Cancer Detection Using Neuro Fuzzy Classifier in CT Images

Abstract

In this study, we have implemented an adaptive neuro fuzzy inference system (ANFIS) for detection of mass in CT images for early diagnosis of lung cancer. After completion of preprocessing and segmentation process four features have been extracted from images and given to ANFIS classifier as an input. The fuzzy system detects the severity of the lung nodules depends on IF-THEN rules. Feature based data set has been created with five fuzzy membership functions of each input. The proposed model is applied on more than 150 images and the computer added diagnosis (CAD) system achieved sensitivity of 97.27% and specificity of 95% with accuracy of 96.66%.