task-specific geometric features (DISK, XFeat, MASt3R.E) and unsupervised foundation features (DIFT, DINOv2) on the three tasks, with the largest margins on semantic and temporal matching.

Load-bearing premise

MATCHA's unified feature is never directly supervised; instead only the intermediate fused semantic and geometric descriptors receive supervision, and the final concatenation is expected to behave well on all tasks.

Editorial extensions

If this is right

  • A single $F_m$ descriptor can replace task-specific feature extractors in pipelines that use nearest-neighbor matching, covering geometric verification, semantic matching, and point tracking.
  • Diffusion-model features that already show emergent correspondences can be pushed past fully unsupervised methods by adding correspondence supervision only to intermediate fused branches.
  • DINOv2's object-level representation carries complementary knowledge for temporal matching, and combining it with supervised diffusion features yields the largest gains on TAP-Vid.
  • The re-purposed TAP-Vid protocol provides a common temporal benchmark for feature models, making it possible to compare geometric, semantic, and unified descriptors under identical conditions.

Reading between the lines

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

  • If MATCHA's claim generalizes, a practical consequence the paper does not spell out is that multi-task systems could cache one descriptor per image and serve geometric, semantic, and temporal matching needs from the same representation, simplifying memory and latency budgets.
  • The ablation's attribution of the joint-supervision failure to data imbalance suggests that a future version trained on a larger, better-balanced semantic-geometric corpus could make direct supervision of the unified feature viable; this is a testable extension the authors leave open.
  • Because the paper only evaluates with nearest-neighbor search, reusing $F_m$ inside dense matchers with cost aggregation or iterative refinement is a natural next test; the descriptor's role might grow where its 8x-downsampled geometry is too coarse.
  • The re-purposed TAP-Vid protocol could become a standard probe for generalization of feature models, but its single-object bias means good scores there may overstate performance on scenes with repetitive structure; future benchmarks should include such scenes.
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 / 6 minor

Summary. The paper proposes MATCHA, a single feature descriptor intended for geometric, semantic, and temporal correspondence matching. It builds on stable diffusion DIFT features (semantic Fh and geometric Fl) and DINOv2 features Fd. A transformer-based fusion module with self- and cross-attention augments Fh and Fl using correspondence supervision (dual-softmax for geometric, contrastive plus flow loss for semantic), yielding enhanced features Fs and Fg. These are concatenated with a strided subsample of DINOv2 features to form the final unified descriptor Fm. The paper evaluates MATCHA on semantic matching (SPair-71k, PF-Pascal, PF-Willow), geometric matching (HPatches, MegaDepth, ScanNet, Aachen for pose estimation), and zero-shot temporal matching on a re-purposed TAP-Vid protocol, and claims state-of-the-art performance across all three tasks with a single descriptor.

Significance. If the results hold, MATCHA provides a single dense descriptor that performs well across geometric, semantic, and temporal matching, potentially replacing task-specific feature extractors in several pipelines. The strength of the paper is its extensive empirical evaluation: ablations (Tab. 3, Tab. 6) show the contribution of each component, and the main evaluations on PF-Willow, Aachen, and TAP-Vid are held-out, reducing circularity concerns. The paper is also transparent about its limitations (e.g., resolution precision and runtime). The central value is the demonstration that limited correspondence supervision can be injected into foundation-model features without destroying their generalization, and that a carefully designed fusion makes otherwise incompatible features complementary.

