LifelongPR: Lifelong knowledge fusion for point cloud place recognition based on replay and prompt learning

2025-07-14

AI Summary

Point cloud place recognition (PCPR) suffers from catastrophic forgetting in continual learning, leading to significant performance degradation on previously encountered scenes and hindering model scalability and real-world deployment. To address this, we propose a lifelong learning framework tailored for PCPR. First, we introduce a dynamic replay sample selection mechanism that identifies representative historical samples based on a joint criterion of spatial diversity and discriminability. Second, we design a two-stage prompt-based training paradigm that decouples feature extraction from task adaptation, thereby enhancing cross-domain generalization. The framework integrates lightweight modules and an efficient knowledge fusion strategy. Extensive experiments on multiple large-scale benchmarks demonstrate substantial improvements: mean Identification Rate@1 (mIR@1) and mean Recall@1 (mR@1) increase by 6.50% and 7.96%, respectively, while the forgetting metric (F) decreases by 8.95%, outperforming existing continual learning approaches for PCPR.

Technology Category

• Machine Learning: Life-Long and Continual Learning
• Computer Vision: Learning & Optimization for CV
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Point cloud place recognition (PCPR) plays a crucial role in photogrammetry and robotics applications such as autonomous driving, intelligent transportation, and augmented reality. In real-world large-scale deployments of a positioning system, PCPR models must continuously acquire, update, and accumulate knowledge to adapt to diverse and dynamic environments, i.e., the ability known as continual learning (CL). However, existing PCPR models often suffer from catastrophic forgetting, leading to significant performance degradation in previously learned scenes when adapting to new environments or sensor types. This results in poor model scalability, increased maintenance costs, and system deployment difficulties, undermining the practicality of PCPR. To address these issues, we propose LifelongPR, a novel continual learning framework for PCPR, which effectively extracts and fuses knowledge from sequential point cloud data. First, to alleviate the knowledge loss, we propose a replay sample selection method that dynamically allocates sample sizes according to each dataset's information quantity and selects spatially diverse samples for maximal representativeness. Second, to handle domain shifts, we design a prompt learning-based CL framework with a lightweight prompt module and a two-stage training strategy, enabling domain-specific feature adaptation while minimizing forgetting. Comprehensive experiments on large-scale public and self-collected datasets are conducted to validate the effectiveness of the proposed method. Compared with state-of-the-art (SOTA) methods, our method achieves 6.50% improvement in mIR@1, 7.96% improvement in mR@1, and an 8.95% reduction in F. The code and pre-trained models are publicly available at https://github.com/zouxianghong/LifelongPR.

Problem

• Addresses catastrophic forgetting in point cloud place recognition models
• Enhances model scalability for diverse and dynamic environments
• Improves performance in continual learning for PCPR applications

Innovation

• Dynamic replay sample selection for knowledge retention
• Prompt learning-based continual learning framework
• Two-stage training for domain adaptation