由于输入视频序列的每一帧被分割成任意形状的视频对象平面(VOP)，这样每个VOP描述了一个语义意义的对象或所感兴趣的视频内容。

Each frame of the input sequence is segmented into arbitrarily shaped image regions (VOP's) such that each VOP describes one semantically meaningful object or video content of interest.

youdao

对于视频序列，则介绍了镜头检测、镜头内容表示、场景的语义描述等技术。

For video sequences, we introduce such techniques as shot detection, representation of shot content, semantic scene description.

youdao

最后，结合以上新方法，提出了一种基于MPEG - 7视频语义内容的层次化描述方案。

Finally, according to the new method, we propose a hierarchical video content description scheme under the international standard MPEG-7.

youdao