| The liver is one of the most important organ for human beings,liver disease,poses a great threat to human life and health,if you can before its deterioration,take corresponding measures,selection and use of scientific and reasonable means to implement efficient research and treatment for liver CT images,strict control of the liver disease misdiagnosis probability,has very important clinical value.This article in view of the current existing in the application of liver CT images in clinical liver disease extent and scope of lesions tell the fuzzy,it is difficult to identify the problems such as study,fuzzy,acquisition of liver CT images,and the actual situation of complex and the problems of image segmentation,the CT image preprocessing,and based on the random walk of the mainstream of image segmentation and image segmentation algorithm based on region growing and has carried on the comparative analysis,based on region growing image segmentation algorithm to improve the efficiency and effect of image segmentation;At the same time,there is low contrast of CT images for the liver,liver tissue and lesion area of gray distribution,overlapping phenomenon similar problems,proposed the enhancement algorithm based on the field transformation,solved the lesions of liver hemangioma grayscale and normal tissues of gray level difference is very small;Based on this,a diagnosis system of hepatic hemangioma was designed,and the results of experiment and application showed that the algorithm and system can achieve better recognition effect for liver hemangioma.The main contents of this paper are as follows:(1)collect liver CT image samples and make routine noise to noise.This paper analyzes the characteristics and effects of regional growth algorithm and random wandering algorithm,and finally determines the segmentation of liver parenchyma by using regional growth algorithm.(2)on hepatic hemangioma and CT of the liver parenchyma image difference is not obvious,liver hemangioma lesions area relatively fuzzy problems,this article first proposed in the traditional feature extraction on the basis of liver lesion area,combining with the field transformation algorithm,can effectively highlight the liver hemangioma lesions area,effectively improve the quality of liver hemangioma lesions area feature extraction.Shape feature of this paper,by using traditional simpler algorithm to extract the features of liver hemangioma,and extraction of liver hemangioma lesions of four characteristics,and to extract the characteristics of the principal component analysis,the first two principal components as the feature vector,so as to realize the feature vector dimension reduction,and compared with normal liver characteristics,and then choose to support vector machine classifier for feature recognition,in order to validate the liver hemangioma recognition effect in this paper,and combined with cyst of liver CT image recognition as a comparison,finally,according to the characteristics of three kinds of liver correct recognition rate is higher than 90%,classification recognition effect is better.Therefore,in this paper,on the basis of the force field conversion algorithm,using the traditional enhancement algorithm can effectively extract the simpler liver hemangioma lesions area,the feature information of the high quality,so as to solve the liver disease extent and scope of lesions tell the problem such as fuzzy,it is difficult to identify.(3)transform based on the field of liver hemangioma feature extraction and recognition technology applied to the diagnosis of liver hemangioma in the system,in the operation of the diagnosis system,can remove the noise,the liver parenchyma segmentation,effectively improving the quality of liver hemangioma lesions area feature extraction,effectively extract the characteristics of liver hemangioma,and achieve good recognition effect on the liver hemangioma. |
