Metraipolis › archive › dc_d3b65002f888
DK-VLM: Geometry as Position for Spatial Reasoning in Vision-Language Models — arXiv:2610.22356
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.22356
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T10:39:23.289Z
- SHA-256 of the archived text
79ace208c4b02a4bb04802ff0282285d15baf20448ba80e3f1223ef3064a7071- Stable URL of this page
https://metraipolis.com/documento/dc_d3b65002f888
How to cite
DK-VLM: Geometry as Position for Spatial Reasoning in Vision-Language Models — arXiv:2610.22356 — arXiv (en), https://arxiv.org/abs/2610.22356 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d3b65002f888 (Metraipolis archive dc_d3b65002f888, sha256 79ace208c4b02a4bb04802ff0282285d15baf20448ba80e3f1223ef3064a7071)
BibTeX
@misc{dcd3b65002f888,
title = {DK-VLM: Geometry as Position for Spatial Reasoning in Vision-Language Models — arXiv:2610.22356},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.22356},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d3b65002f888; sha256 79ace208c4b02a4bb04802ff0282285d15baf20448ba80e3f1223ef3064a7071},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.22356. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Vision-language models still struggle to understand 3D spatial relations across views and update spatial judgments as the observer moves. A common approach fuses 3D reconstruction features into visual tokens, but aligning these features with pretrained semantic representations remains challenging under limited spatial supervision. We propose DK-VLM with Dual-Kernel Attention, which computes attention from both a visual token's input position and its geometric position in the scene. A semantic kernel retains the pretrained model's native attention and positional encoding, while a separate geometric kernel uses estimated 3D points and camera poses to directly modulate attention through point displacements and relative view transformations. Then, a learned gate fuses their outputs. Geometry thus enters attention as position rather than as fused reconstruction features. We further introduce Reference-Frame Composition (RC) supervision to train the model to update its reference frame across successive judgments. It provides stepwise turn and heading targets derived from verified single-step annotations. Among the compared methods, DK-VLM-8B achieves the highest average score across six spatial benchmarks and the highest overall scores on VSI-Bench (74.33) and ReVSI-32 (58.28). RC supervision further improves multi-step spatial reasoning, while the full model preserves general video understanding. Project page: this https URL . Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2610.22356 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-16T08:52:02.744Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).