Locatability-Guided Adaptive Reasoning for Image Geo-Localization with Vision-Language Models

Bo Yu; Fengze Yang; Yiming Liu; Chao Wang; Xuewen Luo; Taozhe Li; Ruimin Ke; Xiaofan Zhou

cs.CVMar 13, 2026

Locatability-Guided Adaptive Reasoning for Image Geo-Localization with Vision-Language Models

Authors: Bo Yu, Fengze Yang, Yiming Liu, Chao Wang, Xuewen Luo, Taozhe Li, Ruimin Ke, Xiaofan Zhou,

Organizations: The University of Utah · The University of Oklahoma · Worcester Polytechnic Institute · Rensselaer Polytechnic Institute ·

Abstract

The emergence of Vision-Language Models (VLMs) has introduced new paradigms for global image geo-localization through retrieval-augmented generation (RAG) and reasoning-driven inference. However, RAG methods are constrained by retrieval database quality, while reasoning-driven approaches fail to internalize image locatability, relying on inefficient, fixed-depth reasoning paths that increase hallucinations and degrade accuracy. To overcome these limitations, we introduce an Optimized Locatability Score that quantifies an image's suitability for deep reasoning in geo-localization. Using this metric, we curate Geo-ADAPT-51K, a locatability-stratified reasoning dataset enriched with augmented reasoning trajectories for complex visual scenes. Building on this foundation, we propose a two-stage Group Relative Policy Optimization (GRPO) curriculum with customized reward functions that regulate adaptive reasoning depth, visual grounding, and hierarchical geographical accuracy. Our framework, Geo-ADAPT, learns an adaptive reasoning policy, achieves state-of-the-art performance across multiple geo-localization benchmarks, and substantially reduces hallucinations by reasoning both adaptively and efficiently.

Figures & tables

Figure 1: Illustration of locatability-guided adaptive reasoning. Current reasoning VLMs for image geo-localization are trained on fixed chain of thoughts, generating predetermined reasoning trajectories step by step without considering the actual image locatability during inference. In our work, we propose Geo-ADAPT, an adaptive reasoning framework that dynamically adjusts reasoning depth based on the image’s Optimized Locatability Score $L_{opt}$ .

Figure 2: Overview of Geo-ADAPT framework. Our approach comprises three components: (1) an Optimized Locatability Score $L_{opt}$ quantifying reasoning feasibility, (2) locatability-stratified dataset curation Geo-ADAPT-51K with enriched implicit reasoning trajectories, and (3) two-stage GRPO training with adaptive depth, grounding, and hierarchical accuracy rewards for dynamic reasoning allocation.

Figure 3: Motivation for Optimized Locatability Score $L_{opt}$ and locatability-guided adaptive reasoning strategies. (a) RAG retrieves visually similar candidates but fails due to being unable to convert implicit visual cues to rich semantic priors. (b) Standard reasoning initially fails similarly, but integrating implicit cues from retrieved image candidates into deep reasoning enables correct localization. Images where $d_{Reason}>d_{RAG}+\tau_{margin}$ receive lower $L_{opt}$ , triggering adaptive deep reasoning.

	IM2GPS3K					YFCC4K
Methods	Street 1km	City 25km	Region 200km	Country 750km	Continent 2500km	Street 1km	City 25km	Region 200km	Country 750km	Continent 2500km
[L]kNN, sigma=4 [ 52 ]	7.5	18.3	27.7	38.7	57.2	2.4	6.2	9.8	23.9	43.1
PlaNet [ 58 ]	7.8	25.8	33.7	48.7	63.2	6.5	14.2	22.4	36.0	56.4
CPlaNet [ 45 ]	11.3	25.7	34.7	47.7	65.3	6.6	15.4	21.6	36.9	54.5
ISNs [ 36 ]	10.1	29.2	35.5	50.5	64.5	6.7	15.7	25.1	36.7	55.1
Translocator [ 39 ]	12.1	30.5	47.6	57.7	80.5	7.6	19.7	25.6	41.8	60.2

Table 1: Performance comparison for Geo-ADAPT-8B on IM2GPS3K and YFCC4K with baseline methods. For all metrics, higher is better. The best performances are highlighted in bold , while the best performances among baselines are underlined . Change represents the relative improvement

\uparrow

or decline

\downarrow

of our method over the best baseline.

	Geo-ADAPT-51K Test
Methods	City Name Acc.	Country Name Acc.
Gemma3-27B [ 49 ]	37.6	71.9
InternVL3.5-38B [ 56 ]	36.1	79.4
Gemini 2.5 Flash † [ 9 ]	54.1	87.2
Qwen3-VL-30B [ 3 ]	43.9	83.5
GeoReasoner [ 25 ]	31.8	67.9

Table 2: Localization accuracy for city and country name prediction on the Geo-ADAPT-51K test split compared with baseline reasoning VLMs. † denotes models that are not publicly available.

	IM2GPS3K
Methods	Street 1km	City 25km	Region 200km	Country 750km	Continent 2500km
w/o $\mathcal{D}_{\text{aug}}$	17.7	43.2	60.1	75.4	89.2
w/o $R_{depth}$	17.2	44.3	60.5	76.7	88.9
w/o $R_{vis}$	17.4	43.7	60.8	76.8	89.0
w/o $R_{geo}$	15.3	40.4	58.9	75.7	87.9
w/o $\mathcal{T}_{\text{SFT}}$	17.8	44.9	61.3	76.5	89.1

Table 3: Ablation study on IM2GPS3K. The worst performance results in ablation study are underlined .

