Can Large Vision Language Models Read Maps Like a Human?
Abstract
MapBench challenges LVLMs with pixel-based map navigation problems, highlighting limitations in spatial reasoning and decision-making.
In this paper, we introduce MapBench-the first dataset specifically designed for human-readable, pixel-based map-based outdoor navigation, curated from complex path finding scenarios. MapBench comprises over 1600 pixel space map path finding problems from 100 diverse maps. In MapBench, LVLMs generate language-based navigation instructions given a map image and a query with beginning and end landmarks. For each map, MapBench provides Map Space Scene Graph (MSSG) as an indexing data structure to convert between natural language and evaluate LVLM-generated results. We demonstrate that MapBench significantly challenges state-of-the-art LVLMs both zero-shot prompting and a Chain-of-Thought (CoT) augmented reasoning framework that decomposes map navigation into sequential cognitive processes. Our evaluation of both open-source and closed-source LVLMs underscores the substantial difficulty posed by MapBench, revealing critical limitations in their spatial reasoning and structured decision-making capabilities. We release all the code and dataset in https://github.com/taco-group/MapBench.
Community
In this paper, we introduce MapBench-the first dataset specifically designed for human-readable, pixel-based map-based outdoor navigation, curated from complex path finding scenarios. MapBench comprises over 1600 pixel space map path finding problems from 100 diverse maps. In MapBench, LVLMs generate language-based navigation instructions given a map image and a query with beginning and end landmarks. For each map, MapBench provides Map Space Scene Graph (MSSG) as an indexing data structure to convert between natural language and evaluate LVLM-generated results. We demonstrate that MapBench significantly challenges state-of-the-art LVLMs both zero-shot prompting and a Chain-of-Thought (CoT) augmented reasoning framework that decomposes map navigation into sequential cognitive processes. Our evaluation of both open-source and closed-source LVLMs underscores the substantial difficulty posed by MapBench, revealing critical limitations in their spatial reasoning and structured decision-making capabilities.
This is an automated message from the Librarian Bot. I found the following papers similar to this paper.
The following papers were recommended by the Semantic Scholar API
- EmbodiedVSR: Dynamic Scene Graph-Guided Chain-of-Thought Reasoning for Visual Spatial Tasks (2025)
- MapNav: A Novel Memory Representation via Annotated Semantic Maps for VLM-based Vision-and-Language Navigation (2025)
- Integrating Chain-of-Thought for Multimodal Alignment: A Study on 3D Vision-Language Learning (2025)
- ChatBEV: A Visual Language Model that Understands BEV Maps (2025)
- SmartWay: Enhanced Waypoint Prediction and Backtracking for Zero-Shot Vision-and-Language Navigation (2025)
- CoSpace: Benchmarking Continuous Space Perception Ability for Vision-Language Models (2025)
- NuPlanQA: A Large-Scale Dataset and Benchmark for Multi-View Driving Scene Understanding in Multi-Modal Large Language Models (2025)
Please give a thumbs up to this comment if you found it helpful!
If you want recommendations for any Paper on Hugging Face checkout this Space
You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: @librarian-bot recommend
Get this paper in your agent:
hf papers read 2503.14607 Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash Models citing this paper 0
No model linking this paper
Datasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 0
No Space linking this paper