Liver extraction in the abdominal CT image by watershed segmentation algorithm

Abstract

In this paper, we proposed a liver extracting procedure for computer aided liver diagnosis system. An extraction of liver region is difficult due to interferences of other organs in the abdominal CT image. For this reason, liver region is extracted in ROI. ROI is selected by using the window for measuring the distribution of liver region’s Hounsfield Unit in abdominal CT image. The distribution is measured by using an existential probability of liver’s value in the window. If the probability of any window exceeded 50%, it would be assigned to ROI. Actually, liver region is not clearly discerned from the adjacent organs as muscle, spleen, and pancreas in abdominal CT image. Liver region is extracted by the watershed segmentation algorithm which is effective in this situation. This segmentation algorithm is very sensitive to the slight variance of contrast. So it generally produces over-segmentation regions. For optimal segmentation these regions are required to merge into the significant regions. Finally, a liver region is selected and extracted by prior information based on anatomic information.