Image-Matching-Paper-List

Image-Matching-Paper-List

A personal list of papers and resources for image matching, pose estimation and some other 3D reconstruction tasks, including perspective images and panoramas (marked with :globe_with_meridians:).

---

• Survey
• Sparse matching (detector-based)
  • Keypoints and descriptors
  • Feature matching
• Semi-dense matching (detector-free)
• Dense matching
• Multi-view
• Training framework
• Pose estimation and others
• Similar images disambiguate
• Datasets
• Challenges and workshops
• Resources and toolboxes

---

Survey

• Local Feature Matching Using Deep Learning: A Survey [arXiv 2024] []

• Local feature matching from detector-based to detector-free: a survey [Applied Intelligence 2024] [] ()

• Deep Learning Reforms Image Matching: A Survey and Outlook [arXiv 2025] []

Sparse

Keypoints and descriptors

• ORB: An efficient alternative to SIFT or SURF [ICCV 2011] []

• :globe_with_meridians: SPHORB: A Fast and Robust Binary Feature on the Sphere [IJCV 2015] [SPHORB]

• Working hard to know your neighbor's margins: Local descriptor learning loss [NeurIPS 2017] [hardnet]

• Repeatability Is Not Enough: Learning Discriminative Affine Regions via Discriminability [ECCV 2018] [affnet]

• Efficient adaptive non-maximal suppression algorithms for homogeneous spatial keypoint distribution [Pattern Recognition Letters 2018] [ANMS-Codes]

• SuperPoint: Self-Supervised Interest Point Detection and Description [CVPRW 2018] [SuperPointPretrainedNetwork]

• Key.Net: Keypoint Detection by Handcrafted and Learned CNN Filters [ICCV 2019] [Key.Net-Pytorch]

• D2-net: A trainable cnn for joint description and detection of local features [CVPR 2019] [d2-net]

• R2D2: Repeatable and Reliable Detector and Descriptor [NeurIPS 2019] [r2d2]

• ASLFeat: Learning Local Features of Accurate Shape and Localization [CVPR 2020] [ASLFeat]

• DISK: Learning local features with policy gradient [NeurIPS 2020] [disk]

• Online Invariance Selection for Local Feature Descriptors [ECCV 2020] [LISRD]

• Learning Feature Descriptors using Camera Pose Supervision [ECCV 2020] [CAPS]

• Co-attention for conditioned image matching [CVPR 2021] [coam]

• Rethinking Low-level Features for Interest Point Detection and Description [ACCVC 2022] [lanet]

• ALIKE: Accurate and Lightweight Keypoint Detection and Descriptor Extraction [TMM 2022] [ALIKE]

• Decoupling Makes Weakly Supervised Local Feature Better [CVPR 2022] [PoSFeat]

• Shared Coupling-bridge for Weakly Supervised Local Feature Learning [arXiv 2022] [SCFeat]

• Self-Supervised Equivariant Learning for Oriented Keypoint Detection [CVPR 2022] [REKD]

• Image Matching and Localization Based on Fusion of Handcrafted and Deep Features [IEEE Sensors Journal 2023] [DeFusion]

• Robust feature matching via progressive smoothness consensus [ISPRS 2023] [Robust-feature-matching-via-Progressive-Smoothness-Consensus]

• SiLK: Simple Learned Keypoints [ICCV 2023] [silk]

• ALIKED: A Lighter Keypoint and Descriptor Extraction Network via Deformable Transformation [IEEE Trans Instrum Meas 2023] [ALIKED]

• MTLDesc: Looking Wider to Describe Better [AAAI 2022] [mtldesc]

• Attention Weighted Local Descriptors [TPAMI 2023] [AWDesc]

• FeatureBooster: Boosting Feature Descriptors with a Lightweight Neural Network [CVPR 2023] [FeatureBooster]

• SFD2: Semantic-guided Feature Detection and Description [arXiv 2023] [sfd2]

• :globe_with_meridians: PanoPoint: Self-Supervised Feature Points Detection and Description for 360° Panorama [CVPRW 2023] []

• DeDoDe: Detect, Don't Describe -- Describe, Don't Detect for Local Feature Matching [3DV 2024] [DeDoDe]

