Research And Implementation Of Cancer Risk Prediction Based On Convolutional Neural

With the rapid development of the imaging euqiments and computer science technology,medical imaging plays an increasingly important role in disease diagnosis and prevention.However,it is undeniable that medical images often contain large amouts of data.Professional interpretation often requires long hours of professional accumulation.Deep learning relied on superior storage and computing capabilities,can exceed the accuracy of professional doctors to improve the effiency of disease diagnosis,which providing a new method for the traditional medical diagnosis.Breast cancer is a malignant disease that affects health.And early diagnosis is an effective means to improve the cure rate.This project uses the convolutional neural network algorithm in deep learning area to anlyze the pathological images of these diseases,which can provide a scientific basis for the treatment of breast cancer.First of all,this paper studies the algorithms of convolutional neural network in deep learning and puts up with a solution to the classification of high resolution medical images of breast cancer.When preprocessing the original image data,the traditional digital image processing method is used to obtain small image patches.Then design and train VGG convolutional neural network and the residual convolutional neural network in MXNet deep learning training architecture.By evaluating the evaluation index of two models and analyzing the ROC curves,the paper proposes a medical image classification model which is suitable for this project.Second,in the field of breast cancer diagnostic medicine,the raw data is panoramic slice.In order to make the model in the project more ubterpretable,an end-to-end breast cancer risk screening system is designed with the help of convolutional neural network model and machine learning model.And then evaluates the experimental results.The paper verifies that the deep residual convolutional neural network can achieve the accurate classification in medical image data.With digital image processing technology and logistic regression model,the diagnosis of breast cancer can finally be realized.The research results provide some references for the early prevention and early diagnosis of breast cancer.