# VIABench

From the catalogue. Listed by one researcher from one source page.

- id: `viabench`
- kind: egocentric
- robot or device: first-person video from blind individuals (class: human egocentric)
- how collected: recorded or shared by visually impaired people
- size as stated: 761 videos; 46.9 hours; 14.5K annotations (per awesome list)
- hours: 46.9 | hours only claimed: none | episodes: not stated
- year: 2026
- organisation: Nanjing University
- licence: not stated (class: not stated)
- access: on GitHub (class: not stated)
- link (paper): https://arxiv.org/abs/2607.14660
- note: Assistance benchmark.

Confirm the licence at the link before relying on it.