AI RESEARCH
From Pixels to Places: A Systematic Benchmark for Evaluating Image Geolocalization Ability in Large Language Models
arXiv CS.CV
•
ArXi:2508.01608v2 Announce Type: replace Image geolocalization, the task of identifying the geographic location depicted in an image, is important for applications in crisis response, digital forensics, and location-based intelligence. While recent advances in large language models (LLMs) offer new opportunities for visual reasoning, their ability to perform image geolocalization remains underexplored. In this study, we