Discrimination of Solitary Pulmonary Nodules on CT Images Based on a Novel Automatic Weighted FCM

A novel automatic feature assessment and weighting Fuzzy C-Means (FCM) algorithm was proposed for the classification of solitary pulmonary nodules (SPN). Six pulmonary nodule features were extracted from computed tomography (CT) images, which were normalized and combined into feature sequence. The feature assessment method was used to calculate discriminative criterion of categories, where the sensitive features were selected and weighted to discriminate between benign and suspicious malignant pulmonary nodules. Forty CT slices of twenty three patients are selected to evaluate the proposed method. The experimental results show that the accuracy of discrimination is 86.3 %, the sensitiveness is 87.5 %, and the specificity is 80 %, which illustrate that the method is feasible, and have good accuracy and sensitivity.