Research On Medical Image Clustering Based On Gaussian Mixture Density Model

Image clustering has become a kind of key image recognition technology.Medical image recognition is important content of medical image analysis and understanding,which plays an important part in the field of medical clinical diagnosis.As a result,there is an important significance to research image clustering algorithm which is suitable for image recognition.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 clustering algorithm based on Gaussian mixture density model,which fits for medical image recognition.In this paper,it had researched Gaussian mixture model density and the clustering algorithm based on Gaussian mixture density model and constructed Gaussian mixture density model of medical image and proposed medical image density Gaussian mixture model parameter estimation algorithm based on the EM and its improvement.The main research work in this paper can be summarized as the following three aspects:(1) This paper systematically researched theory and methods of the parameter estimation and nonparametric estimation of the probability density function.In particular,it expanded on the parameter estimation of Gaussian mixture density model belonging to the theory and methods of half of the estimated parameters.It founded that the clustering algorithm based on Gaussian mixture density model was suitable for medical images in the clustering analysis.(2) For the problem of the model choice,this paper put forward the QAIC criterion function improved.The theory and experiments proved that the function was suitable for determining the weight of medical image Gaussian mixture density model.In the process of research,the use of heuristic method authenticated the correctness of the improvement of QAIC criteria.(3) This paper researched the medical image application of Gaussian mixture density model and presented the described method of medical image data distribution based on Gaussian mixture density model.(4) Aiming at the problem that the k-means initialized the parameters of Gaussian mixture density model that is sensitive to the parameter estimation,this paper presented the ant colony algorithm to improve k-means algorithm,which is applied for determining the initialization of the Gaussian mixture density model.Experiments had proved that the initialization algorithm improved for medical images would yield better clustering results.(5) Through the study of medical image data for each pixel on Gaussian mixture density model having the different contribution degree, this paper had proposed the weighted Gaussian mixture density model of the medical image and the medical image clustering algorithm based on the weighted Gaussian mixture density model.