• S-TREK: Sequential Translation and Rotation Equivariant Keypoints for local feature extraction [ICCV 2023] []

• DarkFeat: Noise-Robust Feature Detector and Descriptor for Extremely Low-Light RAW Images [AAAI 2023] [DarkFeat]

• Steerers: A framework for rotation equivariant keypoint descriptors [arXiv 2023] [rotation-steerers]

• NeRF-Supervised Feature Point Detection and Description [arXiv 2024] []

• DeDoDe v2: Analyzing and Improving the DeDoDe Keypoint Detector [CVPRW 2024] [DeDoDe]

• XFeat: Accelerated Features for Lightweight Image Matching [CVPR 2024] [accelerated_features]

• DaD: Distilled Reinforcement Learning for Diverse Keypoint Detection [arXiv 2025] [dad]

• LiftFeat: 3D Geometry-Aware Local Feature Matching [ICRA 2025] [LiftFeat]

• RIPE: Reinforcement Learning on Unlabeled Image Pairs for Robust Keypoint Extraction [ICCV 2025] [RIPE]

Feature matching

Filter

• GMS: Grid-based Motion Statistics for Fast, Ultra-Robust Feature Correspondence [IJCV 2020] [GMS-Feature-Matcher]

• Learning Two-View Correspondences and Geometry Using Order-Aware Network [ICCV 2019] [OANet]

• Learning to Find Good Correspondences [CVPR 2018] [learned-correspondence-release]

• ACNe: Attentive Context Normalization for Robust Permutation-Equivariant Learning [CVPR 2020] [acne]

• Progressive Correspondence Pruning by Consensus Learning [ICCV 2021] [CLNet]

• PGFNet: Preference-Guided Filtering Network for Two-View Correspondence Learning [TIP 2023] [PGFNet]

• Pentagon-Match (PMatch): Identification of View-Invariant Planar Feature for Local Feature Matching-Based Homography Estimation [arXiv 2023] []

• ConvMatch: Rethinking Network Design for Two-View Correspondence Learning [AAAI 2023] [ConvMatch]

• Progressive Neighbor Consistency Mining for Correspondence Pruning [CVPR 2023] [NCMNet]

• A more reliable local-global-guided network for correspondence pruning [Pattern Recognition Letters 2024] [LG-Net]

• MESA: Matching Everything by Segmenting Anything [CVPR 2024] [A2PM-MESA]

• DMESA: Densely Matching Everything by Segmenting Anything [arXiv 2024] [A2PM-MESA]

• FC-GNN: Recovering Reliable and Accurate Correspondences from Interferences [CVPR 2024] [fcgnn]

• CorrMAE: Pre-training Correspondence Transformers with Masked Autoencoder [arXiv 2024] []

• DeMatch: Deep Decomposition of Motion Field for Two-View Correspondence Learning [CVPR 2024] [DeMatch]

• Image Matching Filtering and Refinement by Planes and Beyond [arXiv 2024] [MiHo]

• DeMo: Deep Motion Field Consensus with Learnable Kernels for Two-view Correspondence Learning [AAAI 2025] [DeMo]

Matcher

• SuperGlue: Learning Feature Matching with Graph Neural Networks [CVPR 2020] [SuperGluePretrainedNetwork]

• Learning to Match Features with Seeded Graph Matching Network [ICCV 2021] [SGMNet]

• NCTR: Neighborhood Consensus Transformer for Feature Matching [ICIP 2022] [NCTR]

• HTMatch: An efficient Hybrid Transformer based Graph Neural Network for Local Feature Matching [Signal Processing 2023] []

• ClusterGNN: Cluster-based Coarse-to-Fine Graph Neural Network for Efficient Feature Matching [CVPR 2022] []

• ParaFormer: Parallel Attention Transformer for Efficient Feature Matching [arXiv 2023] []

• AMatFormer: Efficient Feature Matching via Anchor Matching Transformer [TMM 2023] []

• :globe_with_meridians: SphereGlue: Learning Keypoint Matching on High Resolution Spherical Images [CVPRW 2023] [SphereGlue]

• LightGlue: Local Feature Matching at Light Speed [ICCV 2023] [LightGlue]

• ResMatch: Residual Attention Learning for Local Feature Matching [AAAI 2024] [ResMatch]

