arXiv 2601.027592026-01-06cs.CVcs.RO

跨多样尺度、场景和传感器设置的零样本点云配准

Towards Zero-Shot Point Cloud Registration Across Diverse Scales, Scenes, and Sensor Setups

Laboratory for Information and Decision Systems (LIDS), Massachusetts Institute of Technology（信息与决策系统实验室，麻省理工学院）；Department of Computer Science and Engineering, Seoul National University（计算机科学与工程系，首尔国立大学）

Hyungtae Lim, Minkyun Seo, Luca Carlone, Jaesik Park

备注：: 18 pages, 15 figures. Extended version of our ICCV 2025 highlight paper [arXiv:2503.07940]. arXiv admin note: substantial text overlap with arXiv:2503.07940

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

提出BUFFER-X框架，通过几何自举、分布感知采样和坐标归一化实现零样本点云配准，同时引入BUFFER-X-Lite提升效率，适用于多样场景和传感器设置。

摘要：

一些基于深度学习的点云配准方法在零样本泛化方面存在困难，通常需要针对特定数据集进行超参数调整或重新训练以适应新环境。我们识别出三个关键限制：（a）固定的用户定义参数（例如体素大小、搜索半径）无法在不同尺度之间泛化；（b）学习的关键点检测器在跨域转移中表现不佳；（c）绝对坐标加剧了不同数据集之间的尺度不匹配。为了解决这三个问题，我们提出了BUFFER-X，一种无需训练的配准框架，通过：（a）几何自举法实现自动超参数估计；（b）分布感知的最远点采样代替学习检测器；（c）片段级坐标归一化确保尺度一致性。我们的方法采用分层多尺度匹配，在局部、中等和全局感受野中提取对应关系，使在多样化环境中实现稳健配准。对于效率关键的应用，我们引入BUFFER-X-Lite，通过早退出策略和快速姿态求解器将总计算时间减少43%（相对于BUFFER-X），同时保持准确性。我们在涵盖12个数据集的全面基准上进行评估，这些数据集覆盖了物体尺度、室内和室外场景，包括异构激光雷达配置之间的跨传感器配准。结果表明，我们的方法在无需手动调整或先验知识的情况下能够有效泛化。代码：https://github.com/MIT-SPARK/BUFFER-X。

英文摘要：

Some deep learning-based point cloud registration methods struggle with zero-shot generalization, often requiring dataset-specific hyperparameter tuning or retraining for new environments. We identify three critical limitations: (a) fixed user-defined parameters (e.g., voxel size, search radius) that fail to generalize across varying scales, (b) learned keypoint detectors exhibit poor cross-domain transferability, and (c) absolute coordinates amplify scale mismatches between datasets. To address these three issues, we present BUFFER-X, a training-free registration framework that achieves zero-shot generalization through: (a) geometric bootstrapping for automatic hyperparameter estimation, (b) distribution-aware farthest point sampling to replace learned detectors, and (c) patch-level coordinate normalization to ensure scale consistency. Our approach employs hierarchical multi-scale matching to extract correspondences across local, middle, and global receptive fields, enabling robust registration in diverse environments. For efficiency-critical applications, we introduce BUFFER-X-Lite, which reduces total computation time by 43% (relative to BUFFER-X) through early exit strategies and fast pose solvers while preserving accuracy. We evaluate on a comprehensive benchmark comprising 12 datasets spanning object-scale, indoor, and outdoor scenes, including cross-sensor registration between heterogeneous LiDAR configurations. Results demonstrate that our approach generalizes effectively without manual tuning or prior knowledge of test domains. Code: https://github.com/MIT-SPARK/BUFFER-X.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。