A large-scale, high-quality dataset of URL links to approximately 650000 video clips that covers 700 human action classes, including human-object ...
確定! 回上一頁