Metraipolis › archive › dc_75eeb2e56b54
DynaTokens: Teaching Dynamics to Camera-Controlled Video Models at Test Time — arXiv:2609.35704
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.35704
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T23:54:55.776Z
- SHA-256 of the archived text
bd01bcb4cffec89d86b7827fadf4e6b748bae079e026e40ea5d230d844ce542e- Stable URL of this page
https://metraipolis.com/documento/dc_75eeb2e56b54
How to cite
DynaTokens: Teaching Dynamics to Camera-Controlled Video Models at Test Time — arXiv:2609.35704 — arXiv (en), https://arxiv.org/abs/2609.35704 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_75eeb2e56b54 (Metraipolis archive dc_75eeb2e56b54, sha256 bd01bcb4cffec89d86b7827fadf4e6b748bae079e026e40ea5d230d844ce542e)
BibTeX
@misc{dc75eeb2e56b54,
title = {DynaTokens: Teaching Dynamics to Camera-Controlled Video Models at Test Time — arXiv:2609.35704},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.35704},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_75eeb2e56b54; sha256 bd01bcb4cffec89d86b7827fadf4e6b748bae079e026e40ea5d230d844ce542e},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.35704. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Video generation must account for two sources of motion, one induced by the observer's camera path and the other caused by scene dynamics. An ideal camera-controlled video model should account for both motions: let users move the camera while evolving the scene dynamics. While current models handle camera-induced motion well in static settings, they struggle for dynamic scenes: objects are static, move incorrectly, or degrade in generation quality. We introduce DynaTokens, a lightweight set of learnable scene-specific tokens that teach dynamics to an existing camera-controlled world model. Our method is motivated by a simple asymmetry between the two sources of motion: whereas camera motion affects the generated view globally, object dynamics are spatially localized. Through cross-attention, DynaTokens trains the learnable tokens from a few example trajectories for a scene while keeping the base model frozen, and enables dynamics under new query camera paths. DynaTokens achieves a better simultaneous dynamics-camera tradeoff on VBench2 and WorldScore evaluations than LoRA, block finetuning, and specialized trainable-layer baselines. Analyses of token attention, ablations, and motion temporality suggest that matching the trainable interface to the structure of the learning target is important for effective adaptation. Project website: this https URL Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-29 in cs.CV). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2609.35704 — 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 2026-10-07T23:29:27.618Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).