mlsanityData to 7 Oct 2026

GLASS: Geometry-aware Local Alignment and Structure Synchronization Network for 2D-3D Registration

arXiv 2603.26262 ↗First posted 27 Mar 2026cs.CV cs.LGcited by 1

Not on the map. Why?

Abstract (arXiv)

Image-to-point cloud registration methods typically follow a coarse-to-fine pipeline, extracting patch-level correspondences and refining them into dense pixel-to-point matches. However, in scenes with repetitive patterns, images often lack sufficient 3D structural cues and alignment with point clouds, leading to incorrect matches. Moreover, prior methods usually overlook structural consistency, limiting the full exploitation of correspondences. To address these issues, we propose two novel modules: the Local Geometry Enhancement (LGE) module and the Graph Distribution Consistency (GDC) module. LGE enhances both image and point cloud features with normal vectors, injecting geometric structure into image features to reduce mismatches. GDC constructs a graph from matched points to update features and explicitly constrain similarity distributions. Extensive experiments and ablations on two benchmarks, RGB-D Scenes v2 and 7-Scenes, demonstrate that our approach achieves state-of-the-art performance in image-to-point cloud registration.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 44 of its 95 references that are in the map. First posted on arXiv 27 Mar 2026.

Builds on

44 of 95 references.

References · 44

  1. 2017Attention Is All You Needcited by 59,932
  2. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  3. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,282
  4. 2017Graph Attention Networkscited by 8,854
  5. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,764
  6. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,223
  7. 2016Feature Pyramid Networks for Object Detectioncited by 5,021
  8. 2014Unsupervised Domain Adaptation by Backpropagationcited by 1,906
  9. 2019SuperGlue: Learning Feature Matching with Graph Neural Networkscited by 1,437
  10. 2024Depth Anything V2cited by 1,385
  11. 2019KPConv: Flexible and Deformable Convolution for Point Cloudscited by 1,202
  12. 2021LoFTR: Detector-Free Local Feature Matching with Transformerscited by 974
32 more
  1. 2020Circle Loss: A Unified Perspective of Pair Similarity Optimizationcited by 402
  2. 2020Point-GNN: Graph Neural Network for 3D Object Detection in a Point Cloudcited by 335
  3. 2020PREDATOR: Registration of 3D Point Clouds with Low Overlapcited by 287
  4. 2018SO-Net: Self-Organizing Network for Point Cloud Analysiscited by 210
  5. 2022Geometric Transformer for Fast and Robust Point Cloud Registrationcited by 200
  6. 2021CoFiNet: Reliable Coarse-to-fine Correspondences for Robust Point Cloud Registrationcited by 151
  7. 2024Efficient LoFTR: Semi-Dense Local Feature Matching with Sparse-Like Speedcited by 119
  8. 2020Human-centric Spatio-Temporal Video Grounding With Visual Transformerscited by 83
  9. 2023GeoTransformer: Fast and Robust Point Cloud Registration with Geometric Transformercited by 83
  10. 20192D3D-MatchNet: Learning to Match Keypoints Across 2D Image and 3D Point Cloudcited by 62
  11. 2021DeepI2P: Image-to-Point Cloud Registration via Deep Classificationcited by 58
  12. 2020DeepFit: 3D Surface Fitting via Neural Network Weighted Least Squarescited by 43
  13. 2022CorrI2P: Deep Image-to-Point Cloud Registration via Dense Correspondencecited by 43
  14. 2021P2-Net: Joint Description and Detection of Local Features for Pixel and Point Matchingcited by 42
  15. 2023Differentiable Registration of Images and LiDAR Point Clouds with VoxelPoint-to-Pixel Matchingcited by 35
  16. 2021AdaFit: Rethinking Learning-based Normal Estimation on Point Cloudscited by 29
  17. 2023FreeReg: Image-to-Point Cloud Registration Leveraging Pretrained Diffusion Models and Monocular Depth Estimatorscited by 29
  18. 20232D3D-MATR: 2D-3D Matching Transformer for Detection-free Registration between Images and Point Cloudscited by 27
  19. 2022GraphFit: Learning Multi-scale Graph-Convolutional Representation for Point Cloud Normal Estimationcited by 20
  20. 2023SE-ORNet: Self-Ensembling Orientation-aware Network for Unsupervised Point Cloud Shape Correspondencecited by 19
  21. 2024EF-3DGS: Event-Aided Free-Trajectory 3D Gaussian Splattingcited by 15
  22. 2025Bridge 2D-3D: Uncertainty-aware Hierarchical Registration Network with Domain Alignmentcited by 14
  23. 2023Neural Interactive Keypoint Detectioncited by 12
  24. 2023CoFiI2P: Coarse-to-Fine Correspondences for Image-to-Point Cloud Registrationcited by 12
  25. 20213D Hierarchical Refinement and Augmentation for Unsupervised Learning of Depth and Pose from Monocular Videocited by 11
  26. 2023End-to-end 2D-3D Registration between Image and LiDAR Point Cloud for Vehicle Localizationcited by 9
  27. 2025CA-I2P: Channel-Adaptive Registration Network with Global Optimal Selectioncited by 8
  28. 2022CPPF++: Uncertainty-Aware Sim2Real Object Pose Estimation by Vote Aggregationcited by 6
  29. 2025Adaptive Agent Selection and Interaction Network for Image-to-point cloud Registrationcited by 6
  30. 2023Quantity-Aware Coarse-to-Fine Correspondence for Image-to-Point Cloud Registrationcited by 4
  31. 2025EdgeRegNet: Edge Feature-based Multimodal Registration Network between Images and LiDAR Point Cloudscited by 3
  32. 2026VCR: Variance-Driven Channel Recalibration for Robust Low-Light Enhancementcited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026FS-I2P:A Hierarchical Focus-Sweep Registration Network with Dynamically Allocated Depthcited by 0