major comments (3)
  1. [Sec. 4.3, Tab. 4] The temporal matching protocol is underspecified, and the temporal results are load-bearing for the central claim. The paper does not state how query points that are occluded or out-of-frame in the target frame are handled, how the PCK thresholds are normalized (e.g., relative to image dimensions vs. object scale), or whether mutual nearest-neighbor search is used. TAP-Vid's standard evaluation metrics (e.g., average position error, occlusion accuracy, average Jaccard) are not reported, and no comparison with the official TAP-Vid evaluation is provided. Because MATCHA's margin over DIFT.Uni+DINOv2 is only 1.8 points on average PCK (92.3 vs 90.5), small protocol choices could alter the ranking, and with it the claim of being the first single feature to achieve state-of-the-art across all three tasks. Please specify the complete protocol and report the standard TAP-Vid metrics, or restrict the temporal claim accordingly.
  2. [Abstract, Sec. 4.5, Tabs. 1 and 4] The claim that MATCHA 'consistently surpasses state-of-the-art methods across geometric, semantic, and temporal matching tasks' and 'outperforms all other methods across all tasks' is not supported by the reported numbers. In Table 1, supervised GeoASM outperforms MATCHA on both SPair-71k (PCK@0.1: 85.6 vs 79.6) and PF-Pascal (98.0 vs 96.8); MATCHA is best only on PF-Willow, where GeoASM is not reported. In Figure 4, DISK and R2D2 are more accurate than MATCHA on HPatches at error thresholds below about 7px. The claims should be qualified to the setting of feature-based methods using nearest-neighbor matching with downsampled features, or to the average score over the three tasks, and the abstract should be revised to avoid the universal 'consistently surpasses' wording.
  3. [Sec. 3.3, Eqs. (5)-(6), Tab. 6] The final descriptor Fm is never directly supervised; only the intermediate descriptors Fs and Fg receive correspondence supervision, and Fm is obtained by channel-wise concatenation with a frozen DINOv2 feature. The paper's own supplementary ablation (Table 6) shows that adding joint supervision to the unified feature (MATCHA-Light.Uni.S) sharply degrades semantic matching (PF-Willow average PCK drops from 78.0 to 50.7), which the authors attribute to data imbalance. The central claim of a 'single unified feature' therefore rests on the assumption that indirect supervision of Fs and Fg transfers to Fm through concatenation. While the final evaluations provide some evidence for this, the paper does not analyze the scale compatibility of the concatenated features or test alternative lightweight alignment mechanisms. I request a more thorough discussion of why concatenation is an appropriate unification operation, and ideally an experiment that normalizes or jointly calibrates the concatenated features without full joint training.
minor comments (6)
  1. [Fig. 2 caption] The caption contains a typo: 'MACHA' should be 'MATCHA'.
  2. [Sec. 4.3] The text says 'TAPVid dataset consists of 30 highly varying real-world video sequences'; this refers only to TAPVid-Davis. Please clarify which TAPVid subset is used and why other subsets are omitted.
  3. [Eq. (6)] The notation ':: ds' is not defined; please state explicitly that it denotes taking every ds-th channel of the feature tensor.
  4. [Tab. 4] The 'Average Score' column is not defined. Please explain how the score is computed and justify averaging the three task averages, which may have different difficulty scales.
  5. [Sec. 4.1, Tab. 1] The legend says 'Red indicates methods using image pairs as inputs', but color is not visible in the printed table; please use explicit symbols instead.
  6. [Sec. 3.2] The formula for N is typeset ambiguously; please rewrite it as N = (H/(8p)) * (W/(8p)) to avoid confusion.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: MATCHA's SOTA claims rest on external benchmarks and held-out evaluations; the self-defined TAP-Vid protocol is a validity concern, not circular reasoning.

full rationale

MATCHA is an empirical feature-learning paper rather than a formal derivation. The final unified feature Fm is constructed by (i) supervised training of Fs and Fg via Eq. (5) on standard geometric (ScanNet, MegaDepth) and semantic (PF-PASCAL, SPair-71k, AP-10k) training data and (ii) channel-wise concatenation with frozen DINOv2 features in Eq. (6). The geometric and semantic evaluations are against external benchmarks (HPatches, Aachen, PF-Willow, and held-out splits of ScanNet/MegaDepth/SPair-71k), so those numbers are not fitted to the evaluation targets. The temporal evaluation on TAP-Vid is zero-shot: MATCHA was not trained on TAP-Vid, and the reported gain over DIFT.Uni+DINOv2 is an empirical outcome rather than an input to the method. The re-purposed PCK protocol is a benchmark-validity concern (occlusion handling and normalization are not specified), but this is not a circularity: the protocol does not define the feature or the training objective. Self-citations (e.g., Dynomo, Patch2pix, SFD2) appear only in related work and are not load-bearing. No equation reduces to its own input, and no fitted parameter is renamed as a prediction. Hence no significant circularity is present.

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

The central claim is empirical; no new mathematical theory is derived. The listed free parameters are hand-set training hyperparameters and architecture sizes; learned network weights are the model itself. No new physical or conceptual entities in the sense of new particles, forces, or dimensions are introduced.

free parameters (4)
  • Fusion attention blocks k / hidden size / heads = k=8, hidden=512, heads=8
    Chosen without reported ablation; the fusion module capacity directly affects feature quality (Sec. 3.2, Implementation Details).
  • Loss temperatures and weights = tau=0.02, beta=14.3, w_cl=1.0, w_flow=1.0, w_sem=0.1
    Hand-set; these balance semantic versus geometric supervision and are central to the training balance claims (Sec. 3.3, Eqs. 8-12).
  • Output descriptor dimensions and patch size = D_g=256, D_s=768, p=2
    Hand-set; determines the information capacity of F_g and F_s before concatenation (Sec. 3.2, Implementation Details).
  • Training schedule (iterations, batch size, learning rates) = 220k iterations, batch 24/48, lr 1e-4 to 2e-5
    Chosen empirically; the multi-stage training avoids semantic overfitting (Appendix A).
assumptions (4)
  • domain assumption Stable diffusion features (DIFT) encode both geometric and semantic correspondence information in distinct layers and timesteps.
    Loaded from DIFT [60] in Sec. 3.1; if false, the fusion module has nothing useful to fuse.
  • domain assumption DINOv2 features provide object-level semantic knowledge complementary to DIFT without needing adaptation.
    Stated in Sec. 3.1 and Sec. 4.3; relies on DINOv2's object-centric training, and if the features are redundant or misaligned, concatenation could hurt.
  • domain assumption Nearest-neighbor search in the learned feature space is a valid surrogate for solving geometric, semantic, and temporal matching.
    All evaluations use nearest-neighbor matching (Sec. 4); a different matching head could change the conclusions.
  • ad hoc to paper Direct correspondence supervision on F_s and F_g transfers to the unified F_m without joint supervision.
    Explicit design choice in Sec. 3.3; Table 6 is the only indirect test and shows joint finetuning degrades semantic performance, so the paper leans on this assumption.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MATCHA:Towards Matching Anything." pith.science (2026). https://pith.science/paper/ELYAVGF3

