Summary
International Symposium on Nonlinear Theory and its Applications
2008
Session Number:A3L-D
Session:
Number:A3L-D1
A Hybrid Neural System for ROI Selection in Microcalcification Detection
Laszlo Lasztovicza, Bela Pataki,
pp.-
Publication Date:2008/9/7
Online ISSN:2188-5079
Summary:
Microcalcifications appear as small bright spots varying in size and shape in mammographic images and they are usually grouped into clusters. An image can contain more than one cluster but the area of the clusters is still far smaller than the area of the breast. The aim of ROI (regions of interest) selection is to detect suspicious areas for further analysis. We present here an algorithm for ROI selection consisting three steps; image enhancement, feature selection and classification. The classification step of the proposed algorithm is performed by a hybrid neural system which uses supervised and unsupervised networks.