Lung Segmentation Methods in X-ray CT Images
José Silvestre Silva, Augusto Silva, Beatriz Sousa Santos, Departamento de Física, Rua Larga · 2000
In this paper we present three methods for the automatic segmentation of pulmonary regions in X-ray CT images. Before applying any method, each image is pre–processed. The first two methods rely on intensity oriented image processing techniques based on the mapping between the peaks and valleys of the histogram and groups of organs located in the thoracic region. The second method also uses morphologic filters to remove spurious structures. The third method explores the technique of active contours applied to each one of the lungs. The resulting contours obtained by each method are qualitatively compared with corresponding reference contours drawn by two expert radiologists.