PUMA
Istituto di Scienza e Tecnologie dell'Informazione     
Di Bona S., Niemann H., Salvetti O. The brain matcher. Internal note IEI-B4-17, 1997.
 
 
Abstract
(English)
In this paper a method is presented for a data-driven comparison and registration of digitai images. The method is based on the definition of a novel neural network model. The 'Self Organizing Maps' idea, introduced by Teuvo Kohonen, is extended in order to modify an image considering both its morphometric and densitometric characteristics. An image, considered as a regular two-dimensional array, is manipulated assuming that its pixels are nodes of a network. A main goal is to compute the amount and thè progression of thè transformation events necessary for thè matching of two homologous images taken in different times and conditions. Two optimizing models are also presented. The algorithms developed have been tested on real complex cases: pairs of MR brain images of same patients, containing both normal and pathologic conditions (such as cancer or hematom), have been processed in order to characterise thè lesions themselves and to derive thè deformation laws interesting anatomical structures and their relationships.
Subject Signal processing
Segmentation
I.2.10 Vision and Scene Understanding : Representations, data structures, and transforms
I.4.6 Segmentation : Edge and feature detection
I.5.1 Pattern recognition


Icona documento 1) Download Document PDF


Icona documento Open access Icona documento Restricted Icona documento Private

 


Per ulteriori informazioni, contattare: Librarian http://puma.isti.cnr.it

Valid HTML 4.0 Transitional