Metraipolis › archive › dc_49b9dc6f2e71

EgoExo-Next:Benchmarking Vision-Language Models on Visual-Option Next-State and Cross-View Reasoning — arXiv:2610.04506

Source
arXiv
Original
https://arxiv.org/abs/2610.04506
License (as archived)
CC0-1.0
Archived
2026-10-06T05:26:17.224Z
SHA-256 of the archived text
8be78aad41ac0c747d7792aff5d387487df49a16318db463aa51370d12802fe1
Stable URL of this page
https://metraipolis.com/documento/dc_49b9dc6f2e71

How to cite

EgoExo-Next:Benchmarking Vision-Language Models on Visual-Option Next-State and Cross-View Reasoning — arXiv:2610.04506 — arXiv, https://arxiv.org/abs/2610.04506 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_49b9dc6f2e71 (Metraipolis archive dc_49b9dc6f2e71, sha256 8be78aad41ac0c747d7792aff5d387487df49a16318db463aa51370d12802fe1)
BibTeX
@misc{dc49b9dc6f2e71,
  title = {EgoExo-Next:Benchmarking Vision-Language Models on Visual-Option Next-State and Cross-View Reasoning — arXiv:2610.04506},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.04506},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_49b9dc6f2e71; sha256 8be78aad41ac0c747d7792aff5d387487df49a16318db463aa51370d12802fe1},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.04506. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Vision-language models (VLMs) are increasingly evaluated for egocentric and cross-view video reasoning, yet existing benchmarks largely focus on semantic event understanding, temporal relations, or correspondence between already observed views, leaving their ability to reason directly about future visual states underexplored. We introduce EgoExo-Next, a visual-option benchmark for dynamic visual-state reasoning, where models must identify how an observed action trajectory subsequently appears rather than predict only an action label or textual description. EgoExo-Next contains 2,503 human-curated four-choice questions from six public egocentric and ego--exo video sources and comprises four interconnected subtasks that evaluate egocentric next-state prediction, bidirectional ego--exo state correspondence, exocentric next-state prediction, and their composition in Ego-to-Exo Next-State. Extensive evaluation of proprietary, open-source, and spatial reasoning VLMs reveals a substantial human--model gap, with the best model achieving 43.81\% average accuracy compared with 98.55\% for humans, and the largest degradation occurring on the composed Ego-to-Exo task. These results suggest that current VLMs remain substantially limited in dynamic visual-state reasoning, particularly when temporal progression and cross-view reasoning must be composed. The benchmark is publicly available at \url{ this https URL }.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.04506 — 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 2026-10-06T05:24:05.869Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).