Plaque tissue classification in intracoronary ultrasound images using hirarchical and flat neural classifiers

摘要

In this paper we present a tissue classification system for introacoronary ultrasound images.The presented project aims to develop a viewing stationwhich supports the physician by labeling the ultrasound image regions with their coresponding tissue types.After a feature generationstep with the help of texture analysis and frequency features a neural classification is done.Here we used flat neural classifiers like multilayer-perceptrons and radial basis -function nets.As the systme will have to distinguish many different tissue types,we also used a hierarchical and modular apporach.This is very promising especially if different classifiers can be integrated.