Distributed Semantic Video Tagging for Peer-to-Peer Authoring System

摘要

In the context of a project aimed at the development of a cooperative and distributed Authoring System to facilitate the production, sharing, and integration of new and existing content from various domains, this work addresses problems related to locating, annotating and reusing audio and video streams distributed over the Internet. The Authoring System allows partial annotations (applied to a part and not to the whole document) and time-based tagging of audio and video streams (where the semantic information is applied only to portions of the stream) and provides tools to extract partial contents and mixes them, creating new resources. These resources are represented with an envelope composed of many references that point to other contents (whole or portions of) within the Internet. This envelope holds only references where multimedia portions, called fragments, are located and they are stored as a meta-data. In such a way, the Authoring System prevents multimedia element duplications. Both original and newly created content are delivered to clients using the state of the art streaming technology that is integrated into the Authoring System to provide a uniform interface both to consume and produce knowledge.