Metraipolis › archive › dc_87cbc987100f

VisER: Visual Evidence and Reliance for Object Hallucination Detection in LVLMs — arXiv:2608.30480

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.30480
License (as archived)
CC0-1.0
Archived
2026-10-17T08:40:00.710Z
SHA-256 of the archived text
9d7b3ab90a25c32072494dd583cbfa07b0bcd6d092dddea57820441ab13fbe00
Stable URL of this page
https://metraipolis.com/documento/dc_87cbc987100f

How to cite

VisER: Visual Evidence and Reliance for Object Hallucination Detection in LVLMs — arXiv:2608.30480 — arXiv (en), https://arxiv.org/abs/2608.30480 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_87cbc987100f (Metraipolis archive dc_87cbc987100f, sha256 9d7b3ab90a25c32072494dd583cbfa07b0bcd6d092dddea57820441ab13fbe00)
BibTeX
@misc{dc87cbc987100f,
  title = {VisER: Visual Evidence and Reliance for Object Hallucination Detection in LVLMs — arXiv:2608.30480},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.30480},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_87cbc987100f; sha256 9d7b3ab90a25c32072494dd583cbfa07b0bcd6d092dddea57820441ab13fbe00},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.30480. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Object hallucination remains a persistent reliability issue in large vision-language models, where generated object mentions may sound plausible but lack visual grounding. Recent training-free detectors use internal signals such as token likelihood, attention, visual confidence, or image-text similarity to identify hallucinated objects. These signals are useful, but they are often source-confounded. They measure how strongly an object is supported inside the model without distinguishing whether that support comes from object-specific visual evidence or the generated text prefix. In difficult cases, a hallucinated object can still receive high internal support because it fits the scene, is associated with nearby visual cues, or follows naturally from the generated text prefix. We propose VisER, a training-free two-sided metric for object-level hallucination detection. VisER evaluates each generated object mention from two complementary views. Visual Evidence measures whether object-context compatibility is backed by object-specific evidence from image tokens. Visual Reliance measures whether the object is supported more by the image than by the generated prefix. Combining these views gives a more source-aware grounding score, while avoiding additional object-level verification generations. Across multiple LVLMs and benchmarks, VisER improves AUROC and AUPR over a range of baselines.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-01 in cs.LG).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2608.30480 — 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 2026-10-17T08:13:26.943Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).