The linear attenuation coefficients as features of multiple energy CT image classification

We present in this paper an analysis of the linear attenuation coefficients as useful features of single and multiple energy CT images with the use of statistical pattern classification tools. We analyzed four CT images through two pointwise classifiers (the first classifier is based on the maximum-likelihood criterion and the second classifier is based on the k-means clustering algorithm) and one contextual Bayesian classifier (ICM algorithm – Iterated Conditional Modes) using an a priori Potts–Strauss model. A feature extraction procedure using the Jeffries–Matusita (J–M) distance and the Karhunen–Loève transformation was also performed. Both the classification and the feature selection procedures were found to be in agreement with the predicted discrimination given by the separation of the linear attenuation coefficient curves for different materials.