Hepatic Vein Segmentation Methods Based On CT Images

Hepatic CT(Computed Tomography) scan is the principal means to diagnose the liver disease. The distribution of the hepatic vessels in the CT sequence images is the important basis for diagnosing diseases, developing a treatment and surgery plan. Extract the hepatic vessels by the image segmentation technology automatically and display the distribution of the hepatic vessels, which provide doctors more intuitive and convenient diagnostic information.Present a method of hepatic vessels segmentation and hepatic vein extraction based on the volume ratio of the vessels and liver,and use the ITK-SNAP software to realize the three-dimensional display of the hepatic vein. According to the frame of our hepatic vein segmentation algorithm, it mainly includes four parts: the preprocessing of the CT sequence images, the multi-scale Hessian enhancement, the vessels segmentation based on the volume ratio of vessels and liver, and the extraction of hepatic vein. The series of preprocessing operations consist of finding the range of slices needing to deal with, repairing the mask of the liver area, extracting the candidate hepatic vessels pixel and “Difference Enhancement”. After the preprocessing, use the multi-scale Hessian enhancement algorithm to enhance the tubular structure, which heighten the contrast between the hepatic vessels and the live parenchyma. In the content of the vessels segmentation, introduce the way to find the best threshold using binary search based on the volume ratio of the vessels and liver in the range of vessels gray. And use the best threshold to segment all the vessels in the liver automatically. What's more, use the morphology to thicken the vessels and connect the split. Finally, separate the hepatic vein from portal vein according to the topology structure and anatomical features so that we can extract the hepatic vein easily.Apply our method to many examples scanned by different CT scanners from different companies. The experiments show that the proposed method can segment the hepatic vessels and extract the hepatic vein automatically with good segment results.