• SDGMNet: Statistic-based Dynamic Gradient Modulation for Local Descriptor Learning [AAAI 2024] [SDGMNet]

• Learning Feature Matching via Matchable Keypoint-Assisted Graph Neural Network [arXiv 2023] []

• IMP: Iterative Matching and Pose Estimation with Adaptive Pooling [CVPR 2023] [imp-release]

• Scene-Aware Feature Matching [ICCV 2023] []

• DynamicGlue: Epipolar and Time-Informed Data Association in Dynamic Environments using Graph Neural Networks [arXiv 2024] []

• OmniGlue: Generalizable Feature Matching with Foundation Model Guidance [CVPR 2024] [omniglue]

• DiffGlue: Diffusion-Aided Image Feature Matching [ACM MM 2024] [DiffGlue]

• MambaGlue: Fast and Robust Local Feature Matching With Mamba [ICRA 2025] [MambaGlue]

• CoMatcher: Multi-View Collaborative Feature Matching [arXiv 2025] []

• Matching While Perceiving: Enhance Image Feature Matching with Applicable Semantic Amalgamation [AAAI 2025] [SemaGlue]

• GIMS: Image Matching System Based on Adaptive Graph Construction and Graph Neural Network [Neural Networks 2025] [GIMS]

Refinement

• Learning to Make Keypoints Sub-Pixel Accurate [ECCV 2024] [keypt2subpx]

---

Semi-dense

• Neighbourhood Consensus Networks [NeurIPS 2018] []

• Efficient neighbourhood consensus networks via submanifold sparse convolutions [ECCV 2020] [sparse-ncnet]

• Dual-resolution correspondence networks [NeurIPS 2020] []

• XResolution Correspondence Network [BMVC 2021] [xrcnet]

• Patch2Pix: Epipolar-Guided Pixel-Level Correspondences [CVPR 2021] [patch2pix]

• DFM: A Performance Baseline for Deep Feature Matching [CVPR 2021] [DFM]

• LoFTR: Detector-Free Local Feature Matching with Transformers [CVPR 2021] [LoFTR]

• A case for using rotation invariant features in state of the art feature matchers [CVPRW 2022] [se2-loftr]

• 3DG-STFM: 3D Geometric Guided Student-Teacher Feature Matching [ECCV 2022] [3DG-STFM]

• Local Feature Matching with Transformers for low-end devices [arXiv 2022] [Coarse_LoFTR_TRT]

• QuadTree Attention for Vision Transformers [ICLR 2022] [QuadTreeAttention]

• MatchFormer: Interleaving Attention in Transformers for Feature Matching [ACCV 2022] [MatchFormer]

• ASpanFormer: Detector-Free Matching with Adaptive Span Transformer [ECCV 2022] [ml-aspanformer]

• TopicFM: Robust and Interpretable Topic-Assisted Feature Matching [AAAI 2023] [TopicFM]

• DeepMatcher: A Deep Transformer-based Network for Robust and Accurate Local Feature Matching [arXiv 2023] [DeepMatcher]

• OAMatcher: An Overlapping Areas-based Network for Accurate Local Feature Matching [arXiv 2023] [OAMatcher]

• PATS: Patch Area Transportation with Subdivision for Local Feature Matching [CVPR 2023] [pats]

• PA-LoFTR: Local Feature Matching with 3D Position-Aware Transformer [arXiv 2023] []

• Improving Transformer-based Image Matching by Cascaded Capturing Spatially Informative Keypoints [arXiv 2023] []

• Structured Epipolar Matcher for Local Feature Matching [CVPR 2023] [SEM]

• Adaptive Spot-Guided Transformer for Consistent Local Feature Matching [CVPR 2023] [astr]

• GlueStick: Robust Image Matching by Sticking Points and Lines Together [ICCV 2023] [GlueStick]

• E3CM: Epipolar-Constrained Cascade Correspondence Matching [Neurocomputing 2023] [E3CM]

• MAIM: a mixer MLP architecture for image matching [Unknown 2023] []

• Searching from Area to Point: A Hierarchical Framework for Semantic-Geometric Combined Feature Matching [arXiv 2023] [SGAM]

• Adaptive Assignment for Geometry Aware Local Feature Matching [CVPR 2023] [AdaMatcher]

