Turning Generators into Retrievers: Unlocking MLLMs for Natural Language-Guided Geo-Localization

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Chen, Yuqi, Zhang, Xiaohan, Arrabi, Ahmad, Sultani, Waqas, Chen, Chen, Wshah, Safwan
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910122741071872
author Chen, Yuqi
Zhang, Xiaohan
Arrabi, Ahmad
Sultani, Waqas
Chen, Chen
Wshah, Safwan
author_facet Chen, Yuqi
Zhang, Xiaohan
Arrabi, Ahmad
Sultani, Waqas
Chen, Chen
Wshah, Safwan
contents Natural-language Guided Cross-view Geo-localization (NGCG) aims to retrieve geo-tagged satellite imagery using textual descriptions of ground scenes. While recent NGCG methods commonly rely on CLIP-style dual-encoder architectures, they often suffer from weak cross-modal generalization and require complex architectural designs. In contrast, Multimodal Large Language Models (MLLMs) offer powerful semantic reasoning capabilities but are not directly optimized for retrieval tasks. In this work, we present a simple yet effective framework to adapt MLLMs for NGCG via parameter-efficient finetuning. Our approach optimizes latent representations within the MLLM while preserving its pretrained multimodal knowledge, enabling strong cross-modal alignment without redesigning model architectures. Through systematic analysis of diverse variables, from model backbone to feature aggregation, we provide practical and generalizable insights for leveraging MLLMs in NGCG. Our method achieves SOTA on GeoText-1652 with a 12.2% improvement in Text-to-Image Recall@1 and secures top performance in 5 out of 12 subtasks on CVG-Text, all while surpassing baselines with far fewer trainable parameters. These results position MLLMs as a robust foundation for semantic cross-view retrieval and pave the way for MLLM-based NGCG to be adopted as a scalable, powerful alternative to traditional dual-encoder designs. Project page and code are available at https://yuqichen888.github.io/NGCG-MLLMs-web/.
format Preprint
id arxiv_https___arxiv_org_abs_2604_10721
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Turning Generators into Retrievers: Unlocking MLLMs for Natural Language-Guided Geo-Localization
Chen, Yuqi
Zhang, Xiaohan
Arrabi, Ahmad
Sultani, Waqas
Chen, Chen
Wshah, Safwan
Computer Vision and Pattern Recognition
Artificial Intelligence
Natural-language Guided Cross-view Geo-localization (NGCG) aims to retrieve geo-tagged satellite imagery using textual descriptions of ground scenes. While recent NGCG methods commonly rely on CLIP-style dual-encoder architectures, they often suffer from weak cross-modal generalization and require complex architectural designs. In contrast, Multimodal Large Language Models (MLLMs) offer powerful semantic reasoning capabilities but are not directly optimized for retrieval tasks. In this work, we present a simple yet effective framework to adapt MLLMs for NGCG via parameter-efficient finetuning. Our approach optimizes latent representations within the MLLM while preserving its pretrained multimodal knowledge, enabling strong cross-modal alignment without redesigning model architectures. Through systematic analysis of diverse variables, from model backbone to feature aggregation, we provide practical and generalizable insights for leveraging MLLMs in NGCG. Our method achieves SOTA on GeoText-1652 with a 12.2% improvement in Text-to-Image Recall@1 and secures top performance in 5 out of 12 subtasks on CVG-Text, all while surpassing baselines with far fewer trainable parameters. These results position MLLMs as a robust foundation for semantic cross-view retrieval and pave the way for MLLM-based NGCG to be adopted as a scalable, powerful alternative to traditional dual-encoder designs. Project page and code are available at https://yuqichen888.github.io/NGCG-MLLMs-web/.
title Turning Generators into Retrievers: Unlocking MLLMs for Natural Language-Guided Geo-Localization
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2604.10721