High-level concept detection based on mid-level semantic information and contextual adaptation

Mylonas, P; Spyrou, E; Avrithis, Y

dc.contributor.author	Mylonas, P	en
dc.contributor.author	Spyrou, E	en
dc.contributor.author	Avrithis, Y	en
dc.date.accessioned	2014-03-01T02:44:39Z
dc.date.available	2014-03-01T02:44:39Z
dc.date.issued	2007	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/31930
dc.subject	Feature Extraction	en
dc.subject	Fuzzy Set	en
dc.subject	Hierarchical Clustering	en
dc.subject	Integrated Approach	en
dc.subject	Semantic Information	en
dc.subject	Unsupervised Learning	en
dc.subject	Neural Network	en
dc.subject.other	Algebra	en
dc.subject.other	Feature extraction	en
dc.subject.other	Fuzzy logic	en
dc.subject.other	Fuzzy sets	en
dc.subject.other	Multimedia services	en
dc.subject.other	Neural networks	en
dc.subject.other	Ontology	en
dc.subject.other	Semantics	en
dc.subject.other	Set theory	en
dc.subject.other	Concept detection	en
dc.subject.other	Contextual knowledge	en
dc.subject.other	Fuzzy algebra	en
dc.subject.other	Hierarchical clustering methods	en
dc.subject.other	Integrated approach	en
dc.subject.other	International (CO)	en
dc.subject.other	Low-level features	en
dc.subject.other	Media adaptation	en
dc.subject.other	Mid level features	en
dc.subject.other	P rior work	en
dc.subject.other	Personalization	en
dc.subject.other	Semantic information	en
dc.subject.other	Semantic multimedia analysis	en
dc.subject.other	Information theory	en
dc.title	High-level concept detection based on mid-level semantic information and contextual adaptation	en
heal.type	conferenceItem	en
heal.identifier.primary	10.1109/SMAP.2007.4414409	en
heal.identifier.secondary	http://dx.doi.org/10.1109/SMAP.2007.4414409	en
heal.identifier.secondary	4414409	en
heal.publicationDate	2007	en
heal.abstract	In this paper we propose the use of enhanced mid-level information, such as information obtained from the application of supervised or unsupervised learning methodologies on low-level characteristics, in order to improve semantic multimedia analysis. High-level, a priori contextual knowledge about the semantic meaning of objects and their low-level visual descriptions are combined in an integrated approach that handles in a uniform way the gap between semantics and low-level features. Prior work on low-level feature extraction is extended and a region thesaurus containing all mid-level features is constructed using a hierarchical clustering method. A model vector that contains the distances from each mid-level element is formed and a neural network-based detector is trained for each semantic concept. Contextual adaptation improves the quality of the produced results, by utilizing fuzzy algebra, fuzzy sets and relations. The novelty of the presented work is the contextdriven mid-level manipulation of region types, utilizing a domain-independent ontology infrastructure to handle the knowledge. Early experimental results are presented using data derived from the beach domain. © 2007 IEEE.	en
heal.journalName	SMAP07 - Second International Workshop on Semantic Media Adaptation and Personalization	en
dc.identifier.doi	10.1109/SMAP.2007.4414409	en
dc.identifier.spage	193	en
dc.identifier.epage	198	en