Spatial-Conditioned Reasoning in Long-Egocentric Videos
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914454706323456 |
|---|---|
| author | Tribble, James Wang, Hao Hong, Si-En Zhou, Chaoyi Bastola, Ashish Huang, Siyu Razi, Abolfazl |
| author_facet | Tribble, James Wang, Hao Hong, Si-En Zhou, Chaoyi Bastola, Ashish Huang, Siyu Razi, Abolfazl |
| contents | Long-horizon egocentric video presents significant challenges for visual navigation due to viewpoint drift and the absence of persistent geometric context. Although recent vision-language models perform well on image and short-video reasoning, their spatial reasoning capability in long egocentric sequences remains limited. In this work, we study how explicit spatial signals influence VLM-based video understanding without modifying model architectures or inference procedures. We introduce Sanpo-D, a fine-grained re-annotation of the Google Sanpo dataset, and benchmark multiple VLMs on navigation-oriented spatial queries. To examine input-level inductive bias, we further fuse depth maps with RGB frames and evaluate their impact on spatial reasoning. Our results reveal a trade-off between general-purpose accuracy and spatial specialization, showing that depth-aware and spatially grounded representations can improve performance on safety-critical tasks such as pedestrian and obstruction detection. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_18100 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Spatial-Conditioned Reasoning in Long-Egocentric Videos Tribble, James Wang, Hao Hong, Si-En Zhou, Chaoyi Bastola, Ashish Huang, Siyu Razi, Abolfazl Computer Vision and Pattern Recognition Long-horizon egocentric video presents significant challenges for visual navigation due to viewpoint drift and the absence of persistent geometric context. Although recent vision-language models perform well on image and short-video reasoning, their spatial reasoning capability in long egocentric sequences remains limited. In this work, we study how explicit spatial signals influence VLM-based video understanding without modifying model architectures or inference procedures. We introduce Sanpo-D, a fine-grained re-annotation of the Google Sanpo dataset, and benchmark multiple VLMs on navigation-oriented spatial queries. To examine input-level inductive bias, we further fuse depth maps with RGB frames and evaluate their impact on spatial reasoning. Our results reveal a trade-off between general-purpose accuracy and spatial specialization, showing that depth-aware and spatially grounded representations can improve performance on safety-critical tasks such as pedestrian and obstruction detection. |
| title | Spatial-Conditioned Reasoning in Long-Egocentric Videos |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2601.18100 |