| title | Reason, Then Re-reason: Cross-view Revisiting Improves Spatial Reasoning |
|---|---|
| authors | Chaofan Ma, Zhenjie Mao, Yuhuan Yang, Fanqin Zeng, Yue Shi, Yingjie Zhou, Xiaofeng Cao, Jiangchao Yao |
| date | 2026-06-10 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2606.11683 |
| affiliation | Shanghai Jiao Tong University |
| upvotes | 26 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf |
| abstract | Spatial reasoning from egocentric videos is inherently challenging because the observable evidence is constrained by the camera trajectory. Existing methods rely on single-turn inference, forcing models to resolve geometric ambiguity through semantic priors rather than verifiable evidence. We argue that spatial reasoning should be revisitable: conclusions formed under limited evidence should remain open to revision when complementary viewpoints become available. Building on this insight, we propose Reason, then Re-reason (ReRe), a training-free, inference-time framework with two phases: in the Reason Phase, an MLLM forms a spatial hypothesis from the original video; in the Re-reason Phase, it verifies or revises the hypothesis by observing a synthesized novel-view video. To enable effective cross-view revisiting, we design a Geometry-to-Video pipeline that renders strategically complementary novel views from predicted 3D geometry. These views feature an elevated, oblique perspective with scene-spanning coverage, while preserving the MLLM's native video interface without architectural modifications. Extensive evaluations on VSI-Bench and STI-Bench demonstrate that ReRe substantially boosts open-source MLLMs to rival proprietary state-of-the-art performance. Project page: https://zhenjiemao.github.io/ReRe/ |
cite
@misc{ma2026reason,
title={Reason, Then Re-reason: Cross-view Revisiting Improves Spatial Reasoning},
author={Chaofan Ma and Zhenjie Mao and Yuhuan Yang and Fanqin Zeng and Yue Shi and Yingjie Zhou and Xiaofeng Cao and Jiangchao Yao},
year={2026},
eprint={2606.11683},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2606.11683}
}