Cross Media Content Based Information Retrieval Facility at DISIT

The trial of DISIT for Content Based Information Retrieval allows you to test the usage of different CBIR algorithms (selecting them from the list box that appears on the left side, once you selected the first image from the entry page) in finding similarities and thus in producing recommendations among images, video, documents, animations, and thus among CROSS MEDIA coming from ECLAP service and archive (Http://www.eclap.eu).


The demonstrator is going to compare against those presented other recommendations systems that you can find from ECLAP itself and which are presently based on text analysis. The full data set presently consists of about 100.000 objects coming from: images, documents and video cross media objects of ECLAP. The indexing model adopted an accelerator that allows you to obain the similar objects in a very fast manner and can scale up to millions of content items.

click here to GO at the DEMONSTRATOR


The algorithms used in this phase are mainly:


Paolo Nesi
DISIT: Distributed Systems and Internet Technology Lab
DINFO: Dipartimento di Ingegneria dell'Informazione
Universita' degli Studi di Firenze
Faculty of Engineering
Via S. Marta, 3; 50139 Firenze, ITALY 
E-mail: paolo.nesi@unifi.it
Office: +39-055-4796523, DISIT Lab: +39-055-4796567, +39-055-4796425
Fax.: +39-055-4796363 or +39-055-4796730

Il tuo voto: Nessuno