Research On Intelligent Evaluation System Of Magnetic Resonance Quality Control

Magnetic resonance imaging(MRI)quality control(QC)refers to detection of MRI image quality parameters through a series of technical procedures,the test result reflects the performance of state imaging equipment,which enables clinical engineer to make a timely response in the face of the imaging system's performance change,ensuring that the MRI system was operated in normal,stable and efficient environment.Traditional quality control procedures relys on manual inspection which has many disadvantages such as time-consuming and low accuracy,and the procedure is not conducive to the development and popularization of QC.In view of the above problems,this paper develops the intelligent evaluation system of magnetic resonance quality control,which is based on the MATLAB platform and combined with digital image processing theory.This system realizes the accurate,automatic and fast detection of MRI image quality parameters,and reflects the performance status of MRI system objectively.Compared with the traditional detection methods,it has great superiority and potential.First of all,this paper expounds the importance of the quality control of MRI system based on the operation requirements of clinical MRI system,and makes an analysis of the domestic and foreign research status.Some important MRI image quality parameters are illustrated in this paper,and the traditional detection method of quality control is introduced briefly.In view of the deficiency of traditional detection methods,this paper discusses the purpose and significance of this study.Meanwhile,the theory and the concept of digital image processing are introduced in this paper.Before the design of intelligent evaluation system,this paper discusses the image preprocessing technology which was commonly used in system design,and makes the improvement based on the actual application environment.For image binarization,this paper introduces the global threshold and local threshold binarization method respectively.In the light of the quality problem for gray local high image,an image enhancement method based on Curvelet transform is proposed.Through the adjustment of Curvelet coefficient set,the distribution of image gray level was improved,and then the Otsu is used for image binarization processing.In this paper,the image processing technology,including mathematical morphology processing,is also introduced in detail.Secondly,contraposing the SNR and degree of homogeneity of intelligent detection submodule,this paper explores the distribution character of magnetic resonance signal and noise,and gives the correction factor of the noise standard deviation calculation according to the MRI image data acquisition and the principle of reconstruction.On this basis,the automatic detection of SNR and degree of homogeneity is achieved by a series of image processing techniques such as seeded region growing,linear smoothing filter and so on.In the sub module of spatial linearity detection,this article uses Hough transform circle detection to extract the target circular hole,and puts forward the improved method of random Hough transform circle detection,which uses the gradient to reduce some problems like invalid accumulation in order to raise efficiency of circle detection.And in the sub module of spatial resolution detection,this paper adopts template matching technology to realize the recognition of the strip shape objection,and uses the prior knowledge like position to improve the matching efficiency.In the last,this paper integrates the above-mentioned submodule of parameters intelligent detection into a complete intelligent evaluation system of MRI quality control.This system is able to using related image processing technology to carry on the intelligent analysis to the imported images of MRI quality control,it successfully replaces the traditional manual testing process,and realizes the accurate,automatic and fast detection to the image quality parameters.Besides,the intelligent evaluation system overcomes the defects of traditional detection methods,while liberating the clinic impact posttreatment workstation.It plays an positive and important role in development and popularization of related work of the quality control.