@misc{pith2026250114945,
  author       = {Pith},
  title        = {Pith review of: MATCHA:Towards Matching Anything},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ELYAVGF3}},
  note         = {Machine review of arXiv:2501.14945}
}
read the original abstract

Establishing correspondences across images is a fundamental challenge in computer vision, underpinning tasks like Structure-from-Motion, image editing, and point tracking. Traditional methods are often specialized for specific correspondence types, geometric, semantic, or temporal, whereas humans naturally identify alignments across these domains. Inspired by this flexibility, we propose MATCHA, a unified feature model designed to ``rule them all'', establishing robust correspondences across diverse matching tasks. Building on insights that diffusion model features can encode multiple correspondence types, MATCHA augments this capacity by dynamically fusing high-level semantic and low-level geometric features through an attention-based module, creating expressive, versatile, and robust features. Additionally, MATCHA integrates object-level features from DINOv2 to further boost generalization, enabling a single feature capable of matching anything. Extensive experiments validate that MATCHA consistently surpasses state-of-the-art methods across geometric, semantic, and temporal matching tasks, setting a new foundation for a unified approach for the fundamental correspondence problem in computer vision. To the best of our knowledge, MATCHA is the first approach that is able to effectively tackle diverse matching tasks with a single unified feature.

Figures

Figures reproduced from arXiv: 2501.14945 by the authors.

