Research And Implementation Of 3d Medical Image Segmentation Algorithm In Radiofrequency Ablation For Liver Tumor

Liver tumor has been a serious threat to the health and life of public as a high incidence of cancer in China.The anatomical structure of human body is complex and the incidence of various diseases is very high,therefore the study of computer-assisted intervention is of great significance.The technical process of virtual navigation system for liver tumor radiofrequency ablation has been firstly designed in this paper.By analyzing the advantages and disadvantages of the commonly segmentation algorithms,the skin,bone,tumor and blood vessel segmentation algorithms of abdominal threedimensional medical image have been designed and realized then.Finally design and realize the radiofrequency ablation system of liver tumor,and complete the function of various segmentation algorithms needed in the system.The innovation of the article is described in this paragraph.For the problem of incomplete bone tissue obtained by traditional regional growing method,seed point of the missing region is searched through comparing the initial segmented image with the original image for the iterative region growing,so that the traditional region growth method is improved,and this improved method has better effect on bone segmentation by detecting and iterating region growing.In addition,aiming at the problem that the evolution of the traditional level set curve may stagnate at the local extreme point and the evolution rate may be slow,a liver tumor segmentation algorithm based on the narrowed gradient correction level set is proposed.This method can improve the accuracy and rate of the algorithm through using morphological gradient correction and narrowband.The main research work of this paper is as follows:(1)Design the technical process of computer-aided interventional virtual navigation system under the background of radiofrequency ablation of liver tumor.The advantages and disadvantages of the commonly medical image segmentation algorithm are analyzed through experiments,including the threshold method,the region growing method,the level set method and the edge detection method.According to the analysis,the three-dimensional medical image segmentation module of the surgical virtual navigation system is designed,including skin,bone,tumor and blood vessel segmentation algorithm.(2)The design and experimental analysis of skin,bone,tumor and blood vessel segmentation algorithms are completed,including the skin segmentation algorithm based on edge tracing,the bone segmentation algorithm based on self-detecting iterative region growing,the liver tumor segmentation algorithm based on narrow-band morphological gradient correction level set,and the blood vessel segmentation algorithm based on Hessian matrix and multiple sub-point isolated connections.In this part,the traditional method of region growing is improved by using the method of detection and iteration,and the level set method is improved by combining morphological gradient correction and narrowband method.The practicability of the algorithm is also verified by experiments.(3)Design and realize the virtual navigation system of radiofrequency ablation for liver tumor.Including system function module design,system core function design and implementation,system interface design and implementation,besides the basic functions and algorithm functions has been implemented and tested too.