Atomic Visual Actions (AVA)

Emerging

1papers using it

2018first seen

The Atomic Visual Actions (AVA) dataset contains video clips annotated with human actions and is used to evaluate the recognition and localization of these actions in video clips.

🔎 Find this dataset

Papers using Atomic Visual Actions (AVA) (1)

Video Action Transformer Network2018

Atomic Visual Actions (AVA) — datasets — computer-vision