The AVA dataset densely annotates 80 atomic visual actions in 57.6k movie clips with actions
localized in space and time, resulting in 210k action labels with multiple labels per human
occurring frequently. The main differences with existing video datasets are: (1) the
definition of atomic visual actions, which avoids collecting data for each and every complex
action; (2) precise spatio-temporal annotations with possibly multiple annotations for each
human; (3) the use of diverse, realistic video material (movies). Our goal is to accelerate
research on video action recognition.
AVA v1.0 is now available for download. We are preparing to release AVA v2.0 soon, which will include more densely sampled annotations and person links. Details can be found at our recent arXiv paper update.