Automatic Segmentation of Liver Region Employing Rib Cage and Its 3-D Displaying
Shinya Maeda, Masafumi Komatsu, Hyoungseop Kim, Akiyoshi Yamamoto, Koji Okuda · 2006 SICE-ICASE International Joint Conference · 2006
In recent years, computer aided diagnosis system is developed by using medical image processing techniques. Especially, computed tomography images have been used for diagnosis of liver disease or volume measurement for liver surgery. In CAD system, segmentation of a region of interests is one of the important techniques as pre-processing for extraction of abnormal area, in the abdominal CT image field, there are some regions that having ambiguous boundary such as liver, spleen, and gallbladder. It is difficult to segment such regions including some other soft tissue because one organ is very similar to the other, if a part of liver region is close to the tissue around rib bone, the boundary of region of interest area is very ambiguous. In this paper, we propose a technique for automatic segmentation of the liver region employing the rib cage area and active contour model on the abdominal CT images. By employing the rib cage information instead of ambiguous boundary, we can reduce the segmentation errors. Furthermore, we develop a system for 3D displaying of the segmented region. Experiment is performed employing 4 abdominal MDCT image sets and the obtained results are shown with some discussion