Technical reports

Search command : =""

Current Record: 12 of 13

[Add to Basket]
Identifier	uch.csd.tcl//2000sifakis
Title	Τμηματοποίηση εικόνων από χαρακτηριστικά χρώματος και υφής με χρήση στατιστικών μοντέλων και μεθόδων ΜΔΕ
Alternative Title	Image segmentation by color and texture feature using statistical models and PDE methods
Author	Σηφάκης, Ευτύχης Δ
Abstract	This report presents the design and impementation of an image segmentation system baeed on the joint usage of intensity, color and texture features. Intensity and color features consist of the respective normalized local histograms, while the local variances of the Disrete Wavelet Frames analysis components are used as the texture descriptors. A hierarchical clustering technique is applied on square image blocks to estimate the features of the prototype classes. Following the statistical modeling of the corresponding features, the Bhattacharya metric is used to yield high confidence classifications of image pixels into the extracted classes. The level set based Multi-Label Fast Marching algorithm is used for the propagation of initial classification decisions towards the rest of the image.
Language	Greek
Issue date	2000-07-13
Collection	School/Department--School of Sciences and Engineering--Department of Computer Science--Technical reports
	Type of Work--Technical reports
Permanent Link	https://elocus.lib.uoc.gr//dlib/b/9/a/metadata-dlib-2000sifakis.tkl
Views	798

Digital Documents
	Download document View document Views : 10