The Application of Fuzzy Reasoning and Biclustering in Ultrasound Breast Tumor Classification

Fan Zhang, Qinghua Huang, Xuelong Li

Abstract: The breast tumor is a common disease for females. In the paper, we designed a computer-aided diagnosis (CAD) method for breast tumor which employed the biclustering algorithm and fuzzy inference. The specialist of Sun Yat-sen University helped us collect 500 breast tumor cases. The BI-RADS feature scoring standard was utilized to extract the useful information from ultrasound images. Compared with previous method, the diagnosis process of the proposed method is more easy to be understood. Meanwhile, the performance of the method in this paper is not restricted by the type of image source. The experiment shows that the performance of our method exceeded the previous method which also employed the BI-RADS information in accuracy, positive predictive value (PPV), negative predictive value (NVP), and sensitivity. According to the experiment, the average accuracy achieves 88. S7%, PPV is 86.45%, NVP is 94.24%, specificity reaches 76.55% and sensitivity is 96.85%. The result means that the method in this paper can distinguish the malignant breast tumor precisely.