In this paper, we propose a generic framework that can accurately detect a wide variety of activities from untrimmed videos. Our first ...
確定! 回上一頁