摘要 |
<p>A method and device for describing and capturing video object are disclosed in the embodiments of the present invention. The method includes concretely the following steps: generating a video sequence by capturing a video image; and generating a video object tracking sequence by the video sequence; then generating a video object descriptor by the generated video object tracking sequence and the video sequence. As a result of capturing and tracking the video object by a video object area tracking number in the structured video object tracking sequence, it does not need to establish the video object descriptor for each video object frame by frame, thus the number of the video object descriptor is reduced, and not only the application demand of the intelligent video interactive is well adapted, but also the search speed of the video data is accelerated.</p> |