AN AUTOMATED SYSTEM FOR CLASSIFYING COMPUTED TOMOGRAPHIC (CT) LUNG IMAGES USING MARKOV Model (Report‪)‬

Byline: Hanan M. Amer, Fatma E.Z. Abou-Chadi and Marwa Obayya ABSTRACT: In this paper, the computed topographic (CT) images were utilized to discriminate different lung abnormalities. The performance of traditional feature extraction techniques and wavelet-domain based estimators was compared using Hidden Markov Model (HMM). This was done by analyzing data recorded for healthy subjects and patients suffering from lung cancer and emphysema diseases. The techniques utilized included statistical, intensity, and morphological features as well as features derived from texture analysis, and wavelet domain-based features. Results have shown that using wavelet domain estimators gives higher rates for classifying lung abnormalities. Classification rate reaches about 99.3%.