Figure 1
Figure 1. Geometric correspondences identify 2D points in images of static scenes that represent the same physical 3D *This work was done when Fei Xue was an intern at NVIDIA. Semantic Geometric Temporal [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Heatmap of features from DINOv2, DIFT, and MATCHA. Given a query point from the source image (1st col￾umn), DINOv2 features give more accurate correspondences on single object (1st and 2nd row) but struggle when multiple in￾stances of the same class (3rd row) or similar structures (4th row) exist. Both geometric and semantic features of DIFT perform re￾versely. By unifying knowledge in the three foundation features,… view at source ↗
Figure 3
Figure 3. Architecture of MATCHA. Given an RGB image, MATCHA produces a single feature for geometric, semantic and temporal matching with nearest neighbor searching. MATCHA is built on top of stable diffusion (SD) models [53] and DINOv2 [44]. Specifically, original geometric and semantic features extracted from SD are first fused dynamically with a transformer [64] consists of self and cross attention blocks. In this dynamic … view at source ↗
Figures from the paper (14 more)
Figure 4
Figure 4. Figure 4: Geometric Matching on HPatches. We report Mean Matching Accuracy (MMA) at error thresholds ranging from 1-10 pixel. Concrete and dash lines denote methods with and without supervision, respectively. Megadepth [32] to evaluate the relative pose estimation. We further cr…
Figure 5
Figure 5. Figure 5: Visualization of temporal matches on TapVID￾Davis [12]. Here we visualize several challenging cases for exstablishing temporal correspondences, where MATCHA gen￾erally achieves the best performance in handling extreme scale and viewpoint changes, as well as scenes with…
Figure 6
Figure 6. Figure 6: Visualization of heatmap. Given a source point (top), we visualize the heatmap and predicted matches of MASt3R.E [30], DISK [63], DINOv2 [44], DIFT [60], DIFT.S (fully supervised version of DIFT), and our models MATCHA-Light and MATCHA [PITH_FULL_IMAGE:figures/full_fi…
Figure 7
Figure 7. Figure 7: Geometric matches on outdoor scenes. We visualize the inliers after RANSAC of MASt3R.E [30], DISK [63], DINOv2 [44], DIFT [60], DIFT.S (fully supervised version of DIFT), and our models MATCHA-Light and MATCHA. DISK produces many inliers on local patches but is not rob…
Figure 8
Figure 8. Figure 8: Geometric matches on indoor scenes. We visualize the inliers after RANSAC of MASt3R.E [30], DISK [63], DINOv2 [44], DIFT [60], DIFT.S (fully supervised version of DIFT), and our models MATCHA-Light and MATCHA. Almost all previous methods fail to find sufficient inliers…
Figure 9
Figure 9. Figure 9: Semantic matches on bus category. We visualize the inliers and outliers of MASt3R.E [30], DISK [63], DINOv2 [44], DIFT [60], DIFT.S (fully supervised version of DIFT), and our models MATCHA-Light and MATCHA [PITH_FULL_IMAGE:figures/full_fig_p016_9.png]
Figure 10
Figure 10. Figure 10: Semantic matches on plant category. We visualize the inliers and outliers of MASt3R.E [30], DISK [63], DINOv2 [44], DIFT [60], DIFT.S (fully supervised version of DIFT), and our models MATCHA-Light and MATCHA [PITH_FULL_IMAGE:figures/full_fig_p017_10.png]
Figure 11
Figure 11. Figure 11: Semantic matches on sheep category. We visualize the inliers and outliers of MASt3R.E [30], DISK [63], DINOv2 [44], DIFT [60], DIFT.S (fully supervised version of DIFT), and our models MATCHA-Light and MATCHA [PITH_FULL_IMAGE:figures/full_fig_p018_11.png]
Figure 12
Figure 12. Figure 12: Semantic matches on chair category. We visualize the inliers and outliers of MASt3R.E [30], DISK [63], DINOv2 [44], DIFT [60], DIFT.S (fully supervised version of DIFT), and our models MATCHA-Light and MATCHA [PITH_FULL_IMAGE:figures/full_fig_p019_12.png]
Figure 13
Figure 13. Figure 13: Semantic matches on motorbike category. We visualize the inliers and outliers of MASt3R.E [30], DISK [63], DINOv2 [44], DIFT [60], DIFT.S (fully supervised version of DIFT), and our models MATCHA-Light and MATCHA [PITH_FULL_IMAGE:figures/full_fig_p020_13.png]
Figure 14
Figure 14. Figure 14: Temporal matches on goldfish sequence. We visualize the inliers and outliers of MASt3R.E [30], DISK [63], DINOv2 [44], DIFT.Uni [60], DIFT.S.Uni (fully supervised version of DIFT), and our models MATCHA-Light and MATCHA [PITH_FULL_IMAGE:figures/full_fig_p021_14.png]
Figure 15
Figure 15. Figure 15: Temporal matches on horsejumphigh sequence. We visualize the inliers and outliers of MASt3R.E [30], DISK [63], DI￾NOv2 [44], DIFT.Uni [60], DIFT.S.Uni (fully supervised version of DIFT), and our models MATCHA-Light and MATCHA [PITH_FULL_IMAGE:figures/full_fig_p022_15.png]
Figure 16
Figure 16. Figure 16: Temporal matches on soapbox sequence. We visualize the inliers and outliers of MASt3R.E [30], DISK [63], DINOv2 [44], DIFT.Uni [60], DIFT.S.Uni (fully supervised version of DIFT), and our models MATCHA-Light and MATCHA [PITH_FULL_IMAGE:figures/full_fig_p023_16.png]
Figure 17
Figure 17. Figure 17: Temporal matches on scooterblack sequence. We visualize the inliers and outliers of MASt3R.E [30], DISK [63], DI￾NOv2 [44], DIFT.Uni [60], DIFT.S.Uni (fully supervised version of DIFT), and our models MATCHA-Light and MATCHA [PITH_FULL_IMAGE:figures/full_fig_p024_17.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

76 extracted references · 56 canonical work pages

  1. [1]

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

    Vassileios Balntas, Karel Lenc, Andrea Vedaldi, and Krys- tian Mikolajczyk. Hpatches: A benchmark and evaluation of handcrafted and learned local descriptors. In Proceedings of the IEEE conference on computer vision and pattern recog- nition, pages 5173–5182, 2017. 5

    work page 2017

  2. [2]

    Speeded-up robust features (surf)

    Herbert Bay, Andreas Ess, Tinne Tuytelaars, and Luc Van Gool. Speeded-up robust features (surf). Computer vi- sion and image understanding, 110(3):346–359, 2008. 2

    2008

  3. [3]

    Emerg- ing properties in self-supervised vision transformers

    Mathilde Caron, Hugo Touvron, Ishan Misra, Herv ´e J´egou, Julien Mairal, Piotr Bojanowski, and Armand Joulin. Emerg- ing properties in self-supervised vision transformers. In Pro- ceedings of the IEEE/CVF international conference on com- puter vision, pages 9650–9660, 2021. 3

    2021

  4. [4]

    Aspanformer: Detector-free image matching with adaptive span transformer

    Hongkai Chen, Zixin Luo, Lei Zhou, Yurun Tian, Ming- min Zhen, Tian Fang, David Mckinnon, Yanghai Tsin, and Long Quan. Aspanformer: Detector-free image matching with adaptive span transformer. In European Conference on Computer Vision, pages 20–36. Springer Nature Switzerland Cham, 2022. 2

    work page 2022

  5. [5]

    Cats++: Boosting cost aggregation with convolutions and transformers

    Seokju Cho, Sunghwan Hong, and Seungryong Kim. Cats++: Boosting cost aggregation with convolutions and transformers. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(6):7174–7194, 2022. 3, 5

    work page 2022

  6. [6]

