Ferri, Massimo
(2013)
Graphic-based concept retrieval.
[Preprint]
Full text disponibile come:
Abstract
Two ways of expressing concepts in the context of image retrieval are presented. One, Keypics, is on the side of an image owner, who wants the image itself to be found on the Web; the second, Trittico, is on the side of the image searcher. Both are based on the paradigm of human intermediation for overcoming the semantic gap. Both require tools capable of qualitative analysis, and have been experimented by using persistent homology.
Abstract
Two ways of expressing concepts in the context of image retrieval are presented. One, Keypics, is on the side of an image owner, who wants the image itself to be found on the Web; the second, Trittico, is on the side of the image searcher. Both are based on the paradigm of human intermediation for overcoming the semantic gap. Both require tools capable of qualitative analysis, and have been experimented by using persistent homology.
Tipologia del documento
Preprint
Autori
Parole chiave
concept retrieval; image retrieval; persistent homology; size functions; keypics
Settori scientifico-disciplinari
DOI
Data di deposito
09 Mar 2017 08:32
Ultima modifica
09 Mar 2017 08:32
URI
Altri metadati
Tipologia del documento
Preprint
Autori
Parole chiave
concept retrieval; image retrieval; persistent homology; size functions; keypics
Settori scientifico-disciplinari
DOI
Data di deposito
09 Mar 2017 08:32
Ultima modifica
09 Mar 2017 08:32
URI
Statistica sui download
Statistica sui download
Gestione del documento: