Design And Implementation Of 3D Liver Segmentation System Based On Graph Cuts

Accurate segmentation of the liver is one of the hot areas of medical image processing.But the automatic segmentation often can not accurately split the liver in the current.Because compared to the automatic segmentation, interactive segmentation can achieve more accurate results,and the algorithm theory are more and more attention.Design and implementation of an interactive system for liver Computed Tomography( CT) image segmentation. Interactive graph cuts algorithm analysis the characteristic of Medical Digital Imaging and Communications file(Digital Imaging Communication in Medicine, DICOM) abdominal liver image and other merits of image segmentation algorithm, and combines the advantages of its own algorithm, by improving the algorithm, realization of 3D liver segmentation, can be more accurate. By converting the image to S-T diagram, the liver segmentation into energy minimization problem, then use the maximum flow / minimum cut to segment liver.The system can transform CT image sequence into volume data and parameter files and can be unified on the CT sequence reads.It implements a user-friendly and easy to use interface, through coordinate transformation of raw data,the image of cross-sectional, coronal plane, and Sagittal plane can be displayed separately.It allows users to use the interactive sketch tool easily,rough outline the objectives of the foreground and the background,and accurate segmentation of the liver out using graph cut algorithm.Finally, 3D reconstruction of the liver through Visualization ToolKit(VTK) volume rendering is convenient for users to observe the segmentation results in time.After initial experiments and tests,the system can interactive segmentation of the liver from the 3D CT images and results are more accurate.