Metraipolis › archive › dc_cf0bbef10da1

Putting Captions to the Test: Evaluating Video Caption Quality through Multiple-Choice Question Answering — arXiv:2609.09973

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.09973
License (as archived): CC0-1.0
Archived: 2026-10-16T10:59:54.925Z
SHA-256 of the archived text: 2f04f232a67b17f81dcb2d294882a6b7279d502e69ae5e6478dac025059fe3a2
Stable URL of this page: https://metraipolis.com/documento/dc_cf0bbef10da1

How to cite

Putting Captions to the Test: Evaluating Video Caption Quality through Multiple-Choice Question Answering — arXiv:2609.09973 — arXiv (en), https://arxiv.org/abs/2609.09973 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_cf0bbef10da1 (Metraipolis archive dc_cf0bbef10da1, sha256 2f04f232a67b17f81dcb2d294882a6b7279d502e69ae5e6478dac025059fe3a2)

BibTeX

@misc{dccf0bbef10da1,
  title = {Putting Captions to the Test: Evaluating Video Caption Quality through Multiple-Choice Question Answering — arXiv:2609.09973},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.09973},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_cf0bbef10da1; sha256 2f04f232a67b17f81dcb2d294882a6b7279d502e69ae5e6478dac025059fe3a2},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.09973. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Evaluating video captioning remains a critical challenge for Visual Large Language Models (VLLMs). Existing metrics primarily rely on matching generated text against ground-truth references. This paradigm suffers from the ``one-to-many'' nature of video description, where high-quality captions are often penalized for lexical mismatches or valid shifts in visual focus. Furthermore, such assessments are typically one-dimensional, failing to provide a fine-grained analysis of caption quality. To address this, we redefine caption quality through the lens of information fidelity: A caption must maximize the coverage of salient visual information while ensuring strict factuality. We introduce CapQuiz, a novel reference-free benchmark that assesses captions based on their utility in answering human-verified, fine-grained, multiple-choice questions derived from the video. CapQuiz features a hierarchical taxonomy of 10 question types (spanning Descriptive and Inferential categories) across 24 diverse video domains. Extensive experiments demonstrate that CapQuiz correlates significantly better with human judgments than existing metrics and offers interpretable insights into model performance.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-10 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.09973 — 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 2026-10-16T10:02:34.344Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-10?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).