Design For Image Segmentation Module Of Living Donor Liver Transplantation Surgery Planning System

Liver disease is one of the major diseases in China. Clinical treatment for liver disease has become a challenging problem. Living donor liver transplantation is considered to be an effective method to treat severe or end-stage liver disease. It is very important to assess the risks of the surgery and make an operative planning before the surgery. The traditional method of planning a living donor liver transplantation surgery uses a manual way to analyze the CT images, which is complicated, time consuming and inaccurate. With the development of medical visualization techniques, it is technically possible to reconstruct the 3D model of the liver according to the 2D CT slices, browse it interactively and analysis it. This is significantly in improving the successful rate of liver transplantation. However, domestic research institutes which engaged in the visual analysis and preoperative planning of liver transplantation are rare. Key technique and software are controlled by foreign countries.As one of the major modules of living donor liver transplantation surgery planning system developed by ourselves, this paper undertook the important task of segmenting 3D liver from the original CT slices. In computer vision, segmentation refers to the process of partitioning a digital image into multiple segments. The segmentation of the liver is the prerequisite of liver volume measurement and block segmentation. This paper fully investigated the functional requirements and divided this module hierarchically. We designed the high-level architecture and low-level state machine, then used several open source toolkits to implement 2D and 3D segmentation efficiently and accurately. At the same time, image annotation and measurement functions are provided. The annotation and segmentation results are compared and analyzed, perfectly accomplished the image segmentation function of the system.