Automated Segmentation of Suspicious Regions in Liver CT using FCM

In this study, an automated segmentation approach is proposed that segments the liver from input CT images of the liver, and cluster the components into different clusters, separating the liver from any detected suspicious region for further analysis and diagnostics. The proposed methodology relies on two stages mainly; firstly, the liver is segmented from the input CT image by separating it from other organs in the CT scan. This is done via a group of preprocessing stages such as morphological filtering, thresholding and boundary extraction algorithms. Secondly, the output from this stage is further processed using a FCM clustering technique to segment the pixels into different clusters. The proposed methodology is tested on a wide range of dataset liver CT scan images of both normal and infected livers. The success rate of the proposed methodology reached 93. 3% for infected livers, which proves the reliability of the proposed methodology and paves the way for further investigation.