mlsanityData to 7 Oct 2026

EDM: Equirectangular Projection-Oriented Dense Kernelized Feature Matching

arXiv 2502.20685 ↗First posted 28 Feb 2025cs.CVcited by 7

Not on the map. Why?

Abstract (arXiv)

We introduce the first learning-based dense matching algorithm, termed Equirectangular Projection-Oriented Dense Kernelized Feature Matching (EDM), specifically designed for omnidirectional images. Equirectangular projection (ERP) images, with their large fields of view, are particularly suited for dense matching techniques that aim to establish comprehensive correspondences across images. However, ERP images are subject to significant distortions, which we address by leveraging the spherical camera model and geodesic flow refinement in the dense matching method. To further mitigate these distortions, we propose spherical positional embeddings based on 3D Cartesian coordinates of the feature grid. Additionally, our method incorporates bidirectional transformations between spherical and Cartesian coordinate systems during refinement, utilizing a unit sphere to improve matching performance. We demonstrate that our proposed method achieves notable performance enhancements, with improvements of +26.72 and +42.62 in AUC@5deg on the Matterport3D and Stanford2D3D datasets.

Key facts. Cited by 7 papers in the map (data through 7 Oct 2026). Builds on 42 of its 74 references that are in the map. First posted on arXiv 28 Feb 2025.

Builds on

42 of 74 references.

References · 42

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,764
  3. 2019SuperGlue: Learning Feature Matching with Graph Neural Networkscited by 1,437
  4. 2017Matterport3D: Learning from RGB-D Data in Indoor Environmentscited by 1,354
  5. 2017SuperPoint: Self-Supervised Interest Point Detection and Descriptioncited by 1,313
  6. 2021LoFTR: Detector-Free Local Feature Matching with Transformerscited by 974
  7. 2018Spherical CNNscited by 452
  8. 2017Joint 2D-3D-Semantic Data for Indoor Scene Understandingcited by 404
  9. 2020DISK: Learning local features with policy gradientcited by 310
  10. 2017Learning SO(3) Equivariant Representations with Spherical CNNscited by 244
  11. 2022DKM: Dense Kernelized Feature Matching for Geometry Estimationcited by 174
  12. 2022CroCo: Self-Supervised Pre-training for 3D Vision Tasks by Cross-View Completioncited by 158
30 more
  1. 2019GLU-Net: Global-Local Universal Network for Dense Flow and Correspondencescited by 139
  2. 2020Patch2Pix: Epipolar-Guided Pixel-Level Correspondencescited by 128
  3. 2018Learning a Discriminative Feature Network for Semantic Segmentationcited by 114
  4. 2021Learning Accurate Dense Correspondences and When to Trust Themcited by 113
  5. 2020HoHoNet: 360 Indoor Holistic Understanding with Latent Horizontal Featurescited by 104
  6. 2018OmniDepth: Dense Depth Estimation for Indoors Spherical Panoramascited by 98
  7. 2022PanoFormer: Panorama Transformer for Indoor 360 Depth Estimationcited by 94
  8. 2021UniFuse: Unidirectional Fusion for 360circ Panorama Depth Estimationcited by 91
  9. 2019Tangent Images for Mitigating Spherical Distortioncited by 79
  10. 2019Spherical CNNs on Unstructured Gridscited by 76
  11. 2018DGC-Net: Dense Geometric Correspondence Networkcited by 67
  12. 2021360MonoDepth: High-Resolution 360deg Monocular Depth Estimationcited by 60
  13. 2022OmniFusion: 360 Monocular Depth Estimation via Geometry-Aware Fusioncited by 60
  14. 2019Orientation-aware Semantic Segmentation on Icosahedron Spherescited by 52
  15. 2019GIFT: Learning Transformation-Invariant Dense Visual Descriptors via Group CNNscited by 47
  16. 2023Balanced Spherical Grid for Egocentric View Synthesiscited by 29
  17. 2022Decoupling Makes Weakly Supervised Local Feature Bettercited by 28
  18. 2024FastScene: Text-Driven Fast 3D Indoor Scene Generation via Panoramic Gaussian Splattingcited by 24
  19. 2021Improving 360 Monocular Depth Estimation via Non-local Dense Prediction Transformer and Joint Supervised and Self-supervised Learningcited by 19
  20. 2020OmniSLAM: Omnidirectional Localization and Dense Mapping for Wide-baseline Multi-camera Systemscited by 18
  21. 2022ECO-TR: Efficient Correspondences Finding Via Coarse-to-Fine Refinementcited by 18
  22. 2019What's in my Room? Object Recognition on Indoor Panoramic Imagescited by 16
  23. 20243D Scene Geometry Estimation from 360^circ Imagery: A Surveycited by 11
  24. 2022PanoDepth: A Two-Stage Approach for Monocular Omnidirectional Depth Estimationcited by 9
  25. 2022Panoramic Video Salient Object Detection with Ambisonic Audio Guidancecited by 9
  26. 2023Graph-CoVis: GNN-based Multi-view Panorama Global Pose Estimationcited by 8
  27. 2024OmniSDF: Scene Reconstruction using Omnidirectional Signed Distance Functions and Adaptive Binoctreescited by 6
  28. 2024OmniLocalRF: Omnidirectional Local Radiance Fields from Dynamic Videoscited by 5
  29. 2021Robust 360-8PA: Redesigning The Normalized 8-point Algorithm for 360-FoV Imagescited by 3
  30. 2022Direct Triangulation with Spherical Projection for Omnidirectional Camerascited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 1; 2026: 6.

7 papers.

  1. 2025PANORAMA: The Rise of Omnidirectional Vision in the Embodied AI Eracited by 12
  2. 2026Wid3R: Wide Field-of-View 3D Reconstruction via Camera Model Conditioningcited by 3
  3. 2026360deg Image Perception with MLLMs: A Comprehensive Benchmark and a Training-Free Methodcited by 2
  4. 2026SO3UFormer: Learning Intrinsic Spherical Features for Rotation-Robust Panoramic Dense Predictioncited by 1
  5. 2026OmniDS: Dual-Stream Context Fusion for Omnidirectional Depth from Fisheye Camerascited by 0
  6. 2026SCCM: Spherically Consistent Coarse Matching for ERP Dense Feature Correspondencecited by 0
  7. 2026EPIC: Epipolar-Consistent 360deg Immersive Stereo Video Generationcited by 0