Classification of Cirrhotic Liver in MRI Images Using Texture Features

概要

論文の詳細を見る
Cirrhosis of the liver is characterized by the presence of widespread nodules and fibrosis in the liver. The fibrosis and nodule formation causes distortion of the normal liver architecture, resulting in the exhibit of characteristic texture patterns. The purpose of this paper is to investigate texture features that can discriminate cirrhotic and non-cirrhotic liver in MRI images. A total of 128 features, including 126 co-occurrence matrix based texture features and two statistical measures are measured on ROIs extracted from the liver regions. The performance of the scheme is evaluated by using the ROC methodology. Based on a database consisting of 74 patient cases (25 cirrhotic; 49 non-cirrhotic), the sensitivity and specificity were found to be 80% and 60% for training and 80% and 63% for testing with a patient-hold-out method.
社団法人電子情報通信学会の論文
2007-01-20