Unsupervised segmentation of HRCT lung images using FDK clustering

Image segmentation is a prerequisite process for image content understanding in HRCT lung images for the development of a computer aided diagnosis (CAD) system. An unsupervised segmentation method is proposed in this paper. Initially, lung regions in HRCT lung images are separated and then feature vectors using the deviation in local variance of DCT coefficients are determined for each pixel of lung regions. A reduced set of feature vector is used for unsupervised classification using a rule based Fisher discriminant K-means (FDK) clustering algorithm.