Diagnosis And Analysis Of Diabetic Retinopathy Based On Deep Learning

At present,due to the relatively serious pollution of human living environment and the rapid pace of people’s life,the number of patients is increasing,and the etiology of diseases is becoming more and more complicated.As a result,human health is seriously threatened,and whether the patient is diagnosed with the disease is correctly diagnosed.It is particularly important.Taking type 2 diabetic retinopathy as an example,this paper proposes to combine traditional machine learning methods with deep learning methods,and designs and implements three models to solve two major problems.One is to analyze the characteristics of the disease and focus on outputting the disease.The importance of the characteristics of the sort;Second,the diagnosis of the disease.At present,in the field of disease diagnosis,deep learning methods rely more on acquired images for diagnostic analysis.This topic uses electronic medical record information to conduct related experiments.The main tasks are as follows:a)The establishment of a logistic regression model of disease characteristics,focusing on the importance of the characteristics of the order,to identify the most critical indicators of the impact of the disease,provide doctors with diagnostic reference.b)A diagnostic model based on a multi-layer perceptron was established to automatically and accurately diagnose the patient’s condition.The experimental results show that the model has achieved a diagnostic accuracy of 91.01%.c)A diagnosis model based on convolutional neural network-BNCNN is proposed.This model can effectively prevent gradient dispersion and accelerate network training.Experimental results show that the BNCNN model achieved a diagnostic accuracy of 97.6%.