Global Cross-Modal Geo-Localization: A Million-Scale Dataset and a Physical Consistency Learning Framework

Global Cross-Modal Geo-Localization: A Million-Scale Dataset and a Physical Consistency Learning Framework explores CORE is a million-scale dataset for cross-modal geo-localization, enabling a physical-law-aware network (PLANET) that significantly outperforms state-of-the-art methods, with code and dataset release for immediate application.. Commercial viability score: 8/10 in Cross-Modal Geo-Localization.