Automatic screening method for diabetic macular edema based on transfer learning

Abstract

The invention provides an automatic screening method for diabetic macular edema based on transfer learning, which comprises the following steps: marking whether diabetic macular edema exists through an optical coherence tomography image and fundus color photograph; extracting eye features of a patient by using the trained deep residual neural artificial intelligence model, classifying whether diabetic macular edema exists according to the eye features, and performing polarity adjustment on the deep residual neural artificial intelligence model by using fundus color photograph data set trainingthrough a transfer learning technology; and obtaining a final intelligent screening model of diabetic macular edema based on fundus color photograph. The optical coherence tomography image and the fundus color photograph are used for identifying and classifying patients suffering from diabetic macular edema or not through the fundus color photograph-based intelligent screening model for diabeticmacular edema, so that the screening accuracy and the working efficiency of diabetic macular edema high-risk groups can be improved, and the expenditure of disease diagnosis and treatment of the patients is also reduced; therefore, more patients can be diagnosed and treated in time.

Description

technical field [0001] The invention relates to the technical field of medical application of artificial intelligence, in particular to an automatic screening method for diabetic macular edema based on transfer learning. Background technique [0002] Diabetic macular edema refers to the destruction of the blood-retinal barrier function in the macular area caused by diabetes, resulting in retinal thickening or hard exudative deposits caused by the accumulation of extracellular fluid in the macular area. The lesion first involves retinal capillary endothelial cells, that is, the blood-retinal inner barrier, and in severe cases damages the outer barrier, that is, pigment epithelium damage, resulting in edema that is difficult to eliminate. It is one of the main causes of vision loss in patients with diabetic retinopathy. Incidence was positively associated with diabetes duration, with a 4-year incidence of 5.4% and a clinically significant macular edema incidence of 7.2%. [...

AI Technical Summary

Problems solved by technology

Optical coherence tomography images in ophthalmology can be used in various aspects such as diagnosis and treatment of eye diseases, but its inspection equipment is more than one million, a

Images