Invariant image classification using triple-correlation-based neural networks

Delopoulos, Anastasios; Tirakis, Andreas; Kollias, Stefanos

dc.contributor.author	Delopoulos, Anastasios	en
dc.contributor.author	Tirakis, Andreas	en
dc.contributor.author	Kollias, Stefanos	en
dc.date.accessioned	2014-03-01T01:09:56Z
dc.date.available	2014-03-01T01:09:56Z
dc.date.issued	1994	en
dc.identifier.issn	1045-9227	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/11251
dc.subject	Additive Noise	en
dc.subject	Efficient Implementation	en
dc.subject	Image Classification	en
dc.subject	Image Representation	en
dc.subject	Simulation Study	en
dc.subject	Neural Network	en
dc.subject.classification	Computer Science, Artificial Intelligence	en
dc.subject.classification	Computer Science, Hardware & Architecture	en
dc.subject.classification	Computer Science, Theory & Methods	en
dc.subject.classification	Engineering, Electrical & Electronic	en
dc.subject.other	Algorithms	en
dc.subject.other	Classification (of information)	en
dc.subject.other	Computer architecture	en
dc.subject.other	Correlation methods	en
dc.subject.other	Image analysis	en
dc.subject.other	Invariance	en
dc.subject.other	Mathematical models	en
dc.subject.other	Pattern recognition	en
dc.subject.other	Signal distortion	en
dc.subject.other	Invariant image classification	en
dc.subject.other	Multilayer perceptrons	en
dc.subject.other	Perceptron theory	en
dc.subject.other	Triple correlation based neural networks	en
dc.subject.other	Two dimensional gray scale images	en
dc.subject.other	Neural networks	en
dc.title	Invariant image classification using triple-correlation-based neural networks	en
heal.type	journalArticle	en
heal.identifier.primary	10.1109/72.286911	en
heal.identifier.secondary	http://dx.doi.org/10.1109/72.286911	en
heal.language	English	en
heal.publicationDate	1994	en
heal.abstract	Triple-correlation-based neural networks are introduced and used in this paper for invariant classification of two-dimensional gray scale images. Third-order correlations of an image are appropriately clustered, in spatial or spectral domain, to generate an equivalent image representation that is invariant with respect to translation, rotation, and dilation. An efficient implementation scheme is also proposed, which is robust to distortions, insensitive to additive noise, and classifies the original image using adequate neural network architectures applied directly to 2-D image representations. Third-order neural networks are shown to be a specific category of triple-correlation-based networks, applied either to binary or gray-scale images. A simulation study is given, which illustrates the theoretical developments, using synthetic and real image data.	en
heal.publisher	Publ by IEEE, Piscataway, NJ, United States	en
heal.journalName	IEEE Transactions on Neural Networks	en
dc.identifier.doi	10.1109/72.286911	en
dc.identifier.isi	ISI:A1994NR36000007	en
dc.identifier.volume	5	en
dc.identifier.issue	3	en
dc.identifier.spage	392	en
dc.identifier.epage	408	en