mlsanityData to 7 Oct 2026

GS-CPR: Efficient Camera Pose Refinement via 3D Gaussian Splatting

arXiv 2408.11085 ↗First posted 20 Aug 2024cs.CVcited by 31

Not on the map. Why?

Abstract (arXiv)

We leverage 3D Gaussian Splatting (3DGS) as a scene representation and propose a novel test-time camera pose refinement (CPR) framework, GS-CPR. This framework enhances the localization accuracy of state-of-the-art absolute pose regression and scene coordinate regression methods. The 3DGS model renders high-quality synthetic images and depth maps to facilitate the establishment of 2D-3D correspondences. GS-CPR obviates the need for training feature extractors or descriptors by operating directly on RGB images, utilizing the 3D foundation model, MASt3R, for precise 2D matching. To improve the robustness of our model in challenging outdoor environments, we incorporate an exposure-adaptive module within the 3DGS framework. Consequently, GS-CPR enables efficient one-shot pose refinement given a single RGB query and a coarse initial pose estimation. Our proposed approach surpasses leading NeRF-based optimization methods in both accuracy and runtime across indoor and outdoor visual localization benchmarks, achieving new state-of-the-art accuracy on two indoor datasets. The project page is available at https://xrim-lab.github.io/GS-CPR/.

Key facts. Cited by 31 papers in the map (data through 7 Oct 2026). Builds on 45 of its 60 references that are in the map. First posted on arXiv 20 Aug 2024.

Builds on

45 of 60 references.

References · 45

  1. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  2. 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,141
  3. 2021Masked-attention Mask Transformer for Universal Image Segmentationcited by 2,338
  4. 2023DUSt3R: Geometric 3D Vision Made Easycited by 1,526
  5. 2019SuperGlue: Learning Feature Matching with Graph Neural Networkscited by 1,437
  6. 2017SuperPoint: Self-Supervised Interest Point Detection and Descriptioncited by 1,313
  7. 2021LoFTR: Detector-Free Local Feature Matching with Transformerscited by 974
  8. 2024Grounding Image Matching in 3D with MASt3Rcited by 960
  9. 2015NetVLAD: CNN architecture for weakly supervised place recognitioncited by 935
  10. 2023Scaffold-GS: Structured 3D Gaussians for View-Adaptive Renderingcited by 684
  11. 2023LightGlue: Local Feature Matching at Light Speedcited by 583
  12. 2018From Coarse to Fine: Robust Hierarchical Localization at Large Scalecited by 556
33 more
  1. 2020INeRF: Inverting Neural Radiance Fields for Pose Estimationcited by 381
  2. 2019D2-Net: A Trainable CNN for Joint Detection and Description of Local Featurescited by 324
  3. 2016DSAC - Differentiable RANSAC for Camera Localizationcited by 264
  4. 2016Large-Scale Image Retrieval with Attentive Deep Local Featurescited by 238
  5. 2017Geometric Loss Functions for Camera Pose Regression with Deep Learningcited by 224
  6. 2019Understanding the Limitations of CNN-based Absolute Camera Pose Regressioncited by 177
  7. 2020Visual Camera Re-Localization from RGB and RGB-D Images Using DSACcited by 176
  8. 2015Modelling Uncertainty in Deep Learning for Camera Relocalizationcited by 133
  9. 2016End-to-end Learning of Deep Visual Representations for Image Retrievalcited by 122
  10. 2021LENS: Localization enhanced by NeRF synthesiscited by 116
  11. 2020Self-supervising Fine-grained Region Similarities for Large-scale Image Localizationcited by 110
  12. 2021Learning Multi-Scene Absolute Pose Regression with Transformerscited by 108
  13. 2023Accelerated Coordinate Encoding: Learning to Relocalize in Minutes using RGB and Posescited by 107
  14. 2022LaMAR: Benchmarking Localization and Mapping for Augmented Realitycited by 98
  15. 2019AtLoc: Attention Guided Camera Localizationcited by 92
  16. 2022DFNet: Enhance Absolute Pose Regression with Direct Feature Matchingcited by 82
  17. 2021On the Limits of Pseudo Ground Truth in Visual Camera Re-localisationcited by 63
  18. 2016Learning to Navigate the Energy Landscapecited by 62
  19. 2021Direct-PoseNet: Absolute Pose Regression with Photometric Consistencycited by 61
  20. 2022Parallel Inversion of Neural Radiance Fields for Robust Pose Estimationcited by 55
  21. 2024GLACE: Global Local Accelerated Coordinate Encodingcited by 48
  22. 2023CROSSFIRE: Camera Relocalization On Self-Supervised Features from an Implicit Representationcited by 45
  23. 2024Map-Relative Pose Regression for Visual Re-Localizationcited by 42
  24. 2023Neural Refinement for Absolute Pose Regression with Feature Synthesiscited by 37
  25. 2024The NeRFect Match: Exploring NeRF Features for Visual Localizationcited by 32
  26. 2023NeRF-Loc: Visual Localization with Conditional Neural Radiance Fieldcited by 29
  27. 20246DGS: 6D Pose Estimation from a Single Image and a 3D Gaussian Splatting Modelcited by 28
  28. 2022Investigating the Role of Image Retrieval for Visual Localization -- An exhaustive benchmarkcited by 27
  29. 2023PNeRFLoc: Visual Localization with Point-based Neural Radiance Fieldscited by 22
  30. 2024The Unreasonable Effectiveness of Pre-Trained Features for Camera Pose Refinementcited by 21
  31. 2024HR-APR: APR-agnostic Framework with Uncertainty Estimation and Hierarchical Refinement for Camera Relocalisationcited by 16
  32. 2024Learning Neural Volumetric Pose Features for Camera Localizationcited by 11
  33. 2024AIR-HLoc: Adaptive Retrieved Images Selection for Efficient Visual Localisationcited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 3; 2025: 15; 2026: 13.

