Font Size: a A A

Gaussian Mixture Model And Clustering For Medical Image

Posted on:2011-04-07Degree:MasterType:Thesis
Country:ChinaCandidate:S W ZhangFull Text:PDF
GTID:2178360302993719Subject:Biomedical engineering
Abstract/Summary:PDF Full Text Request
Medical image recognition is the vital content of automatic diagnosis by medical images and also an important research direction of medical fields all over the world. Medical image clusting is important content of image recognition,which plays an important part in the field of medical clinical diagnosis.It is realistically significant to research image clustering algorithm which fit to medical image. At present, the medical image clustering algorithm has not yet achieved the desired effect of identification and can not fully meet the requirements of the medical image analysis and understanding. In this paper, it attempted to research the initialization of EM algorithm and clustering algorithm based on double Gaussian mixture model, which fits for medical image recognition.In this paper,we research on the initialization of EM algorithm and double Gaussian mixture model. The approximate density initialization of EM algorithm and the medical image clusting based on double gaussian mixture model were proposed.The main contents include following several aspects.(1) A systematic study on the EM algorithm,the theory and method of Gaussian mixture model,indicates that the parameter estimation based on Gaussian mixture density model is semi-parametric estimation.(2) In this paper ,we researched theory and methods of the approximate density function and mixture density function systematically. The performance of EM algorithm heavily depends on the initial values of the parameters in EM. In this paper,The approximate density function is adopted to initialize EM.(3) Made a comparison to approximate density initialization,Kmeans initialization and random initialization method. The application of these parameters in analysis of Gaussian Mixture Desity Mode based on real human abdomen medical images and the results of experiments show that it can achieve better effect than Kmeans and random initialization.(4)This paper,the clustering methods of medical images are deeply researched.The medical clusting methed based on double gaussian mixture mode is proposed. The results of experiments show that this method is better for medical image clustering.
Keywords/Search Tags:double gaussian mixture model, initialization, Expectation Maximization algorithm, approximate density
PDF Full Text Request
Related items