A fully automated Liver and HCC Tumor segmentation using cell density and Morphological Operations

A fully automated Liver and HCC Tumor segmentation using cell density and Morphological Operations

An early detection and diagnosis of hepatocellular carcinoma (HCC) is most discriminating step in liver cancer management. Image processing plays very vital role to early detection and diagnosis of HCC. The fast and accurate CT liver images segmentation is a very challenging in detection, diagnosis, clinical studies and treatment planning of HCC. The purpose of this research is to develop an automatic HCC detection and diagnosis system that detect and segment all type of HCC lesion from liver CT images with maximum sensitivity and minimum specificity. The proposed method planned an automatic liver segmentation and then HCC lesion segmentation. In first step, we created mask based on chosen histogram thresholds to predict best frame and segmentation of liver using prior knowledge of location and shape. Secondly proposed HCC lesion based on 3D with combinations of edge detection techniques, histogram analysis, and morphological processing to detect and segment the HCC lesions. The experiments were carried out with 31 CT cases involving 18 HCC lesion and 13 non lesion. The simulation and experiments result demonstrated high accuracy 98% and at low 2% false positive which compared with manual segmentation done by radiologist.

WE ACCEPT