Statistical shape model of the liver and effective mode selection for classification of liver cirrhosis

In computational anatomy, statistical shape model is used for quantitative evaluation of the variations of an organ shape. Since liver cirrhosis will cause significant hepatic morphological changes, we applied statistical shape model of the liver to capture the morphological changes and recognize whether a liver is normal or abnormal. In this paper, we propose an effective mode selection method to improve the classification accuracy. In addition to the conventional Accumulated Variance Contribution Rate (AVCR) based mode selection, we newly propose a Pearson correlation based mode selection method and combine them to select the effective modes. The coefficients of the selected modes (components) are used as features to recognize whether liver is normal or abnormal. The effectiveness of the proposed method is evaluated by the classification accuracy of normal and abnormal. Experimental results show that our proposed method is superior than conventional methods.