Streaming video media using semantic information

在互联网上传输视频是一个大问题。视频文件非常大，需要几个小时甚至几十个小时才能下载。等待内容太长了。在编码、缓冲区、协议、代理和服务器等底层结构上进行了大量的研究。这些技术要么不可扩展，要么不切实际。该项目旨在与CyberView合作开发一个使用语义信息的流媒体视频的高级方案。它将视频分成多个镜头，并允许对内容进行语义浏览。它还引入了具有各种行为的智能代理。

Delivering video over the Internet has posted a big problem. Video files are extremely big and need hours or even tens of hours to download. This is too long to wait for the content. Much research has been attempted over low-level structures such as coding, buffer, protocol, agent and server. These techniques are either not scalable or impractical. This project aims to collaborate with CyberView on developing a high-level scheme of streaming video using semantic information. It breaks videos into shots and allows semantic browsing of the content. It also introduces intelligent agents with various behaviours.