Metraipolis › archive › dc_7c23cd3852fe

MVAgent: Multi-Agent Video Generation via Consistent Condition Construction and Shot-Level Policy Optimization — arXiv:2609.30609

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.30609
License (as archived): CC0-1.0
Archived: 2026-10-08T03:12:29.296Z
SHA-256 of the archived text: 1c7e0e4889ef38ec39e858c41b1caf1ea1621e8005fecd07f00dafd8fe8d4bd3
Stable URL of this page: https://metraipolis.com/documento/dc_7c23cd3852fe

How to cite

MVAgent: Multi-Agent Video Generation via Consistent Condition Construction and Shot-Level Policy Optimization — arXiv:2609.30609 — arXiv (en), https://arxiv.org/abs/2609.30609 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7c23cd3852fe (Metraipolis archive dc_7c23cd3852fe, sha256 1c7e0e4889ef38ec39e858c41b1caf1ea1621e8005fecd07f00dafd8fe8d4bd3)

BibTeX

@misc{dc7c23cd3852fe,
  title = {MVAgent: Multi-Agent Video Generation via Consistent Condition Construction and Shot-Level Policy Optimization — arXiv:2609.30609},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.30609},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7c23cd3852fe; sha256 1c7e0e4889ef38ec39e858c41b1caf1ea1621e8005fecd07f00dafd8fe8d4bd3},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.30609. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Multi-shot agentic video generation requires consistent character appearance, stable spatial layout across camera angles, and continuous character state between shots. When every shot is a separate request to a frozen generator, repeated text does not determine appearance, layout or state. We therefore recast the problem as condition construction and present MVAgent, a multi-agent pipeline whose agents collaborate through typed conditioning inputs. Because an environment image shows one viewpoint, a Spatial Grounding agent samples views from generated camera-traversal clips and anchors each shot to the view matching its framing. As generated shots drift from the plan, an Observer records how each shot ends in a continuity memory, from which a Transition agent builds character action and spatial references for the next shot. An Orchestrator composes these inputs into each request. Since a request reveals its effect only after rendering, we train it by agentic reinforcement learning with Trunk-GDPO, which compares rendered candidates at every shot rather than once per video and continues the best as the trunk. With generator and judges frozen, MVAgent attains the highest cross-shot consistency and narrative-planning quality among the compared methods on ViMax-Bench and is preferred over the strongest agentic baseline in human evaluation.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-28 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.30609 — 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 2026-10-08T02:50:19.029Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).