Dual-level Progressive Hardness-Aware Reweighting for Cross-View Geo-Localization

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zheng, Guozheng, Guan, Jian, Xie, Mingjie, Zhao, Xuanjia, Fan, Congyi, Zhang, Shiheng, Feng, Pengming
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914133292613632
author Zheng, Guozheng
Guan, Jian
Xie, Mingjie
Zhao, Xuanjia
Fan, Congyi
Zhang, Shiheng
Feng, Pengming
author_facet Zheng, Guozheng
Guan, Jian
Xie, Mingjie
Zhao, Xuanjia
Fan, Congyi
Zhang, Shiheng
Feng, Pengming
contents Cross-view geo-localization (CVGL) between drone and satellite imagery remains challenging due to severe viewpoint gaps and the presence of hard negatives, which are visually similar but geographically mismatched samples. Existing mining or reweighting strategies often use static weighting, which is sensitive to distribution shifts and prone to overemphasizing difficult samples too early, leading to noisy gradients and unstable convergence. In this paper, we present a Dual-level Progressive Hardness-aware Reweighting (DPHR) strategy. At the sample level, a Ratio-based Difficulty-Aware (RDA) module evaluates relative difficulty and assigns fine-grained weights to negatives. At the batch level, a Progressive Adaptive Loss Weighting (PALW) mechanism exploits a training-progress signal to attenuate noisy gradients during early optimization and progressively enhance hard-negative mining as training matures. Experiments on the University-1652 and SUES-200 benchmarks demonstrate the effectiveness and robustness of the proposed DPHR, achieving consistent improvements over state-of-the-art methods.
format Preprint
id arxiv_https___arxiv_org_abs_2510_27181
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Dual-level Progressive Hardness-Aware Reweighting for Cross-View Geo-Localization
Zheng, Guozheng
Guan, Jian
Xie, Mingjie
Zhao, Xuanjia
Fan, Congyi
Zhang, Shiheng
Feng, Pengming
Computer Vision and Pattern Recognition
Artificial Intelligence
Cross-view geo-localization (CVGL) between drone and satellite imagery remains challenging due to severe viewpoint gaps and the presence of hard negatives, which are visually similar but geographically mismatched samples. Existing mining or reweighting strategies often use static weighting, which is sensitive to distribution shifts and prone to overemphasizing difficult samples too early, leading to noisy gradients and unstable convergence. In this paper, we present a Dual-level Progressive Hardness-aware Reweighting (DPHR) strategy. At the sample level, a Ratio-based Difficulty-Aware (RDA) module evaluates relative difficulty and assigns fine-grained weights to negatives. At the batch level, a Progressive Adaptive Loss Weighting (PALW) mechanism exploits a training-progress signal to attenuate noisy gradients during early optimization and progressively enhance hard-negative mining as training matures. Experiments on the University-1652 and SUES-200 benchmarks demonstrate the effectiveness and robustness of the proposed DPHR, achieving consistent improvements over state-of-the-art methods.
title Dual-level Progressive Hardness-Aware Reweighting for Cross-View Geo-Localization
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2510.27181