    Local All-Pair Correspondence for Point Tracking

    Seokju Cho, Jiahui Huang, Jisu Nam, Honggyu An, Seun- gryong Kim, and Joon-Young Lee. Local all-pair correspon- dence for point tracking. arXiv preprint arXiv:2407.15420,

    work page Pith review arXiv

  7. [7]

    Universal correspondence network

    Christopher B Choy, JunYoung Gwak, Silvio Savarese, and Manmohan Chandraker. Universal correspondence network. Advances in neural information processing systems , 29,

  8. [8]

    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, pages 236–243. Springer, 2003. 6

    work page 2003

Show all 76 references
  1. [9]

    Scannet: Richly-annotated 3d reconstructions of indoor scenes

    Angela Dai, Angel X Chang, Manolis Savva, Maciej Hal- ber, Thomas Funkhouser, and Matthias Nießner. Scannet: Richly-annotated 3d reconstructions of indoor scenes. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 5828–5839, 2017. 5, 6, 9

    2017

  2. [10]

    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 Proceedings of the IEEE conference on computer vision and pattern recognition workshops , pages 224–236, 2018. 1, 2, 6, 9

    2018

  3. [11]

    Diffusion models beat gans on image synthesis

    Prafulla Dhariwal and Alexander Nichol. Diffusion models beat gans on image synthesis. Advances in neural informa- tion processing systems, 34:8780–8794, 2021. 1, 3

    2021

  4. [12]

    Tap-vid: A benchmark for track- ing any point in a video

    Carl Doersch, Ankush Gupta, Larisa Markeeva, Adria Re- casens, Lucas Smaira, Yusuf Aytar, Joao Carreira, Andrew Zisserman, and Yi Yang. Tap-vid: A benchmark for track- ing any point in a video. Advances in Neural Information Processing Systems, 35:13610–13626, 2022. 1, 2, 3, 6, 7, 10

    2022

  5. [13]

    D2- net: A trainable cnn for joint description and detection 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 description and detection of local features. In Proceedings of the ieee/cvf conference on computer vision and pattern recognition, ...

  6. [14]

    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 Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 17765–17775, 2023. 2

    2023

  7. [15]

    Roma: Robust dense fea- ture matching

    Johan Edstedt, Qiyu Sun, Georg B ¨okman, M ˚arten Wadenb¨ack, and Michael Felsberg. Roma: Robust dense fea- ture matching. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 19790– 19800, 2024. 3

    2024

  8. [16]

    S2dnet: Learning accurate correspondences for sparse-to- dense feature matching

    Hugo Germain, Guillaume Bourmaud, and Vincent Lepetit. S2dnet: Learning accurate correspondences for sparse-to- dense feature matching. arXiv preprint arXiv:2004.01673 ,

    2004 arXiv

  9. [17]

    Proposal flow

    Bumsub Ham, Minsu Cho, Cordelia Schmid, and Jean Ponce. Proposal flow. In Proceedings of the IEEE Con- ference on Computer Vision and Pattern Recognition, pages 3475–3484, 2016. 5, 7

    2016

  10. [18]

    Proposal flow: Semantic correspondences from ob- ject proposals

    Bumsub Ham, Minsu Cho, Cordelia Schmid, and Jean Ponce. Proposal flow: Semantic correspondences from ob- ject proposals. IEEE transactions on pattern analysis and machine intelligence, 40(7):1711–1725, 2017. 5, 9

    2017

  11. [19]

    Harley, Zhaoyuan Fang, and Katerina Fragkiadaki

    Adam W. Harley, Zhaoyuan Fang, and Katerina Fragkiadaki. Particle video revisited: Tracking through occlusions using point trajectories. In ECCV, 2022. 1, 3

    2022

  12. [20]

    Unsupervised semantic correspondence using stable diffu- sion

    Eric Hedlin, Gopal Sharma, Shweta Mahajan, Hossam Isack, Abhishek Kar, Andrea Tagliasacchi, and Kwang Moo Yi. Unsupervised semantic correspondence using stable diffu- sion. Advances in Neural Information Processing Systems , 36, 2024. 3, 5

    2024

  13. [21]

    Learning semantic corre- spondence with sparse annotations

    Shuaiyi Huang, Luyu Yang, Bo He, Songyang Zhang, Xum- ing He, and Abhinav Shrivastava. Learning semantic corre- spondence with sparse annotations. In European Conference on Computer Vision, pages 267–284. Springer, 2022. 3, 5

    2022

  14. [22]

    Openclip, 2021

    Gabriel Ilharco, Mitchell Wortsman, Ross Wightman, Cade Gordon, Nicholas Carlini, Rohan Taori, Achal Dave, Vaishaal Shankar, Hongseok Namkoong, John Miller, et al. Openclip, 2021. 3

    2021

  15. [23]