	Geo-ADAPT-51K Test
Methods	City Name Acc.	Country Name Acc.
w/o $\mathcal{D}_{\text{aug}}$	53.1	87.5
w/o $R_{depth}$	54.7	88.6
w/o $R_{vis}$	53.9	87.8
w/o $R_{geo}$	51.4	86.3
w/o $\mathcal{T}_{\text{SFT}}$	54.9	88.4

Table 4: Ablation study on Geo-ADAPT-51K test. The worst performance results in ablation study are underlined .

Appendix figures & tables4 assets

Supplementary material from the paper’s appendix.

Appendix

Parameter	Setting
GPU	NVIDIA H200 $\times$ 4
Dataset Samples	51K
Batch Size	64
Batch Size per Device	16
Training GPU Memory	90 GB / GPU
Base Model	Qwen3-VL-8B-Instruct

Appendix

Table 5: More Details on Training and Inference.

Optimizer Settings
Hyper-parameter	Value
Optimizer	AdamW
Adam $\beta_{1}$	0.9
Adam $\beta_{2}$	0.95
Weight Decay	0.1
Learning Rate Schedule

Appendix

Table 6: Hyper-parameter Settings of Proposed Geo-ADAPT.

Figure 4: Qualitative comparison on a weakly localizable sample ( $L_{opt}=0.42$ ). Facing visual ambiguity, Geo-ADAPT adaptively expands its reasoning depth (202 tokens), conducting a detailed forensic scan to extract the address “17510” and localize the exact shopping center. In contrast, the baseline generates a longer but less grounded response (318 tokens) and fails to identify the city.

Figure 5: Qualitative comparison on a highly localizable sample ( $L_{opt}=0.67$ ). Recognizing the strong explicit cues (Olympic Rings), Geo-ADAPT adaptively shortens its reasoning path for efficiency (156 tokens), rapidly converging on the correct location. This contrasts with the baseline, which engages in verbose description (329 tokens) without improving precision.

VLM Reasoning

Vision-language model reasoning studies how multimodal systems use visual evidence and language to answer questions, plan, and explain.

See all 846 papers

Skill-Conditioned Visual Geolocation for Vision-Language Models
Apr 10, 2026Chenjie Yang, Yutian Jiang, Yutong Deng +1VLM Reasoning Vision-Language Models
Southwest Jiaotong University Chengdu, China · The Hong Kong University of Science and Technology (Guangzhou) Guangzhou, China · Zhejiang University Hangzhou, China
Where Do Vision-Language Models Fail? World Scale Analysis for Image Geolocalization
Apr 17, 2026Siddhant Bharadwaj, Ashish Vashist, Fahimul Aleem +1VLM Evaluation Large Vision-Language Models
University of Central Florida
VLM-Guided Visual Place Recognition for Planet-Scale Geo-Localization
Jul 23, 2025Sania Waheed, Na Min An, Michael Milford +2Cross-View Geo-Localization Visual Place Recognition
Univ. of Southampton, UK · KAIST, South Korea · QUT, Australia +1

Apr 10, 2026J/K move · Enter open · S save

Skill-Conditioned Visual Geolocation for Vision-Language Models

Chenjie Yang, Yutian Jiang, Yutong Deng, Chenyu Wu

Southwest Jiaotong University Chengdu, China · The Hong Kong University of Science and Technology (Guangzhou) Guangzhou, China · Zhejiang University Hangzhou, China

VLM Reasoning Vision-Language Models Faithfulness in VLMs

Open paper AskPDF arXiv

Vision-language models (VLMs) have shown a promising ability in image geolocation, but they still lack structured geographic reasoning and the capacity for autonomous self-evolution. Existing methods predominantly rely on implicit parametric memory, which often exploits outdated knowledge and generates hallucinated reasoning. Furthermore, current inference is a "one-off" process, lacking the feedback loops necessary for self-evolution based on reasoning outcomes. To address these issues, we propose GeoSkill, a training-free framework based on an evolving Skill-Graph. We first initialize the graph by refining human expert trajectories into atomic, natural-language skills. For execution, GeoSkill employs an inference model to perform direct reasoning guided by the current Skill-Graph. For continuous growth, an Autonomous Evolution mechanism leverages a larger model to conduct multiple reasoning rollouts on image-coordinate pairs sourced from web-scale data and verified real-world reasoning. By analyzing both successful and failed trajectories from these rollouts, the mechanism iteratively synthesizes and prunes skills, effectively expanding the Skill-Graph and correcting geographic biases without any parameter updates. Experiments demonstrate that GeoSkill achieves promising performance in both geolocation accuracy and reasoning faithfulness on GeoRC, while maintaining superior generalization across diverse external datasets. Furthermore, our autonomous evolution fosters the emergence of novel, verifiable skills, significantly enhancing the system's cognition of real-world geographic knowledge beyond isolated case studies.

Abstract

VLM Reasoning

Explore similar work

Skill-Conditioned Visual Geolocation for Vision-Language Models

Where Do Vision-Language Models Fail? World Scale Analysis for Image Geolocalization

VLM-Guided Visual Place Recognition for Planet-Scale Geo-Localization