| 摘要 |
A method and a device for describing and capturing video objects are provided in the embodiments of the present application, including: capturing video images to generate video sequences, generating a video object tracking sequence (OTS) according to the video sequences, and generating video object descriptors (ODs) according to the video OTS and video sequences. Therefore, in a generated video object tracking sequence (OTS), the video object region tracking number (TID) is used to capture and track video objects. This makes it unnecessary to create a video object descriptor (OD) for each video object on a frame-by-frame basis. Therefore, the quantity of video ODs is reduced, thereby suiting the application requirement for the intelligent video interaction and accelerating the search of the video materials. |