• TopicFM+: Boosting Accuracy and Efficiency of Topic-Assisted Feature Matching [arXiv 2023] [TopicFM]

• TKwinFormer: Top k Window Attention in Vision Transformers for Feature Matching [arXiv 2023] [TKwinFormer]

• Occ2Net: Robust Image Matching Based on 3D Occupancy Estimation for Occluded Regions [ICCV 2023] []

• FMRT: Learning Accurate Feature Matching with Reconciliatory Transformer [arXiv 2023] []

• SAM-Net: Self-Attention based Feature Matching with Spatial transformers and Knowledge Distillation [ESWA 2023] [SAM-Net]

• Are Semi-Dense Detector-Free Methods Good at Matching Local Features ? [arXiv 2024] []

• Efficient LoFTR: Semi-Dense Local Feature Matching with Sparse-Like Speed [CVPR 2024] [efficientloftr]

• HCPM: Hierarchical Candidates Pruning for Efficient Detector-Free Matching [arXiv 2024] []

• Affine-based Deformable Attention and Selective Fusion for Semi-dense Matching [arXiv 2024] []

• Raising the Ceiling: Conflict-Free Local Feature Matching with Dynamic View Switching [ECCV 2024] []

• Eto: Efficient transformer-based local feature matching by organizing multiple homography hypotheses [arXiv 2024] []

• JamMa: Ultra-lightweight Local Feature Matching with Joint Mamba [CVPR 2025] [JamMa]

• EDM: Efficient Deep Feature Matching [ICCV 2025] [EDM]

• HomoMatcher: Achieving Dense Feature Matching with Semi-Dense Effciency by Homography Estimation [AAAI 2025] []

• CoMatch: Dynamic Covisibility-Aware Transformer for Bilateral Subpixel-Level Semi-Dense Image Matching [arXiv 2025] []

• Focus What Matters: Matchability-Based Reweighting for Local Feature Matching [arXiv 2025] []

• Mind the Gap: Aligning Vision Foundation Models to Image Feature Matching [ICCV 2025] []

Dense

• Dgc-net: Dense geometric correspondence network [WACV 2019] [DGC-Net]

• Ransac-flow: generic two-stage image alignment [ECCV 2020] [RANSAC-Flow]

• GLU-Net: Global-local universal network for dense flow and correspondences [CVPR 2020] [GLU-Net]

• DenseGAP: Graph-Structured Dense Correspondence Learning with Anchor Points [ICPR 2022] [DenseGAP]

• Learning accurate dense correspondences and when to trust them [CVPR 2021] [PDCNet]

• Pdc-net+: Enhanced probabilistic dense correspondence network [TPAMI 2023] [DenseMatching]

• COTR: Correspondence Transformer for Matching Across Images [ICCV 2021] [COTR]

• ECO-TR: Efficient Correspondences Finding Via Coarse-to-Fine Refinement [ECCV 2022] [ECO-TR]

• PUMP: Pyramidal and Uniqueness Matching Priors for Unsupervised Learning of Local Descriptors [CVPR 2022] [pump]

• DKM: Dense Kernelized Feature Matching for Geometry Estimation [CVPR 2023] [DKM]

• PMatch: Paired Masked Image Modeling for Dense Geometric Matching [CVPR 2023] [PMatch]

• RoMa: Revisiting Robust Losses for Dense Feature Matching [CVPR 2024] [RoMa]

• RGM: A Robust Generalist Matching Model [arXiv 2023] [RGM]

• Learning Affine Correspondences by Integrating Geometric Constraints [CVPR 2025] [DenseAffine]

• Learning Dense Feature Matching via Lifting Single 2D Image to 3D Space [ICCV 2025] [L2M]

• ArgMatch: Adaptive Refinement Gathering for Efficient Dense Matching [ICCV 2025] [argmatch]

• RoMa v2: Harder Better Faster Denser Feature Matching [arXiv 2025] [romav2]

---

Multi-view

• End2End Multi-View Feature Matching with Differentiable Pose Optimization [ICCV 2023] [e2e_multi_view_matching]

• CoMatcher: Multi-View Collaborative Feature Matching [CVPR 2025] [CoMatcher]

