mlsanityData to 7 Oct 2026

Image-to-Point Cloud Registration Made Easy with Rectified Flow-based LiDAR Upsampling

arXiv 2607.14639 ↗First posted 16 Jul 2026cs.RO cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Image-to-Point Cloud Registration (I2P) is essential for integrating camera and LiDAR in perception and autonomous systems, yet the modality gap between images and point clouds makes it difficult to achieve both high accuracy and strong generalization. In this paper, we propose a simple yet effective I2P method that treats LiDAR as an imaging sensor: from a single sparse LiDAR scan, we generate a dense LiDAR intensity image using Conditional Rectified Flow, match it with a camera image using a pre-trained feature matcher, and estimate the 6-DoF relative pose via PnP-RANSAC. The proposed model is pre-trained through a self-supervised image completion task and fine-tuned on a small amount of LiDAR data (neither image-point cloud pairs nor ground-truth sensor poses are required), enabling it to scale to diverse LiDAR and camera configurations. Experiments on the R3LIVE dataset show that the proposed method achieves a mean error of 4.89deg / 1.63 m, outperforming existing methods, while completing a single registration in approximately 0.68 s.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 25 of its 32 references that are in the map. First posted on arXiv 16 Jul 2026.

Builds on

25 of 32 references.

References · 25

  1. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  2. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  3. 2020Denoising Diffusion Implicit Modelscited by 9,655
  4. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flowcited by 3,550
  5. 2023DUSt3R: Geometric 3D Vision Made Easycited by 1,526
  6. 2021Image Super-Resolution via Iterative Refinementcited by 1,506
  7. 2025VGGT: Visual Geometry Grounded Transformercited by 1,475
  8. 2015ORB-SLAM: a Versatile and Accurate Monocular SLAM Systemcited by 983
  9. 2021LoFTR: Detector-Free Local Feature Matching with Transformerscited by 974
  10. 2023LightGlue: Local Feature Matching at Light Speedcited by 583
  11. 2025MapAnything: Universal Feed-Forward Metric 3D Reconstructioncited by 295
  12. 2023ALIKED: A Lighter Keypoint and Descriptor Extraction Network via Deformable Transformationcited by 120
13 more
  1. 2022GSV-Cities: Toward Appropriate Supervised Visual Place Recognitioncited by 92
  2. 20192D3D-MatchNet: Learning to Match Keypoints Across 2D Image and 3D Point Cloudcited by 62
  3. 2021DeepI2P: Image-to-Point Cloud Registration via Deep Classificationcited by 58
  4. 2023LiDAR Data Synthesis with Denoising Diffusion Probabilistic Modelscited by 53
  5. 2024XoFTR: Cross-modal Feature Matching Transformercited by 45
  6. 2022CorrI2P: Deep Image-to-Point Cloud Registration via Dense Correspondencecited by 43
  7. 2023General, Single-shot, Target-less, and Automatic LiDAR-Camera Extrinsic Calibration Toolboxcited by 41
  8. 2021CT-ICP: Real-time Elastic LiDAR Odometry with Loop Closurecited by 38
  9. 2023FreeReg: Image-to-Point Cloud Registration Leveraging Pretrained Diffusion Models and Monocular Depth Estimatorscited by 29
  10. 20232D3D-MATR: 2D-3D Matching Transformer for Detection-free Registration between Images and Point Cloudscited by 27
  11. 2024Fast LiDAR Data Generation with Rectified Flowscited by 13
  12. 2023CoFiI2P: Coarse-to-Fine Correspondences for Image-to-Point Cloud Registrationcited by 12
  13. 2024GLIM: 3D Range-Inertial Localization and Mapping with GPU-Accelerated Scan Matching Factorscited by 7

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026DPA-I2P: Depth-Guided Projective Alignment for Image-to-Point-Cloud Registration in Autonomous Driving
  2. 2025Register Any Point: Scaling 3D Point Cloud Registration by Flow Matching
  3. 2024EADReg: Probabilistic Correspondence Generation with Efficient Autoregressive Diffusion Model for Outdoor Point Cloud Registration
  4. 2025Efficient Point Clouds Upsampling via Flow Matching
  5. 2025LiNeXt: Revisiting LiDAR Completion with Efficient Non-Diffusion Architectures
  6. 2025PVNet: Point-Voxel Interaction LiDAR Scene Upsampling Via Diffusion Models
  7. 2026Towards Zero-Shot Point Cloud Registration Across Diverse Scales, Scenes, and Sensor Setups
  8. 2025Diff2I2P: Differentiable Image-to-Point Cloud Registration with Diffusion Prior