Project description
The main goal of the project was the research and development of new video asset management systems, which considerably simplifies the handling of and retrieval from large sized video databases, suitable for professional users, but also for home users. The core technology used for this approach is based on descriptions of audio and video data derived from the MPEG-7 standard.
Very fast automatic analysis procedures decompose the video sequence in its components, the shots. A storyboard, displaying representative keyframes of the shots, gives a summary about the content of a video.
An important research field is the search for semantically similar shots. Many video segments, e.g. the weather forecast, are characterized by typical contents. These contents should be "learned" by the system. With the help of similarity searches previously "learned" video segments should be located and played back within a split second.

Project Data
Supported by: Bundesministerium für Wirtschaft und Arbeit
Project period: 01.04.2002 - 31.03.2005

Project partners