Research map / Point clouds / Registration and matching

$\pi^2$: A Simple Framework for 2D-to-3D Registration

NeurIPS 2026 · Point clouds › Registration and matching · approach: Non-rigid point set registration

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes π², an image-to-point-cloud registration method that directly regresses camera-centered coordinates for each LiDAR point instead of using intermediate correspondences or proxies.

Problem

Image-to-point-cloud (I2P) registration, i.e., estimating the camera pose of a target image relative to an unaligned LiDAR point cloud, which existing methods typically tackle via intermediate targets such as 2D-3D correspondences, matching confidence, visibility masks, or overlap regions.

Approach

Reformulates I2P registration as point-wise camera-centered coordinate regression: each point of the unaligned LiDAR cloud is regressed directly into the camera frame from the target image, avoiding intermediate correspondence, confidence, visibility, or overlap targets.

Closest papers by meaning

• GraphI2P: Image-to-Point Cloud Registration with Exploring Pattern of Correspondence via Graph Learning (CVPR 2025 · similarity 0.89)
• 3D Focusing-and-Matching Network for Multi-Instance Point Cloud Registration (NeurIPS 2024 · similarity 0.89)
• Implicit Correspondence Learning for Image-to-Point Cloud Registration (CVPR 2025 · similarity 0.88)
• Partial Point Cloud Registration with Multi-view 2D Image Learning (AAAI 2025 · similarity 0.88)
• Mining and Transferring Feature-Geometry Coherence for Unsupervised Point Cloud Registration (NeurIPS 2024 · similarity 0.87)
• Rectified Point Flow: Generic Point Cloud Pose Estimation (NeurIPS 2025 · similarity 0.87)