NAVIG: Natural Language-guided Analysis with Vision Language Models for Image Geo-localization
This addresses the problem of accurate image location prediction for applications like mapping and navigation, but it is incremental as it builds on existing Vision Language Models.
The paper tackles image geo-localization by creating NaviClues, a dataset from GeoGuessr, and introducing Navig, a framework that uses language reasoning to reduce average distance error by 14% compared to previous state-of-the-art models with fewer than 1000 training samples.
Image geo-localization is the task of predicting the specific location of an image and requires complex reasoning across visual, geographical, and cultural contexts. While prior Vision Language Models (VLMs) have the best accuracy at this task, there is a dearth of high-quality datasets and models for analytical reasoning. We first create NaviClues, a high-quality dataset derived from GeoGuessr, a popular geography game, to supply examples of expert reasoning from language. Using this dataset, we present Navig, a comprehensive image geo-localization framework integrating global and fine-grained image information. By reasoning with language, Navig reduces the average distance error by 14% compared to previous state-of-the-art models while requiring fewer than 1000 training samples. Our dataset and code are available at https://github.com/SparrowZheyuan18/Navig/.