Metric-Bench: Exploring In-context Spatial Metric Reasoning in VLMs for Indoor Scenes

原文：Metric-Bench: Exploring In-context Spatial Metric Reasoning in VLMs for Indoor Scenes

作者：Yuling Xi, Haokai Zhang, Muzhi Zhu, Hao Zhong, Zongze Du, Hengyu Zhao, Chenchen Jing, Yufei Yin, Bin Qin, Yongjie Yang, Zhenbo Luo, Hao Chen, Chunhua Shen

来源：arXiv cs.CV（计算机视觉）

正文

Computer Science > Computer Vision and Pattern Recognition

arXiv:2609.25841v1 (cs)

[Submitted on 22 Sep 2026]

Title:Metric-Bench: Exploring In-context Spatial Metric Reasoning in VLMs for Indoor Scenes

Authors:Yuling Xi, Haokai Zhang, Muzhi Zhu, Hao Zhong, Zongze Du, Hengyu Zhao, Chenchen Jing, Yufei Yin, Bin Qin, Yongjie Yang, Zhenbo Luo, Hao Chen, Chunhua Shen

View a PDF of the paper titled Metric-Bench: Exploring In-context Spatial Metric Reasoning in VLMs for Indoor Scenes, by Yuling Xi and 12 other authors

View PDF

HTML (experimental)

Abstract:Metric reasoning is a critical and challenging task for Vision Language Models (VLMs), playing a pivotal role in embodied AI tasks such as robotic manipulation and autonomous navigation. However, current spatial reasoning remains bottlenecked by rigid pixel-level supervision; such localized optimization often compromises general multimodal intelligence, triggering performance degradation or catastrophic forgetting of broad reasoning capabilities. To address these limitations, we introduce Metric-Bench, a focused benchmark designed to guide metric-spatial reasoning using contextual information. By incorporating in-image reference objects with known physical dimensions, Metric-Bench guides models to implicitly learn the 2D-to-3D mapping without camera intrinsics. We further present MetricReasoner, a task-adapted reinforcement fine-tuning recipe for reference-grounded metric reasoning, using structured prompts and verifiable numerical rewards. Extensive experiments on Metric-Bench demonstrate that our approach significantly enhances spatial metric understanding, outperforming existing and even larger proprietary models by 43.1%, while improving downstream embodied performance over a spatial-specialized counterpart by 30.4% on RoboSpatial overall accuracy and 9.3% on ERQA, and additionally delivering consistent gains on general benchmarks (15.9% on V$\star$Bench, 88.9% on BLINK), indicating that the proposed adaptation does not necessarily compromise general VLM capabilities.

Comments:

Accepted to ECCV

Subjects:

Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)

Cite as:

arXiv:2609.25841 [cs.CV]

(or

arXiv:2609.25841v1 [cs.CV] for this version)

https://doi.org/10.48550/arXiv.2609.25841

Focus to learn more

arXiv-issued DOI via DataCite (pending registration)

主题

计算机视觉 · cs.MM

由「前沿雷达」于 2026-09-22 采集。正文取自原文页面，已保留出处链接。标题与正文版权归原作者所有。