Metraipolis › archive › dc_de647a34a005
EgoActionBench: Benchmarking Dense Action Captioning for Egocentric Human Videos — arXiv:2611.03646
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.03646
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T18:56:42.932Z
- SHA-256 of the archived text
8fcbd4a9710575f58f09bc45bc5b622b384f7445c05eabdcef1483f09dbf60ea- Stable URL of this page
https://metraipolis.com/documento/dc_de647a34a005
How to cite
EgoActionBench: Benchmarking Dense Action Captioning for Egocentric Human Videos — arXiv:2611.03646 — arXiv (en), https://arxiv.org/abs/2611.03646 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_de647a34a005 (Metraipolis archive dc_de647a34a005, sha256 8fcbd4a9710575f58f09bc45bc5b622b384f7445c05eabdcef1483f09dbf60ea)
BibTeX
@misc{dcde647a34a005,
title = {EgoActionBench: Benchmarking Dense Action Captioning for Egocentric Human Videos — arXiv:2611.03646},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.03646},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_de647a34a005; sha256 8fcbd4a9710575f58f09bc45bc5b622b384f7445c05eabdcef1483f09dbf60ea},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.03646. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Learning from human data offers a scalable path toward embodied intelligence, yet it requires models to accurately understand fine-grained human actions from egocentric observations. However, existing benchmarks remain limited in action atomicity, temporal precision, and hand-action association. We introduce EgoActionBench, a benchmark that represents continuous human manipulation as temporally localized atomic actions associated with the left hand, right hand, or both hands. EgoActionBench contains 1,407 episodes with 41,132 human-refined action annotations. To evaluate such dense predictions, we propose an evaluation protocol that retrieves temporally eligible candidates, selects a prediction by content agreement, and scores semantic fidelity and temporal overlap alongside action precision and recall, jointly capturing what happens, when it happens, and which hand performs it. We benchmark 20 baseline models and systems on EgoActionBench and further adapt two backbones, Qwen3-VL-4B and LLaVA-OneVision-2-8B, through supervised fine-tuning followed by reinforcement learning on automatically annotated training episodes. EgoAction improves EgoActionScore by 15.9 and 23.2 points on the two backbones, and both adapted models outperform all open-source baselines. Our experiments reveal substantial limitations of current models in action atomicity, temporal precision, and hand-action association, highlighting the gap between general video understanding and the action understanding required for embodied intelligence. 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/2611.03646 — 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 2026-11-03T18:16:53.086Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).