TurboReg: TurboClique for Robust and Efficient Point Cloud Registration

2025-07-02

AI Summary

To address the low efficiency of robust estimation in correspondence-based point cloud registration (PCR), this paper proposes TurboClique—a lightweight 3-clique compatibility graph structure—along with a linear-time Pivot-Guided parallel clique search algorithm. By leveraging high SC²-scoring correspondences as pivots, the method enables efficient pruning and parallelized maximal clique enumeration, significantly improving inlier recall and computational speed. Evaluated on real-world datasets including 3DMatch with FCGF features, TurboClique achieves a 208.22× speedup over 3DMAC while attaining higher inlier recall. The approach thus strikes an effective balance between robustness and real-time performance, establishing a new paradigm for efficient PCR.

Technology Category

• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Search and Optimization: Sampling/Simulation-based Search
• Intelligent Robots: State Estimation

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Security and Privacy: Large-scale security measurements

Abstract

Robust estimation is essential in correspondence-based Point Cloud Registration (PCR). Existing methods using maximal clique search in compatibility graphs achieve high recall but suffer from exponential time complexity, limiting their use in time-sensitive applications. To address this challenge, we propose a fast and robust estimator, TurboReg, built upon a novel lightweight clique, TurboClique, and a highly parallelizable Pivot-Guided Search (PGS) algorithm. First, we define the TurboClique as a 3-clique within a highly-constrained compatibility graph. The lightweight nature of the 3-clique allows for efficient parallel searching, and the highly-constrained compatibility graph ensures robust spatial consistency for stable transformation estimation. Next, PGS selects matching pairs with high SC$^2$ scores as pivots, effectively guiding the search toward TurboCliques with higher inlier ratios. Moreover, the PGS algorithm has linear time complexity and is significantly more efficient than the maximal clique search with exponential time complexity. Extensive experiments show that TurboReg achieves state-of-the-art performance across multiple real-world datasets, with substantial speed improvements. For example, on the 3DMatch+FCGF dataset, TurboReg (1K) operates $208.22 imes$ faster than 3DMAC while also achieving higher recall. Our code is accessible at href{https://github.com/Laka-3DV/TurboReg}{ exttt{TurboReg}}.

Problem

• Overcome exponential time complexity in point cloud registration
• Improve robustness and efficiency in correspondence-based PCR
• Achieve fast, parallelizable search with high inlier ratios

Innovation

• TurboClique for efficient parallel searching
• Pivot-Guided Search with linear complexity
• Highly-constrained compatibility graph ensures robustness