@inproceedings{kim-etal-2026-diagnosing,
title = "Diagnosing Spatial Consistency across Perspectives and Viewpoints in Large Vision-Language Models",
author = "Kim, Yoonji and
Kim, Jieun and
Jeong, Yujin and
Cho, Sung-Bae",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://preview.aclanthology.org/ingest-acl/2026.acl-long.1514/",
pages = "32803--32827",
ISBN = "979-8-89176-390-6",
abstract = "Consistent reasoning about 3D spatial relations across changing viewpoints is fundamental for Embodied AI agents operating in dynamic environments. While Large Vision-Language Models (LVLMs) have advanced multimodal perception, their ability to maintain spatial consistency across diverse perspectives remains underexplored. Existing benchmarks primarily assess spatial capabilities from a static, single-view, and egocentric perspective, failing to capture the dynamic nature of real-world spatial cognition.To address this gap, we introduce \textbf{SCOPE} (\textbf{S}patial \textbf{CO}nsistency across \textbf{PE}rspectives and Viewpoints), a comprehensive benchmark designed to rigorously diagnose spatial reasoning capabilities. Grounded in human cognitive theories of dual spatial representations, SCOPE discretizes the $360^\circ$ field into multiview scenarios to systematically evaluate both allocentric and egocentric reasoning capabilities. Our dataset comprises \textbf{20.1K} spatial VQA pairs derived from high-quality 3D environments. Through an extensive evaluation of 26 state-of-the-art LVLMs, we identify two fundamental limitations that prevent consistent spatial understanding across viewpoints.We hope \textbf{SCOPE} facilitates the diagnosis of spatial reasoning, serving as a stepping stone toward reliable embodied action."
}Markdown (Informal)
[Diagnosing Spatial Consistency across Perspectives and Viewpoints in Large Vision-Language Models](https://preview.aclanthology.org/ingest-acl/2026.acl-long.1514/) (Kim et al., ACL 2026)
ACL