Relevance feedback approach for image retrieval combining support vector machines and adapted Gaussian mixture models

Marakakis, A; Siolas, G; Galatsanos, N; Likas, A; Stafylopatis, A

dc.contributor.author	Marakakis, A	en
dc.contributor.author	Siolas, G	en
dc.contributor.author	Galatsanos, N	en
dc.contributor.author	Likas, A	en
dc.contributor.author	Stafylopatis, A	en
dc.date.accessioned	2014-03-01T01:36:42Z
dc.date.available	2014-03-01T01:36:42Z
dc.date.issued	2011	en
dc.identifier.issn	1751-9659	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/21407
dc.subject.classification	Engineering, Electrical & Electronic	en
dc.subject.other	Common models	en
dc.subject.other	Content-Based Image Retrieval	en
dc.subject.other	Gaussian Mixture Model	en
dc.subject.other	Gaussian mixtures	en
dc.subject.other	Image database	en
dc.subject.other	Image representations	en
dc.subject.other	Kernel function	en
dc.subject.other	Kullback-Leibler	en
dc.subject.other	Negative examples	en
dc.subject.other	Numerical experiments	en
dc.subject.other	Relevance feedback	en
dc.subject.other	Feedback	en
dc.subject.other	Image retrieval	en
dc.subject.other	Probability distributions	en
dc.subject.other	Support vector machines	en
dc.title	Relevance feedback approach for image retrieval combining support vector machines and adapted Gaussian mixture models	en
heal.type	journalArticle	en
heal.identifier.primary	10.1049/iet-ipr.2009.0402	en
heal.identifier.secondary	http://dx.doi.org/10.1049/iet-ipr.2009.0402	en
heal.language	English	en
heal.publicationDate	2011	en
heal.abstract	A new relevance feedback (RF) approach for content-based image retrieval (CBIR) is presented, which uses Gaussian mixture (GM) models as image representations. The GM of each image is obtained as an adaptation of a universal GM which models the probability distribution of the features of the image database. In each RF round, the positive and negative examples provided by the user until the current round are used to train a support vector machine (SVM) to distinguish between the relevant and irrelevant images according to the preferences of the user. In order to quantify the similarity between two images represented as GMs, Kullback-Leibler (KL) approximations are employed, the computation of which can be further accelerated taking advantage from the fact that the GMs of the images are all refined from a common model. An appropriate kernel function, based on this distance between GMs, is used to make possible the incorporation of GMs in the SVM framework. Finally, comparative numerical experiments that demonstrate the merits of the proposed RF methodology and the advantages of using GMs for image modelling are provided. © 2011 The Institution of Engineering and Technology.	en
heal.publisher	INST ENGINEERING TECHNOLOGY-IET	en
heal.journalName	IET Image Processing	en
dc.identifier.doi	10.1049/iet-ipr.2009.0402	en
dc.identifier.isi	ISI:000292961300001	en
dc.identifier.volume	5	en
dc.identifier.issue	6	en
dc.identifier.spage	531	en
dc.identifier.epage	540	en