Region Matters: Efficient and Reliable Region-Aware Visual Place Recognition

TL;DR

FoL++ advances visual place recognition by integrating region-aware modeling and adaptive re-ranking, significantly boosting accuracy and speed with a lightweight approach.

Contribution

Introduces a novel reliability estimation and adaptive candidate scheduling framework for improved, efficient visual place recognition without manual annotations.

Findings

Achieves state-of-the-art performance on seven benchmarks.

Improves inference speed by 40% over previous methods.

Maintains a lightweight memory footprint.

Abstract

Visual Place Recognition (VPR) determines a query image's geographic location by matching it against geotagged databases. However, existing methods struggle with perceptual aliasing caused by irrelevant regions and inefficient re-ranking due to rigid candidate scheduling. To address these issues, we introduce FoL++, a method combining robust discriminative region modeling with adaptive re-ranking. Specifically, we propose a Reliability Estimation Branch to generate spatial reliability maps that explicitly model occlusion resistance. This representation is further optimized by two spatial alignment losses (SAL and SCEL) to effectively align features and highlight salient regions. For weakly supervised learning without manual annotations, a pseudo-correspondence strategy generates dense local feature supervision directly from aggregation clusters. Our Adaptive Candidate Scheduler dynamically resizes candidate pools based on global similarity. By weighting local matches by reliability and adaptively fusing global and local evidence, FoL++ surpasses traditional independent matching systems. Extensive experiments across seven benchmarks demonstrate that FoL++ achieves state-of-the-art performance with a lightweight memory footprint, improving inference speed by 40% over FoL. Code and models will be released (and merged with FoL) at https://github.com/chenshunpeng/FoL.