    Omniglue: Generalizable feature match- ing with foundation model guidance

    Hanwen Jiang, Arjun Karpur, Bingyi Cao, Qixing Huang, and Andr´e Araujo. Omniglue: Generalizable feature match- ing with foundation model guidance. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 19865–19875, 2024. 2, 3

    2024

  16. [24]

    Co- tracker: It is better to track together

    Nikita Karaev, Ignacio Rocco, Benjamin Graham, Natalia Neverova, Andrea Vedaldi, and Christian Rupprecht. Co- tracker: It is better to track together. arXiv preprint arXiv:2307.07635, 2023. 1, 3

    2023 arXiv

  17. [25]

    De- formable spatial pyramid matching for fast dense correspon- dences

    Jaechul Kim, Ce Liu, Fei Sha, and Kristen Grauman. De- formable spatial pyramid matching for fast dense correspon- dences. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 2307–2314, 2013. 3

    2013

  18. [26]

    Fcss: Fully con- volutional self-similarity for dense semantic correspondence

    Seungryong Kim, Dongbo Min, Bumsub Ham, Sangryul Jeon, Stephen Lin, and Kwanghoon Sohn. Fcss: Fully con- volutional self-similarity for dense semantic correspondence. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 6560–6569, 2017. 1, 3

    2017

  19. [27]

    Semantic attribute matching networks

    Seungryong Kim, Dongbo Min, Somi Jeong, Sunok Kim, Sangryul Jeon, and Kwanghoon Sohn. Semantic attribute matching networks. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 12339–12348, 2019. 3

    2019

  20. [28]

    Poselib-minimal solvers for camera pose es- timation, 2020

    Viktor Larsson. Poselib-minimal solvers for camera pose es- timation, 2020. 6

    2020

  21. [29]

    Sfnet: Learning object-aware semantic correspon- dence

    Junghyup Lee, Dohyung Kim, Jean Ponce, and Bumsub Ham. Sfnet: Learning object-aware semantic correspon- dence. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition , pages 2278–2287,

  22. [30]

    Ground- ing image matching in 3d with mast3r.European Conference on Computer Vision, 2024

    Vincent Leroy, Yohann Cabon, and J´erˆome Revaud. Ground- ing image matching in 3d with mast3r.European Conference on Computer Vision, 2024. 6, 8, 10, 11, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24

    2024

  23. [31]

    Sd4match: Learning to prompt stable diffu- sion model for semantic matching

    Xinghui Li, Jingyi Lu, Kai Han, and Victor Adrian Prisacariu. Sd4match: Learning to prompt stable diffu- sion model for semantic matching. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 27558–27568, 2024. 1, 3, 5

    2024

  24. [32]

    Megadepth: Learning single- view depth prediction from internet photos

    Zhengqi Li and Noah Snavely. Megadepth: Learning single- view depth prediction from internet photos. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 2041–2050, 2018. 6, 9

    2018

  25. [33]

    Lightglue: Local feature matching at light speed

    Philipp Lindenberger, Paul-Edouard Sarlin, and Marc Polle- feys. Lightglue: Local feature matching at light speed. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 17627–17638, 2023. 2

    2023

  26. [34]

    Sift flow: Dense correspondence across scenes and its applications

    Ce Liu, Jenny Yuen, and Antonio Torralba. Sift flow: Dense correspondence across scenes and its applications. IEEE transactions on pattern analysis and machine intelligence , 33(5):978–994, 2010. 3

    2010

  27. [35]

    Decoupled weight decay regularization

    I Loshchilov. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101, 2017. 10

    2017 arXiv

  28. [36]

    Distinctive image features from scale- invariant keypoints

    David G Lowe. Distinctive image features from scale- invariant keypoints. International journal of computer vi- sion, 60:91–110, 2004. 1, 2

    2004

  29. [37]

    Dynamic 3d gaussians: Tracking by per- sistent dynamic view synthesis

    Jonathon Luiten, Georgios Kopanas, Bastian Leibe, and Deva Ramanan. Dynamic 3d gaussians: Tracking by per- sistent dynamic view synthesis. In 3DV, 2024. 3

    2024

  30. [38]

    Diffusion hyperfeatures: Searching through time and space for semantic correspondence

    Grace Luo, Lisa Dunlap, Dong Huk Park, Aleksander Holyn- ski, and Trevor Darrell. Diffusion hyperfeatures: Searching through time and space for semantic correspondence. Ad- vances in Neural Information Processing Systems, 36, 2024. 1, 3, 5

    2024

  31. [39]

    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 Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 2527–2536, 2019. 2

    2019

  32. [40]

    Spair-71k: A large-scale benchmark for semantic correspon- dence

    Juhong Min, Jongmin Lee, Jean Ponce, and Minsu Cho. Spair-71k: A large-scale benchmark for semantic correspon- dence. arXiv preprint arXiv:1908.10543, 2019. 5, 9

    1908 arXiv

  33. [41]

