Title :
K-DIME: an affective image filtering system
Author :
Bianchi-Berthouze, Nadia
Author_Institution :
Univ. of Aizu, Japan
Abstract :
The Kansei (Japanese for sensitivity) Distributed Information Management Environment, or K-DIME, software environment lets users or software applications fetch multimedia material from the Web on the basis of textual keywords and filter it using a Kansei user model (KUM). The KUM learns a mapping between low-level features of the multimedia data and impression words - such as "romantic" - expressed in natural language. K-DIME acts as a metasearch engine for the Web.
Keywords :
Web sites; image retrieval; information filters; multimedia computing; natural languages; search engines; K-DIME; Kansei Distributed Information Management Environment; Kansei user model; Web; affective image filtering system; impression words; low-level features; metasearch engine; multimedia data; multimedia material; natural language; software environment; textual keywords; Airplanes; Color; Filtering; Image analysis; Image processing; Image retrieval; Neural networks; Neurofeedback; Search engines; Shape;
Journal_Title :
MultiMedia, IEEE
DOI :
10.1109/MMUL.2003.1218262
