A Pseudo Global Fusion Paradigm-Based Cross-View Network for LiDAR-Based Place Recognition

2025-08-12

AI Summary

In cross-view LiDAR point cloud place recognition, conventional metric learning methods neglect the intrinsic structure of feature space and intra-class variability, leading to degraded performance under complex and time-varying conditions. To address this, we propose a pseudo-global fusion-driven multimodal collaborative learning framework. Our key contributions are: (1) a pseudo-global information guidance mechanism that establishes a unified cross-view semantic representation; (2) a manifold-adaptive pairwise variance–locality metric learning scheme, which employs Symmetric Positive Definite (SPD) matrices to model Mahalanobis distance, thereby overcoming Euclidean distance’s limitations in capturing nonlinear distributions; and (3) robust localization via multimodal branch integration and higher-order geometric-aware feature fusion. Experiments on multiple benchmark datasets demonstrate significant improvements in recognition accuracy and generalization capability—particularly under challenging conditions including occlusion, illumination variation, and seasonal changes—outperforming existing state-of-the-art methods.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Security and Privacy: Large-scale security measurements

Abstract

LiDAR-based Place Recognition (LPR) remains a critical task in Embodied Artificial Intelligence (AI) and Autonomous Driving, primarily addressing localization challenges in GPS-denied environments and supporting loop closure detection. Existing approaches reduce place recognition to a Euclidean distance-based metric learning task, neglecting the feature space's intrinsic structures and intra-class variances. Such Euclidean-centric formulation inherently limits the model's capacity to capture nonlinear data distributions, leading to suboptimal performance in complex environments and temporal-varying scenarios. To address these challenges, we propose a novel cross-view network based on an innovative fusion paradigm. Our framework introduces a pseudo-global information guidance mechanism that coordinates multi-modal branches to perform feature learning within a unified semantic space. Concurrently, we propose a Manifold Adaptation and Pairwise Variance-Locality Learning Metric that constructs a Symmetric Positive Definite (SPD) matrix to compute Mahalanobis distance, superseding traditional Euclidean distance metrics. This geometric formulation enables the model to accurately characterize intrinsic data distributions and capture complex inter-class dependencies within the feature space. Experimental results demonstrate that the proposed algorithm achieves competitive performance, particularly excelling in complex environmental conditions.

Problem

• Addresses LiDAR place recognition in GPS-denied environments
• Overcomes limitations of Euclidean distance-based metric learning
• Captures nonlinear data distributions in complex environments

Innovation

• Pseudo-global fusion for multi-modal feature learning
• Manifold Adaptation and Pairwise Variance-Locality Learning
• SPD matrix for Mahalanobis distance computation