GaGA is an MLLM-based interactive geolocation system that improves country-level accuracy by 4.57% and city-level accuracy by 2.92% over OSV-5M-Baseline on a reproduced GWS15k benchmark.
CPlaNet: Enhancing Image Geolocalization by Combinatorial Partitioning of Maps,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Towards Interactive Global Geolocation Assistant
GaGA is an MLLM-based interactive geolocation system that improves country-level accuracy by 4.57% and city-level accuracy by 2.92% over OSV-5M-Baseline on a reproduced GWS15k benchmark.