Metraipolis › archive › dc_afb73e6449b2

PhyMetric: Diagnosing Physical Plausibility in Text-to-Video Generation via Scene-Level QA — arXiv:2610.19823

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.19823
License (as archived)
CC0-1.0
Archived
2026-10-15T21:56:39.587Z
SHA-256 of the archived text
f6289c13ad5020b0a178940e8a0fb9bc0fa6a33d89d54bd316289b42663c86f5
Stable URL of this page
https://metraipolis.com/documento/dc_afb73e6449b2

How to cite

PhyMetric: Diagnosing Physical Plausibility in Text-to-Video Generation via Scene-Level QA — arXiv:2610.19823 — arXiv (en), https://arxiv.org/abs/2610.19823 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_afb73e6449b2 (Metraipolis archive dc_afb73e6449b2, sha256 f6289c13ad5020b0a178940e8a0fb9bc0fa6a33d89d54bd316289b42663c86f5)
BibTeX
@misc{dcafb73e6449b2,
  title = {PhyMetric: Diagnosing Physical Plausibility in Text-to-Video Generation via Scene-Level QA — arXiv:2610.19823},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.19823},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_afb73e6449b2; sha256 f6289c13ad5020b0a178940e8a0fb9bc0fa6a33d89d54bd316289b42663c86f5},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19823. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Text-to-video (T2V) generation models can produce visually realistic videos that nevertheless violate the physical consequences implied by the prompt and scene. Existing benchmarks largely rely on holistic plausibility scores or broad quality dimensions, leaving it unclear which physical behavior is being tested and why a model fails. We introduce PhyMetric, a scene-level diagnostic benchmark that recasts physical plausibility evaluation as verifiable, scene-grounded question answering. PhyMetric contains 1{,}115 scenes spanning 50 physical categories across six domains, paired with 8{,}920 answer-balanced binary questions over 12 evaluation dimensions. Each scene is associated with eight targeted yes/no questions and a structured physical prior specifying the governing law, expected temporal sequence, observable signatures, and common failure modes. At evaluation time, a vision-language model (VLM) judge answers each question from question-aware sampled frames conditioned on the prior, and correctness is hierarchically aggregated into scene, dimension, physical category, and model scores. Experiments on four open-source T2V models and three VLM judges show that scene-specific binary QA is a critical component of the protocol: when each targeted question is replaced by a single generic plausibility query, evaluator accuracy drops to the chance level. Structured physical priors offer a complementary signal that improves human alignment of the resulting scores, with effects that vary across VLM judges of different capacity. These findings establish PhyMetric as a benchmark that turns physical plausibility evaluation from opaque holistic scoring into interpretable, scene-level inspection. Our project page is available at: 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/2610.19823 — 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 2026-10-15T21:38:29.504Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).