from Hacker News

AVA: A Finely Labeled Video Dataset for Human Action Understanding

by hurrycane on 10/19/17, 6:57 PM with 6 comments

  • by SloopJon on 10/20/17, 5:57 PM

    From the download page:

    > The AVA dataset contains 192 videos split into 154 training and 38 test videos. Each video has 15 minutes annotated in 3 second intervals, resulting in 300 annotated segments.

    So basically this is a couple of CSV files annotating 192 videos, which are hosted on YouTube. ava_train_v1.0.csv is about 7 MB.

  • by lifeisstillgood on 10/20/17, 6:51 PM

    the most interesting thing i found was "We use movies as the source of AVA".

    while the datasets will only grow, movies are not realistic - they are by design faked, acted, well lit etc. While that is probably the best thing to do with a starting set i am waiting for the CNN/RNN to start saying (much like the early black female standford researcher who was not identified as human face) that person is not walking - i know walking, it's just like John Cleese.