Automatic Segmentation of Lung Fields from Radiographic Images of SARS Patients Using a New Graph Cuts Algorithm

This paper proposes an approach to the segmentation of lung fields in the severe acute respiratory syndrome (SARS) infected radiographic images, which is the first step towards a computer-aided diagnosis system. To overcome the segmentation difficulty of highly atypical property of SARS in the lung images, our algorithm first uses morphological operations to obtain the initial estimation of the regions where the lung boundaries lie in, and then applies a new graph-based optimization method to find the interested regions. The theoretical analysis shows that our approach is resistant to boundary discontinuity, noise, and large patches that affect the boundary search. Experimental results are given to demonstrate the good performance of our algorithm