Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2605.10106 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910207800508416 |
|---|---|
| author | Mou, Tingshu He, Jiabo Wang, Renying Liu, Ce Yang, Hao Zhang, Tiehua Chen, Jingjing Ma, Xingjun |
| author_facet | Mou, Tingshu He, Jiabo Wang, Renying Liu, Ce Yang, Hao Zhang, Tiehua Chen, Jingjing Ma, Xingjun |
| contents | Recent advances in Multi-modal Large Language Models (MLLMs) target 3D spatial intelligence, yet the progress has been largely driven by post-training on curated benchmarks, leaving the inference-time approach relatively underexplored. In this paper, we take a training-free perspective and introduce ViSRA, a human-aligned Video-based Spatial Reasoning Agent, as a framework to probe the spatial reasoning mechanism of MLLMs. ViSRA elicits spatial reasoning in a modular and extensible manner by leveraging explicit spatial information from expert models, enabling a plug-and-play flexible paradigm. ViSRA offers two key advantages: (1) human-aligned and transferable 3D understanding rather than task-specific overfitting; and (2) no post-training computational cost along with heavy manual curation of spatial reasoning datasets. Experimental results demonstrate consistent improvement across a set of MLLMs on both existing benchmarks and unseen 3D spatial reasoning tasks, with ViSRA outperforming baselines by up to a 15.6% and 28.9% absolute margin respectively. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_10106 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | ViSRA: A Video-based Spatial Reasoning Agent for Multi-modal Large Language Models Mou, Tingshu He, Jiabo Wang, Renying Liu, Ce Yang, Hao Zhang, Tiehua Chen, Jingjing Ma, Xingjun Computer Vision and Pattern Recognition Artificial Intelligence Recent advances in Multi-modal Large Language Models (MLLMs) target 3D spatial intelligence, yet the progress has been largely driven by post-training on curated benchmarks, leaving the inference-time approach relatively underexplored. In this paper, we take a training-free perspective and introduce ViSRA, a human-aligned Video-based Spatial Reasoning Agent, as a framework to probe the spatial reasoning mechanism of MLLMs. ViSRA elicits spatial reasoning in a modular and extensible manner by leveraging explicit spatial information from expert models, enabling a plug-and-play flexible paradigm. ViSRA offers two key advantages: (1) human-aligned and transferable 3D understanding rather than task-specific overfitting; and (2) no post-training computational cost along with heavy manual curation of spatial reasoning datasets. Experimental results demonstrate consistent improvement across a set of MLLMs on both existing benchmarks and unseen 3D spatial reasoning tasks, with ViSRA outperforming baselines by up to a 15.6% and 28.9% absolute margin respectively. |
| title | ViSRA: A Video-based Spatial Reasoning Agent for Multi-modal Large Language Models |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2605.10106 |