Minimum Energy Geometric Hermite Interpolation And Liver Ct Images

This thesis studies the Geometric Hermite Interpolation and the image processing on liver CT images.Geometric Hermite interpolation is an interpolation method that has great application prospects. This paper presents a new Geometric Hermite Interpolation method for the reason that most existed methods may not consider the shape adjustment and the fairing of the interpolation curves. This paper presents a class of quadratic spline functions and gives a G~1 -continuity geometric Hermite interpolation method with minimum energy constrained and adding a knot point in the parameter interval. It also gives the optimal solutions under the condition of the minimum energy and the necessary and sufficient conditions for getting the positive parameters. This method is an extension of geometric Hermite interpolation. It allows for easy computation and local modifications. Examples show this method to be effective.Analysis of CT liver images is another hot topic in nowadays. This thesis studies the segmentation method for CT images with multiple liver regions and the automatic segmentation algorithm of rib cage in CT images. The segmentation of liver CT images is one of the most important parts of the liver computer aided diagnosis system. It also continues to be large challenge in the image segmentation field. This paper presents a new segmentation method for CT images with multiple liver regions for the reason that most existed methods may only obtain single liver areas within such images. Firstly, this method applies an automatic double thresholding technique to perform initial segmentation, followed by the use of adjacent inter CT slice feature similarities as determinants to segment the multiple liver regions. Experiments show the method to be automatic,fast and effective in segmenting multiple liver regions. The rib cage can provide very important information when segmenting the liver on CT images. But the existed methods may not segment the rib cage accurate for the reason of the rib's uneven regional distribution. This thesis presents a new and automatic segmentation method of the rib cage using the information of adjacent several images. Examples show this method can segment the rib cage accurately.