OmniMapBench: Benchmarking Visual-Centric Reasoning on Diverse Map Documents
A critical limitation is identified in many document understanding benchmarks: visual content is often reducible to text, enabling high performance without genuine visual grounding. To address this limitation, OmniMapBench is introduced to foster visual-centric reasoning for map documents. The benchmark comprises 2,096 manually annotated question-answer pairs across 1,603 map documents from nine categories.
- ▪A critical limitation is identified in many document understanding benchmarks: visual content is often reducible to text, enabling high performance without genuine visual grounding.
- ▪To address this limitation, OmniMapBench is introduced to foster visual-centric reasoning for map documents.
- ▪The benchmark comprises 2,096 manually annotated question-answer pairs across 1,603 map documents from nine categories.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Computer Vision and Pattern Recognition arXiv:2607.09068 (cs) [Submitted on 10 Jul 2026] Title:OmniMapBench: Benchmarking Visual-Centric Reasoning on Diverse Map Documents Authors:Yang Chen, Yunwen Li, Yufan Shen, Minghao Liu, Tianyu Zheng, Bin Fu, Qunshu Lin, Zhi Yu, Botian Shi View a PDF of the paper titled OmniMapBench: Benchmarking Visual-Centric Reasoning on Diverse Map Documents, by Yang Chen and Yunwen Li and Yufan Shen and Minghao Liu and Tianyu Zheng and Bin Fu and Qunshu Lin and Zhi Yu and Botian Shi View PDF HTML (experimental) Abstract:Recent advancements in LVLMs necessitate robust benchmarks for complex, visually grounded reasoning.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