31 papers.

Sort:

  1. 2025What Is The Best 3D Scene Representation for Robotics? From Geometric to Foundation Modelscited by 29
  2. 2024GSplatLoc: Grounding Keypoint Descriptors into 3D Gaussian Splatting for Improved Visual Localizationcited by 15
  3. 2025From Sparse to Dense: Camera Relocalization with Scene-Specific Detector from Feature Gaussian Splattingcited by 13
  4. 2025Gaussian Splatting Feature Fields for Privacy-Preserving Visual Localizationcited by 10
  5. 2025A Guide to Structureless Visual Localizationcited by 9
  6. 20253DGEER: 3D Gaussian Rendering Made Exact and Efficient for Generic Camerascited by 8
  7. 2025Digital Twin Generation from Visual Data: A Surveycited by 7
  8. 2025GSVisLoc: Generalizable Visual Localization for Gaussian Splatting Scene Representationscited by 5
  9. 2024Adversarial Exploitation of Data Diversity Improves Visual Localizationcited by 4
  10. 2025AnimateScene: Camera-controllable Animation in Any Scenecited by 3
  11. 2025SmallGS: Gaussian Splatting-based Camera Pose Estimation for Small-Baseline Videoscited by 2
  12. 2025Coca-Splat: Collaborative Optimization for Camera Parameters and 3D Gaussianscited by 1
19 more
  1. 2025A Constrained Optimization Approach for Gaussian Splatting from Coarsely-posed Images and Noisy Lidar Point Cloudscited by 1
  2. 2025Camera Pose Refinement via 3D Gaussian Splattingcited by 1
  3. 2025RayGaussX: Accelerating Gaussian-Based Ray Marching for Real-Time and High-Quality Novel View Synthesiscited by 1
  4. 2025GRLoc: Geometric Representation Regression for Visual Localizationcited by 1
  5. 2026ImLoc: Revisiting Visual Localization with Image-based Representationcited by 1
  6. 2026Rethinking Pose Refinement in 3D Gaussian Splatting under Pose Prior and Geometric Uncertaintycited by 1
  7. 2026DRoPS: Dynamic 3D Reconstruction of Pre-Scanned Objectscited by 1
  8. 2024Image Compression Using Novel View Synthesis Priorscited by 0
  9. 2025Gaussian Alignment for Relative Camera Pose Estimation via Single-View Reconstructioncited by 0
  10. 2026L3:Scene-agnostic Visual Localization in the Wildcited by 0
  11. 2026PlanaReLoc: Camera Relocalization in 3D Planar Primitives via Region-Based Structure Matchingcited by 0
  12. 2026Hierarchical Visual Relocalization with Nearest View Synthesis from Feature Gaussian Splattingcited by 0
  13. 2026LSGS-Loc: Towards Robust 3DGS-Based Visual Localization for Large-Scale UAV Scenarioscited by 0
  14. 2026Disambiguating 2D-3D Correspondences in Gaussian Splatting-based Feature Fields for Visual Localizationcited by 0
  15. 2026SG2Loc: Sequential Visual Localization on 3D Scene Graphscited by 0
  16. 2026Revisiting Pose Sensitivity in Splat-based Computed Tomography under Sparse-view Reconstructioncited by 0
  17. 2026G2ARD-GS: Geometry-Guided Anchor-Regularized Gaussian Splatting Distillationcited by 0
  18. 2026GS-CPE: Unified 6-Degree-of-Freedom Camera Pose Estimation via 3D Gaussian Splattingcited by 0
  19. 2026LightLoc++: Sensor-Robust Representation Learning for Efficient Outdoor LiDAR Localizationcited by 0