Design And Implementation Of Medical Image Segmentation System Based On Convolution Neural Network

Today,when the incidence and mortality of cancer are increasing year by year,people are paying more and more attention to health problems,and the hope for improving the timeliness and accuracy of tumor detection is becoming more and more urgent.With the rapid development of modern computer technology,convolutional neural network has become a research hotspot of experts and scholars at home and abroad.Applying its research results to the medical field helps doctors make efficient and accurate diagnosis of tumors.This has important social Meaning and practical value.Because tumors in different body parts and different examination methods have different shapes,this paper studies the segmentation model of MRI gliomas,and builds a Web system so that doctors can directly use the model.The system also adds functions of patient image browsing,patient image browsing marking,diagnosis result printing and system message.It is convenient for doctors to directly use the system for medical work in clinic.The main contributions of this paper are as follows:1.This paper studies the related theories and methods of MRI segmentation of brain gliomas.Through the research of CNN-based U-Net model and V-Net model,combined with the analysis of the BraTS19 data set,a preprocessing process for the BraTS19 data set is designed.And propose a segmentation model BraTS19-3DVNet based on V-Net.the experiment proves that the BraTS 19-3DVNet model in this paper performs well and can meet the requirements for the segmentation speed and accuracy of glioma in most scenarios and environments.2.This paper combines the doctor’s clinical diagnosis process,completes the system requirements analysis and system design.Paper confirms the system’s business requirements,user requirements,functional requirements,non-functional requirements,software architecture,functional modules and database solutions,and mainly divides the system into There are six functional modules for identity authentication,system management,patient data management,patient image analysis,diagnosis management and message management.This paper uses Spring Boot as the system implementation framework,completes the system function implementation and testing according to the design of each functional module,and performs performance testing on the system,and finally realizes an efficient,reliable and convenient Web system.Doctors can use this system to make quick and accurate judgments on the shape,size and boundary of the patient’s glioma,and improve the efficiency and accuracy of diagnosis.In addition,the system provides a good reference for the practical application of the convolutional neural network model in the medical field.