A deep learning based method for ophthalmic disease diagnosis
Based on the fact of the high prevalence of ophthalmic diseases,especially Diabetic Detinopathy(DR),which is increasing year by year,a new diagnostic method for ophthalmic diseases is designed.This study uses the 2019 Kaggle diabetic retinopathy detection competition data set,adopts the advanced deep learning framework,builds a variety of advanced neural network models.After training,the prediction re-sults are obtained,and finally multiple models are integrated to improve the Kappa score,all owing the final model possessing strong robustness to solve the problem of model over fitting.This method includes func-tions such as disease diagnosis,automatic score analys is,etc.The experiment results show that this meth-od greatly shortens the diagnosis cycle of ophthalmologists,improves the accuracy and efficiency of diagno-sis,and provides a convenient and efficient diagnosis aid for ophthalmologists.
NETL
NSTL
万方数据
