Texture Analysis of Liver Tumor from Abdominal Computed Tomography in Computer Aided Diagnostic System

This paper proposes an automatic computer aided diagnostic system (CAD) for detection of liver diseases like hepatoma and hemangioma from Abdominal Computed tomography (CT) images. Liver and Lesion is segmented using gray level methods and clustering. Histogram analyzer is used to fix the threshold and morphological operation is used for post processing. Rules are applied to remove the obstacles. Fuzzy c-mean (FCM) clustering is used to extract the lesion from the segmented liver. First order statistical features, co-occurrence matrix based texture features, auto covariance features and Zernike moments are extracted from the segmented lesion. The sequential backward selection is applied to get the reduced feature set. The textual information obtained after feature reduction is used to train various neural networks such as a Probabilistic Neural Network (PNN) and Cascade feed forward BPN (CFBPN). The outcome obtained from neural networks is analyzed.