SinGeo：释放单模型潜力实现鲁棒的跨视角地理定位
SinGeo: Unlock Single Model's Potential for Robust Cross-View Geo-Localization
论文提出的课程学习策略和一致性评估方法可被视觉研究者复用于提升模型对视角变化的鲁棒性，例如在跨视角定位、无人机视觉等任务中设计渐进式训练和稳定性度量。
原文依据 SinGeo employs a dual discriminative learning architecture that enhances intra-view discriminability within both ground and satellite branches, and is the first to introduce a curriculum learning strategy to achieve robust CVGL.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论尽管近期进展激增，鲁棒的跨视角地理定位（CVGL）仍然具有挑战性。现有方法仍依赖于特定视场角（FoV）的训练范式，模型在固定FoV下优化，但在未见过的FoV和未知方向下测试时会崩溃。这一局限需要部署多个模型来覆盖多样变化。虽然已有研究通过简单随机化FoV探索动态FoV训练，但它们未能实现跨多样条件的鲁棒性——隐含假设所有FoV难度相同。为解决这一空白，我们提出SinGeo，一个简单而强大的框架，使单一模型无需额外模块或显式变换即可实现鲁棒的跨视角地理定位。SinGeo采用双判别学习架构，增强地面和卫星分支内的视角内判别性，并首次引入课程学习策略以实现鲁棒的CVGL。在四个基准数据集上的大量评估表明，SinGeo在多样条件下达到最优结果，并显著优于专门针对极端FoV训练的方法。除了优越性能，SinGeo还展现出跨架构可迁移性。此外，我们提出一种一致性评估方法，客观评估模型在不同视角下的稳定性，为理解和推进未来CVGL研究中的鲁棒性提供客观视角。代码见：https://github.com/Yangchen-nudt/SinGeo。
英文原文摘要 1570 字符
Robust cross-view geo-localization (CVGL) remains challenging despite the surge in recent progress. Existing methods still rely on field-of-view (FoV)-specific training paradigms, where models are optimized under a fixed FoV but collapse when tested on unseen FoVs and unknown orientations. This limitation necessitates deploying multiple models to cover diverse variations. Although studies have explored dynamic FoV training by simply randomizing FoVs, they failed to achieve robustness across diverse conditions---implicitly assuming all FoVs are equally difficult. To address this gap, we present SinGeo, a simple yet powerful framework that enables a single model to realize robust cross-view geo-localization without additional modules or explicit transformations. SinGeo employs a dual discriminative learning architecture that enhances intra-view discriminability within both ground and satellite branches, and is the first to introduce a curriculum learning strategy to achieve robust CVGL. Extensive evaluations on four benchmark datasets reveal that SinGeo sets state-of-the-art (SOTA) results under diverse conditions, and notably outperforms methods specifically trained for extreme FoVs. Beyond superior performance, SinGeo also exhibits cross-architecture transferability. Furthermore, we propose a consistency evaluation method to objectively assess model stability under varying views, providing an objective perspective for understanding and advancing robustness in future CVGL research. Codes are available at: https://github.com/Yangchen-nudt/SinGeo.