News

The VITALAS project officially ended on March 31 2010.

The Video & image Indexing and reTrievAl in the LArge Scale project is no longer active. VITALAS (2007-2010) aimed to deliver a reliable and efficient pre-industrial prototype, allowing intelligent access to large multimedia professional archives, by addressing three major challenges:

• corss-media indexing (automatic annotation) and retrieval;
• large-scale search techniques; and
• visualization and context adaptation.

The project planned to deliver a pre-industrial search engine designed and validated by audiovisual professionals exhibiting functionalities that would allow interactive indexing and semi-automatic annotation, with interactive and personalized access to large-scale multimedia content resources. Having met all of the objectives, research and technology highlights of the VITALAS project include:

• Visual similarity search: functionality implemented according to VITALAS use case scenarios, including global similarity to an uploaded image across 20 million images;
• Cross-media (Concept, textual, visual) access with fused search including keyword suggestions;
• AV access to news videos, raw video structuring and exploration through video markers detection and annotations; automatic structuring and visualization of news content;
• Media content description & scalability including: similarity search on 1M images and 16,000 hours of video; local search on 100,000 images; speech search on 2,000 hours.

This website is no longer maintained. Information may be out of date. Some of the compenents of the VITALAS prototype have been carried into the large-scale integrating project "Access to Audiovisual Archives" (AXES), with the objective of investigating innovative solutions for audiovisual content exploration. Please visit the AXES website at the following address: http://www.axes-project.eu/