1 dataset found

Groups: Video Action Recognition Organizations: No Organization

Filter Results

  • AVA v2.2

    The AVA v2.2 dataset for spatiotemporal action localization contains the bounding box annotations and the corresponding action labels on keyframes.