mlsanityData to 13 Oct 2026

CRISTAL: Real-time Camera Registration in Static LiDAR Scans using Neural Rendering

arXiv 2511.16349 ↗First posted 20 Nov 2025cs.CV cs.GRcited by 0

Not on the map. Why?

Abstract (arXiv)

Accurate camera localization is crucial for robotics and Extended Reality (XR), enabling reliable navigation and alignment of virtual and real content. Existing visual methods often suffer from drift, scale ambiguity, and depend on fiducials or loop closure. This work introduces a real-time method for localizing a camera within a pre-captured, highly accurate colored LiDAR point cloud. By rendering synthetic views from this cloud, 2D-3D correspondences are established between live frames and the point cloud. A neural rendering technique narrows the domain gap between synthetic and real images, reducing occlusion and background artifacts to improve feature matching. The result is drift-free camera tracking with correct metric scale in the global LiDAR coordinate system. Two real-time variants are presented: Online Render and Match, and Prebuild and Localize. We demonstrate improved results on the ScanNet++ dataset and outperform existing SLAM pipelines.

Key facts. No citing papers in the map yet (data through 13 Oct 2026). Builds on 23 of its 44 references that are in the map. First posted on arXiv 20 Nov 2025.

Builds on

23 of 44 references.

References · 23

  1. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,656
  2. 2023ScanNet++: A High-Fidelity Dataset of 3D Indoor Scenescited by 752
  3. 2024XFeat: Accelerated Features for Lightweight Image Matchingcited by 113
  4. 2019OpenVSLAM: A Versatile Visual SLAM Frameworkcited by 64
  5. 20192D3D-MatchNet: Learning to Match Keypoints Across 2D Image and 3D Point Cloudcited by 62
  6. 2019Neural Point Cloud Rendering via Multi-Plane Projectioncited by 60
  7. 2021DeepI2P: Image-to-Point Cloud Registration via Deep Classificationcited by 59
  8. 2025WildGS-SLAM: Monocular Gaussian Splatting SLAM in Dynamic Environmentscited by 48
  9. 2022CorrI2P: Deep Image-to-Point Cloud Registration via Dense Correspondencecited by 44
  10. 2021P2-Net: Joint Description and Detection of Local Features for Pixel and Point Matchingcited by 42
  11. 2024FAST-LIVO2: Fast, Direct LiDAR-Inertial-Visual Odometrycited by 41
  12. 2023FreeReg: Image-to-Point Cloud Registration Leveraging Pretrained Diffusion Models and Monocular Depth Estimatorscited by 31
11 more
  1. 2019LCD: Learned Cross-Domain Descriptors for 2D-3D Matchingcited by 29
  2. 2019CMRNet: Camera to LiDAR-Map Registrationcited by 25
  3. 2020Monocular Camera Localization in Prior LiDAR Maps with 2D-3D Line Correspondencescited by 23
  4. 2023Pointersect: Neural Rendering with Cloud-Ray Intersectioncited by 23
  5. 2020CMRNet++: Map and Camera Agnostic Monocular Visual Localization in LiDAR Mapscited by 14
  6. 2021Rendering Point Clouds with Compute Shaders and Vertex Order Optimizationcited by 14
  7. 2023Point2Pix: Photo-Realistic Point Cloud Rendering via Neural Radiance Fieldscited by 12
  8. 2022CAD2Render: A Modular Toolkit for GPU-accelerated Photorealistic Synthetic Data Generation for the Manufacturing Industrycited by 10
  9. 2025Diff2I2P: Differentiable Image-to-Point Cloud Registration with Diffusion Priorcited by 8
  10. 2023Colmap-PCD: An Open-source Tool for Fine Image-to-point cloud Registrationcited by 4
  11. 2025Real-time Neural Rendering of LiDAR Point Cloudscited by 1

Built on by

Not cited yet.