Research Of Liver Image Segmentation Method And3D Modeling Method Based On Multilayer Spiral CT

Primary liver cancer is one of the common and harmful tumor to human healthwhich has a quick development speed and poor postoperative recovery. Thetraditional method of manual segmentation is entirely dependent on doctors’knowledge and experience which often lead to the failure of operation. The mainreason is that the the preoperative analysis and surgery are inconsistent. To someextent, the success rate of surgery will be improved, if the doctors can grasp the databefore operation with the accurately segmentation of CT sweep surface image andcalculate the three-dimensional modeling and volume by computer image processingand3D modeling.Thus, the combination in computer technology field of the imageprocessing and three-dimensional modeling with the anatomy and diagnostic inmedical filed has become a domestic hot topics. This topic devotes to achieve theintegration of liver CT image processing and modeling in order to complete the HCCto computer-aided medical.In this paper,the main research is multi-slice CT images of liver segmentation and3D modeling methods which based on the characteristics of computer-aided medicaldiagnosis and liver CT image.Firstly, a brief introduction on the background and purpose of multi-slice spiralCT liver segmentation method and three-dimensional modeling method. In additionalto, the current research situation will be described and analyzed in this chapter.Secondly, this paper has researched how to convert the images windowing of liver CTimage to BMP format images.The efficient filtering method not only retains the liverinternal information, but also enhances the edge portion of the liver. It is benefit to theliver segmentation. Moreover,this paper presents a RBF-CI (RBF-Confidence Interval)of liver segmentation method which based on the research of segmentation method inregion growing of a single liver and sequence of liver image.Based on the use ofwindow technology and algorithm of anisotropic diffusion filtering for pre-processing of the liver image, the RBF neural network learning algorithm is applied tocalculation of confidence intervals constant coefficients. It can reduce user interactionand adapt to liver image segmentation.For the real requirement of the liversegmentation map group,we study the liver image sequences segmentation andpresent a sequence segmentation of region growing algorithm which combinesmorphological dilation and corrosion in the characteristic of a single liver imagesegmentation.Finally, the efficient ray casting algorithm is a combination of spacejump technology of cloud algorithm and fast linear interpolation accelerationalgorithm. Through the improvement of spatial hopping technology, it can get asmuch as possible jump empty voxel. Moreover, through the fast linear differencemethod shortens the calculation time of sample point difference, which makes thewhole algorithm becomes fast and efficient and has no effect on the final imagingquality.