Our research efforts into Content-Based Video Indexing started with a need for the SCOLA project video archive system.
The focus of our work is to integrate a useful indexing and query system. To realize that we need to process video and audio data to extract keywords that provide the descriptive metadata for the associated video clip. These metadata keywords can then be used in a search engine for finding appropriate video clips that match the user's search phrases.

An extension to this project is to mine the metadata catalog, with the goal of establishing other relations out of the existing data. Such data mining systems are particularly useful for the intelligence community.

A particular challenge for our project is to process video in real-time. All processing steps for the data extraction must be performed while the video stream is received by the system.