---
格式版本: 2
标题: "GeoStore: Finding Small Storefronts in Large Scenes -- A Fine-Grained POI Localization Benchmark with Global-to-Local Asymmetric Matching"
原文链接: "https://arxiv.org/abs/2609.02012"
发布日期: "2026-09-02"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:local:strict_original_body"
发布时间证据: "**\\[v1\\]** Wed, 2 Sep 2026 02:39:21 UTC (3,295 KB)"
发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body"
发布时间校准置信度: "high"
发布时间候选数量: 18
发布时间严格候选数量: 6
发布时间原页读取状态: "source template page reused from URL open"
发布时间未找到原因: ""
发布时间校准时间: "2026-09-04T01:38:34+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-09-04T01:32:16+08:00"
入库时间: "2026-09-03T17:38:34.939Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=Scale-up&searchtype=all"
匹配关键词:
  - "Scale-up"
相关厂家:
  []
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 5
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "该页面为arXiv计算机视觉论文，主题为POI定位与图像匹配，与超节点/AI Rack/机柜级AI基础设施完全无关。仅因搜索词命中Scale-up，属明显无关内容。"
AI质检模型: "zj-deepseek-v4-flash"
AI质检时间: "2026-09-04T01:38:53+08:00"
AI主题相关性: 0
AI来源权威性: 3
AI新颖性: 0
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 2
AI摘要: "该文提出GeoStore，首个面向细粒度POI定位的基准，任务是将门店近景照片匹配到街景大场景中的小店面区域；现有全局描述子方法在这种非对称场景下受限。"
AI摘要模型: "ali-deepseek-v4-flash"
AI摘要时间: "2026-09-04T00:17:42.995Z"
采集批次: "2026年9月3日22点43分34秒"
采集批次ID: "20260903-224334-406"
去重键: "https://arxiv.org/abs/2609.02012"
---

## Computer Science > Computer Vision and Pattern Recognition

## Title:GeoStore: Finding Small Storefronts in Large Scenes -- A Fine-Grained POI Localization Benchmark with Global-to-Local Asymmetric Matching

Authors:[Lu Han](https://arxiv.org/search/cs?searchtype=author&query=Han,+L), [Xiting Sun](https://arxiv.org/search/cs?searchtype=author&query=Sun,+X), [Hao Wang](https://arxiv.org/search/cs?searchtype=author&query=Wang,+H), [Zhiqiang Cao](https://arxiv.org/search/cs?searchtype=author&query=Cao,+Z), [Ruihuan Du](https://arxiv.org/search/cs?searchtype=author&query=Du,+R), [Ziquan Zeng](https://arxiv.org/search/cs?searchtype=author&query=Zeng,+Z), [Chunlong Lv](https://arxiv.org/search/cs?searchtype=author&query=Lv,+C)

[View PDF](https://arxiv.org/pdf/2609.02012) [HTML (experimental)](https://arxiv.org/html/2609.02012v1)

> Abstract:Point-of-interest (POI) localization -- matching a user's close-up storefront photograph against large-scale geo-tagged street-view imagery -- underpins map construction, POI verification, and location-based services. Its closest existing paradigm, visual place recognition (VPR), assumes symmetric, whole-image matching of the same scene at a comparable scale; POI localization instead must match a close-up query, in which the target fills the frame, against wide references in which the same POI occupies only a small, off-center region among visually similar shops, under a substantial capture-domain gap. We introduce GeoStore, to our knowledge the first benchmark dedicated to this asymmetric, fine-grained, open-set formulation, and show that global-descriptor methods tuned for symmetric VPR are systematically limited on it, since a single global vector dilutes the small target. We further propose GLAM (Global-to-Local Asymmetric Matching), which couples a retrieval-anchoring global descriptor with an asymmetric local pathway: each reference is kept as a compact set of pooled region tokens and matched against a single query probe through a learnable soft late interaction; at inference, the same tokens enable a lightweight mutual-nearest-neighbor re-ranking. GLAM surpasses strong global and two-stage baselines on Recall@1/5/10 and mAP, with ~5x smaller re-ranking features and ~two orders of magnitude lower per-pair matching cost than prior local re-ranking. The benchmark and code will be publicly released.

| Comments: |  |
| --- | --- |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Information Retrieval (cs.IR) |
| Cite as: | [arXiv:2609.02012](https://arxiv.org/abs/2609.02012) \[cs.CV\] |
|  | (or [arXiv:2609.02012v1](https://arxiv.org/abs/2609.02012v1) \[cs.CV\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2609.02012](https://doi.org/10.48550/arXiv.2609.02012) |

## Submission history

From: Lu Han \[[view email](https://arxiv.org/show-email/c41ef3ff/2609.02012)\]  
**\[v1\]** Wed, 2 Sep 2026 02:39:21 UTC (3,295 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2609.02012) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))