Segmentation of chest X-ray radiographs, a new robust solution

Abstract

The paper presents some new results of a Com puter Aided Diagnosis (CAD) system development project for automatic evaluation of chest X-ray radiographs. The main goal of the R&D project is to develop an intelligent advisory system for analyzing chest X-ray images and detecting lung nodules. This paper deals with an important preprocessing step: how to segment the X-ray images according to the ana tomical parts (determining the contours of the lung, heart, clavicles and the rib cage). For solving these tasks different approaches have been used so far in our research-development work. Here some new, improved and more robust solutions are presented. © 2011 Springer-Verlag Berlin Heidelberg.