Research On Computer-aided Diagnosis Technology Of Nasopharyngeal Carcinoma Based On Magnetic Resonance Images

Early detection and accurate diagnosis is the key to improve the efficacy and survival rate of nasopharyngeal carcinoma.Computer-aided diagnosis technology can improve the diagnosis of nasopharyngeal carcinoma,helping doctors to make faster and more accurate diagnosis.In this paper,we focus on computer-aided diagnosis of nasopharyngeal carcinoma based on magnetic resonance imaging,and research the magnetic resonance image preprocessing,nasal mucosa region segmentation,feature extraction and classification,and achieved some research results.The main work of this paper is as follows:1,Due to the magnetic resonance imaging process may produce noise or artifacts,we select the appropriate method to remove the noise in the magnetic resonance image,while enhancing the target and background contrast.2,We focus on the accurate segmentation of the nasal mucosa region,preparing for subsequent classification.In this paper,we improve the Chan-Vese model.Firstly,we use the threshold segmentation method to construct the local image containing the nasopharyngeal mucosa region.Then,we use Chan-Vese model to segment the local image accurately,which can effectively combine the advantages of the two algorithms,and automatically and accurately divide the nasal mucosa area.The performance of the algorithm is evaluated from three aspects: reliability,accuracy and segmentation efficiency.The experimental results show that the method improves the speed and precision of segmentation.3,We study the feature extraction algorithm from the aspects of texture,shape and grayscale.In order to solve the problem of redundancy,we use the principal component analysis algorithm to reduce the dimensionality of the extracted features.4,We use the feature matrix obtained by dimensionality reduction as input to the support vector machine classifier,then we use the optimal parameters to classify the magnetic resonance images of nasopharyngeal carcinoma patients and normal people,and giving the intelligent diagnosis results.The value of this paper is to realize the intelligent diagnosis of nasopharyngeal carcinoma,which can be applied to the research and development of computer aided diagnosis device software system of nasopharyngeal carcinoma based on magnetic resonance image,which will produce great practical application value.