Computerized Distinction of Improved Fuzzy Support Machine for Imageology Character of Benign and Malignant Pulmonary Nodules

摘要

In order to improve the accuracy of the solitary pulmonary nodule diagnosis with medical signs in medical imaging diagnostics, a novel computer-aided classification method is developed. In the view of the existing problems in the lung cancer diagnosis such as the large number of data and the low diagnose efficiency. In order to solve the problem, a new classification method based on the Fuzzy Support Vector Machine (FSVM) was developed to choose the lung with suspicious lesion at an early stage. In this method, the membership function was improved based on the spectral clustering theory which ensures each sample has two membership degrees that guarantees the class of the specific sample more reasonably. The proposed method was used to classify benign and malignant of the pulmonary nodules, the parameters show this method can distinguish the noise and outliers samples more effectively, compared with the traditional fuzzy support vector machine method. Thus, the results illustrated the robust to noise capability and the effective classification ability of this method.