AI RESEARCH
Spatially-Weighted CLIP for Street-View Geo-localization
arXiv CS.CV
•
ArXi:2604.04357v1 Announce Type: new This paper proposes Spatially-Weighted CLIP (SW-CLIP), a novel framework for street-view geo-localization that explicitly incorporates spatial autocorrelation into vision-language contrastive learning. Unlike conventional CLIP-based methods that treat all non-matching samples as equally negative, SW-CLIP leverages Tobler's First Law of Geography to model geographic relationships through distance-aware soft supervision. Specifically, we