    Scalable nearest neigh- bor algorithms for high dimensional data

    Marius Muja and David G Lowe. Scalable nearest neigh- bor algorithms for high dimensional data. IEEE transactions on pattern analysis and machine intelligence , 36(11):2227– 2240, 2014. 2

    2014

  34. [42]

    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 , 31(5):1147–1163,

  35. [43]

    Neural congealing: Aligning images to a joint se- mantic atlas

    Dolev Ofri-Amar, Michal Geyer, Yoni Kasten, and Tali Dekel. Neural congealing: Aligning images to a joint se- mantic atlas. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 19403– 19412, 2023. 1

    2023

  36. [44]

    Dinov2: Learning robust visual features without supervision

    Maxime Oquab, Timoth ´ee Darcet, Th ´eo Moutakanni, Huy V o, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, et al. Dinov2: Learning robust visual features without supervision. arXiv preprint arXiv:2304.07193, 2023. 3, 4, ...

    2023 arXiv

  37. [45]

    Pytorch: An im- perative style, high-performance deep learning library

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. Pytorch: An im- perative style, high-performance deep learning library. Ad- vances in neural information processing systems ...

    2019

  38. [46]

    Xfeat: Accelerated fea- tures for lightweight image matching

    Guilherme Potje, Felipe Cadar, Andr ´e Araujo, Renato Mar- tins, and Erickson R Nascimento. Xfeat: Accelerated fea- tures for lightweight image matching. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2682–2691, 2024. 1, 5, 6, 8, 9

    2024

  39. [47]

    Learning transferable visual models from natural language supervi- sion

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervi- sion. In International conference on machine learning, ...

    2021

  40. [48]

    R2d2: Reliable and repeatable detec- tor and descriptor

    Jerome Revaud, Cesar De Souza, Martin Humenberger, and Philippe Weinzaepfel. R2d2: Reliable and repeatable detec- tor and descriptor. Advances in neural information process- ing systems, 32, 2019. 1, 2, 5, 6, 9

    2019

  41. [49]

    Convo- lutional neural network architecture for geometric matching

    Ignacio Rocco, Relja Arandjelovic, and Josef Sivic. Convo- lutional neural network architecture for geometric matching. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 6148–6157, 2017. 3

    2017

  42. [50]

    End- to-end weakly-supervised semantic alignment

    Ignacio Rocco, Relja Arandjelovi ´c, and Josef Sivic. End- to-end weakly-supervised semantic alignment. In Proceed- ings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 6917–6925, 2018. 3

    2018

  43. [51]

    Neighbourhood con- sensus networks

    Ignacio Rocco, Mircea Cimpoi, Relja Arandjelovi ´c, Akihiko Torii, Tomas Pajdla, and Josef Sivic. Neighbourhood con- sensus networks. Advances in neural information processing systems, 31, 2018. 2, 3

    2018

  44. [52]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022. 3

    2022

  45. [53]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022. 1, 3, 4

    2022

  46. [54]

    Particle video: Long-range mo- tion estimation using point trajectories.International journal of computer vision, 80:72–91, 2008

    Peter Sand and Seth Teller. Particle video: Long-range mo- tion estimation using point trajectories.International journal of computer vision, 80:72–91, 2008. 1, 3

    2008

  47. [55]

    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 Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 4938–4947, 2020. 1, 2

    2020

  48. [56]

    Benchmarking 6dof outdoor visual localization in changing conditions

    Torsten Sattler, Will Maddern, Carl Toft, Akihiko Torii, Lars Hammarstrand, Erik Stenborg, Daniel Safari, Masatoshi Okutomi, Marc Pollefeys, Josef Sivic, et al. Benchmarking 6dof outdoor visual localization in changing conditions. In Proceedings of the IEEE conference on compu...

    2018

  49. [57]

    Structure- from-motion revisited

    Johannes L Schonberger and Jan-Michael Frahm. Structure- from-motion revisited. In Proceedings of the IEEE con- ference on computer vision and pattern recognition , pages 4104–4113, 2016. 1

    2016

  50. [58]

    Dynomo: Online point tracking by dynamic online monocular gaussian recon- struction

    Jenny Seidenschwarz, Qunjie Zhou, Bardienus Duisterhof, Deva Ramanan, and Laura Leal-Taix ´e. Dynomo: Online point tracking by dynamic online monocular gaussian recon- struction. arXiv preprint arXiv:2409.02104, 2024. 3

    2024 arXiv

  51. [59]

    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 Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition , pages 8922–8931, 2021. 2, 5, 9

    2021

  52. [60]

    Emergent correspondence from image diffusion

    Luming Tang, Menglin Jia, Qianqian Wang, Cheng Perng Phoo, and Bharath Hariharan. Emergent correspondence from image diffusion. Advances in Neural Information Pro- cessing Systems, 36:1363–1389, 2023. 1, 3, 5, 6, 8, 10, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24

    2023

  53. [61]

