When Vision Misleads, Let Location Speak: A Worldwide Image Geo-Localization Method via Location Attention Mechanism and Large Multimodal Models
2026-06-08 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors developed TransGeoCLIP, a system that helps figure out where a photo was taken anywhere in the world. It does this better than before by using both image and location information together and paying special attention to GPS coordinates. Their approach uses advanced AI models called Transformers and large multimodal models to match photos with places more accurately, especially when different places look very similar. Tests on several datasets show that their method improves accuracy in pinpointing exact locations, especially within one kilometer.
image geo-localizationTransformer encoderlocation attention mechanismGPS coordinateslarge multimodal models (LMMs)CLIPretrieval-based frameworkstreet-level localizationimage-text-GPS embedding
Authors
Junchao Cui, Wenqi Shi, Xuanzi Ma, Nan Wu, Shaoyong Du, Xiangyang Luo
Abstract
Worldwide image geo-localization aims to determine the capture location of an image on a global scale. Existing methods often mislocalize images by matching them to visually similar scenes from different geographic regions, which limits reliability in practical applications. To address this issue, we propose TransGeoCLIP, a novel retrieval-based framework that integrates a location attention mechanism and large multimodal models (LMMs). Using the Transformer encoder with location attention to encode GPS coordinates, TransGeoCLIP can effectively distinguish geographic features among visually similar images. The framework consists of two stages: 1) Retrieval database construction, which employs Transformers equipped with location attention mechanisms to encode labeled GPS coordinates and enhance location semantics, subsequently enables joint image-text-GPS embedding through CLIP; 2) Retrieval-augmented inference, which leverages LMMs to infer the final image location prediction from retrieved database results. Extensive experimental results on diverse datasets, including IM2GPS, IM2GPS3k, YFCC4k, and YFCC26k, demonstrate that TransGeoCLIP significantly enhances localization performance for visually similar images. Particularly, street-level localization accuracy (within 1 km error) is substantially improved, surpassing state-of-the-art methods by 1.5%, 1.07%, 7.18%, and 9.75% on these benchmarks, respectively.