Q-GeoMem: Question-Guided Geometric Memory for Video Spatial Reasoning
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914605302808576 |
|---|---|
| author | Gao, Xianqiang Chen, Qizhi Qu, Delin Song, Haoming Wang, Zhigang Zhao, Bin Wang, Dong Li, Xuelong |
| author_facet | Gao, Xianqiang Chen, Qizhi Qu, Delin Song, Haoming Wang, Zhigang Zhao, Bin Wang, Dong Li, Xuelong |
| contents | Video spatial reasoning requires accumulating viewpoint-dependent evidence over time while retaining information useful to the question being asked. Existing spatial video-language models improve geometric perception and long-range context modeling, but often treat memory as a generic temporal cache, which can introduce redundant or irrelevant geometry and weaken long-horizon reasoning. We propose \textbf{\ours}, a question-guided geometric memory framework for video spatial reasoning. \ours injects camera-conditioned geometry into visual tokens and maintains two complementary memories: a Fine-Grained Context Bank for recent dense features and camera states, and a Semantic-Geometric Evidence Bank for compact long-range evidence. Each candidate frame is scored by the product of Q-Former-based question relevance and novelty with respect to the retained bank; this score is stored and reused during reading, while a capacity-based replacement rule keeps the bank compact. During reasoning, both memories are read before update and adaptively fused with the current frame representation. Experiments on VSI-Bench and VSTI-Bench show that \ours achieves state-of-the-art performance among evaluated spatial reasoning models, validating the effectiveness of question-guided geometric memory. Ablations further verify the contribution of the proposed evidence scoring mechanism. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_27318 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Q-GeoMem: Question-Guided Geometric Memory for Video Spatial Reasoning Gao, Xianqiang Chen, Qizhi Qu, Delin Song, Haoming Wang, Zhigang Zhao, Bin Wang, Dong Li, Xuelong Computer Vision and Pattern Recognition Video spatial reasoning requires accumulating viewpoint-dependent evidence over time while retaining information useful to the question being asked. Existing spatial video-language models improve geometric perception and long-range context modeling, but often treat memory as a generic temporal cache, which can introduce redundant or irrelevant geometry and weaken long-horizon reasoning. We propose \textbf{\ours}, a question-guided geometric memory framework for video spatial reasoning. \ours injects camera-conditioned geometry into visual tokens and maintains two complementary memories: a Fine-Grained Context Bank for recent dense features and camera states, and a Semantic-Geometric Evidence Bank for compact long-range evidence. Each candidate frame is scored by the product of Q-Former-based question relevance and novelty with respect to the retained bank; this score is stored and reused during reading, while a capacity-based replacement rule keeps the bank compact. During reasoning, both memories are read before update and adaptively fused with the current frame representation. Experiments on VSI-Bench and VSTI-Bench show that \ours achieves state-of-the-art performance among evaluated spatial reasoning models, validating the effectiveness of question-guided geometric memory. Ablations further verify the contribution of the proposed evidence scoring mechanism. |
| title | Q-GeoMem: Question-Guided Geometric Memory for Video Spatial Reasoning |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2605.27318 |