    Sosnet: Second order similarity reg- ularization for local descriptor learning

    Yurun Tian, Xin Yu, Bin Fan, Fuchao Wu, Huub Heijnen, and Vassileios Balntas. Sosnet: Second order similarity reg- ularization for local descriptor learning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 11016–11025, 2019. 2

    2019

  54. [62]

    Glu- net: Global-local universal network for dense flow and corre- spondences

    Prune Truong, Martin Danelljan, and Radu Timofte. Glu- net: Global-local universal network for dense flow and corre- spondences. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 6258–6268,

  55. [63]

    Disk: Learning local features with policy gradient

    Michał Tyszkiewicz, Pascal Fua, and Eduard Trulls. Disk: Learning local features with policy gradient. Advances in Neural Information Processing Systems , 33:14254–14265,

  56. [64]

    Attention is all you need

    A Vaswani. Attention is all you need. Advances in Neural Information Processing Systems, 2017. 4

    2017

  57. [65]

    Tracking everything everywhere all at once

    Qianqian Wang, Yen-Yu Chang, Ruojin Cai, Zhengqi Li, Bharath Hariharan, Aleksander Holynski, and Noah Snavely. Tracking everything everywhere all at once. InInternational Conference on Computer Vision, 2023. 3

    2023

  58. [66]

    Shape of motion: 4d re- construction from a single video

    Qianqian Wang, Vickie Ye, Hang Gao, Jake Austin, Zhengqi Li, and Angjoo Kanazawa. Shape of motion: 4d re- construction from a single video. In arXiv preprint arXiv:2407.13764, 2024. 3

    2024

  59. [67]

    Croco: Self-supervised pre-training for 3d vision tasks by cross-view completion

    Philippe Weinzaepfel, Vincent Leroy, Thomas Lucas, Ro- main Br´egier, Yohann Cabon, Vaibhav Arora, Leonid Ants- feld, Boris Chidlovskii, Gabriela Csurka, and J ´erˆome Re- vaud. Croco: Self-supervised pre-training for 3d vision tasks by cross-view completion. Advances in Neura...

    2022

  60. [68]

    Spatialtracker: Tracking any 2d pixels in 3d space

    Yuxi Xiao, Qianqian Wang, Shangzhan Zhang, Nan Xue, Sida Peng, Yujun Shen, and Xiaowei Zhou. Spatialtracker: Tracking any 2d pixels in 3d space. In CVPR, 2024. 3

    2024

  61. [69]

    Rnnpose: Recurrent 6-dof object pose refinement with robust correspondence field estimation and pose optimization

    Yan Xu, Kwan-Yee Lin, Guofeng Zhang, Xiaogang Wang, and Hongsheng Li. Rnnpose: Recurrent 6-dof object pose refinement with robust correspondence field estimation and pose optimization. In Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition , pag...

    2022

  62. [70]

    Sfd2: Semantic-guided feature detection and description

    Fei Xue, Ignas Budvytis, and Roberto Cipolla. Sfd2: Semantic-guided feature detection and description. In Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5206–5216, 2023. 2

    2023

  63. [71]

    Lift: Learned invariant feature transform

    Kwang Moo Yi, Eduard Trulls, Vincent Lepetit, and Pascal Fua. Lift: Learned invariant feature transform. In Com- puter Vision–ECCV 2016: 14th European Conference, Am- sterdam, The Netherlands, October 11-14, 2016, Proceed- ings, Part VI 14, pages 467–483. Springer Internationa...

    2016

  64. [72]

    Ap-10k: A benchmark for animal pose esti- mation in the wild

    Hang Yu, Yufei Xu, Jing Zhang, Wei Zhao, Ziyu Guan, and Dacheng Tao. Ap-10k: A benchmark for animal pose esti- mation in the wild. arXiv preprint arXiv:2108.12617, 2021. 9

    2021 arXiv

  65. [73]

    Telling left from right: Identifying geometry-aware semantic corre- spondence

    Junyi Zhang, Charles Herrmann, Junhwa Hur, Eric Chen, Varun Jampani, Deqing Sun, and Ming-Hsuan Yang. Telling left from right: Identifying geometry-aware semantic corre- spondence. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 307...

    2024

  66. [74]

    A tale of two features: Stable diffusion complements dino for zero-shot semantic correspondence

    Junyi Zhang, Charles Herrmann, Junhwa Hur, Luisa Pola- nia Cabrera, Varun Jampani, Deqing Sun, and Ming-Hsuan Yang. A tale of two features: Stable diffusion complements dino for zero-shot semantic correspondence. Advances in Neural Information Processing Systems, 36, 2024. 1, 3, 5

    2024

  67. [75]

    Patch2pix: Epipolar-guided pixel-level correspondences

    Qunjie Zhou, Torsten Sattler, and Laura Leal-Taixe. Patch2pix: Epipolar-guided pixel-level correspondences. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 4669–4678, 2021. 2, 6

    2021

  68. [2020]

    1, 2, 5, 6, 8, 10, 11, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24

Pith tools

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