---

Training framework

• GIM: Learning Generalizable Image Matcher From Internet Videos [ICLR 2024] [gim]

---

Pose estimation and others

• :globe_with_meridians: Structure from motion using full spherical panoramic cameras [ICCVW 2011] []

• PoseNet: A Convolutional Network for Real-Time 6-DOF Camera Relocalization [ICCV 2015] [PoseNet]

• Geometric loss functions for camera pose regression with deep learning [CVPR 2017] []

• Relative Camera Pose Estimation Using Convolutional Neural Networks [ACIVS 2017] [relativeCameraPose]

• DSAC - Differentiable RANSAC for Camera Localization [CVPR 2017] [DSAC]

• Generalized Differentiable RANSAC [arXiv 2022] [differentiable_ransac]

• RPNet: an End-to-End Network for Relative Camera Pose Estimation [ECCVW 2018] [RPNet]

• Camera relocalization by computing pairwise relative poses using convolutional neural network [ICCVW 2017] [RelPoseNet]

• Deep Keypoint-Based Camera Pose Estimation with Geometric Constraints [IROS 2020] [pytorch-deepFEPE]

• Wide-Baseline Relative Camera Pose Estimation with Directional Learning [CVPR 2021] [DirectionNet]

• Learning single and multi-scene camera pose regression with transformer encoders [Computer Vision and Image Understanding 2024] [transposenet]

• :globe_with_meridians: Robust 360-8PA: Redesigning The Normalized 8-point Algorithm for 360-FoV Images [ICRA 2021] [robust_360_8PA]

• :globe_with_meridians: Pose Estimation for Two-View Panoramas: a Comparative Analysis [CVPRW 2022] [Keypoints]

• The 8-Point Algorithm as an Inductive Bias for Relative Pose Prediction by ViTs [3DV 2022] [rel_pose]

• :globe_with_meridians: CoVisPose: Co-visibility Pose Transformer for Wide-Baseline Relative Pose Estimation in 360 Indoor Panoramas [ECCV 2022] []

• Map-free Visual Relocalization: Metric Pose Relative to a Single Image [ECCV 2022] [map-free-reloc]

• :globe_with_meridians: GPR-Net: Multi-view Layout Estimation via a Geometry-aware Panorama Registration Network [arXiv 2022] []

• RelMobNet: End-to-end relative camera pose estimation using a robust two-stage training [arXiv 2022] []

• GRelPose: Generalizable End-to-End Relative Camera Pose Regression [arXiv 2022] [GRelPose]

• A Lightweight Domain Adaptive Absolute Pose Regressor Using BARLOW TWINS Objective [arXiv 2022] []

• Uncertainty-Driven Dense Two-View Structure from Motion [arXiv 2023] []

• CGA-PoseNet: Camera Pose Regression via a 1D-Up Approach to Conformal Geometric Algebra [arXiv 2023] []

• :globe_with_meridians: Graph-CoVis: GNN-based Multi-view Panorama Global Pose Estimation [arXiv 2023] []

• Map-Relative Pose Regression for Visual Re-Localization [CVPR 2024] [marepo]

• SRPose: Two-view Relative Pose Estimation With Sparse Keypoints [ECCV 2024] [SRPose]

---

Similar images disambiguate

• Doppelgangers: Learning to Disambiguate Images of Similar Structures [ICCV 2023] [Doppelgangers]

---

Datasets

• HPatches
• YFCC100M
• MegaDepth
• ScanNet
• :globe_with_meridians: Matterport3D
• :globe_with_meridians: Zillow Indoor Dataset (ZInD)
• :globe_with_meridians: SphereCraft: A Dataset for Spherical Keypoint Detection, Matching and Camera Pose Estimation

---

Challenges and workshops

• Image Matching Challenge 2024
• Image Matching Challenge 2023
• Image Matching Challenge 2022
• Image Matching Challenge 2021
• Image Matching Challenge 2020
• Image Matching Challenge 2019
• Image Matching: Local Features and Beyond workshop at CVPR
• :globe_with_meridians: Omnidirectional Computer Vision workshop at CVPR

---

Resources and toolboxes

• image-matching-webui

• deep-image-matching

• image-matching-toolbox

---

Format:

• Title [journal year] [repo]