Detection method and device for retinopathy of prematurity based on deep neural network

Abstract

The method and device for detecting retinopathy of prematurity based on a deep neural network provided by the embodiments of the present invention belong to the field of image processing. The method collects a plurality of fundus image data first, and then labels the plurality of fundus image data based on preset rules to generate image data to be processed; then divides the image data to be processed into training sets according to a preset ratio and a test set; then set up a deep neural network model; then train the deep neural network model based on the training set; then process the data in the test set by the trained deep neural network model to obtain processed The output data; finally obtain the ROP lesion diagnosis result based on the output data. Therefore, human resources can be reduced, detection efficiency can be improved, part of the work of ophthalmologists can be saved, and it has important guiding significance for clinical detection of retinopathy of prematurity.

Description

Technical field [0001] The present invention relates to the field of image processing, and specifically to methods and devices for detecting retinopathy of prematurity based on deep neural networks. Background technique [0002] Retinopathy of prematurity (ROP) is a retinal vascular proliferative disease that mainly occurs in premature infants or low-birth-weight infants. With the popularity of neonatal intensive care units and the development of intensive care technology, premature infants and low-birth-weight infants The survival rate of underweight infants is gradually improving, and the number of infants suffering from ROP is also gradually increasing. ROP has great harm to children's vision. In addition to retinal degeneration, myopia, amblyopia, etc., severe cases can lead to lifelong blindness in children. Currently, in some hospitals with good medical conditions, for premature infants and low-weight infants who are at risk of ROP, the usual method is to use professi...

AI Technical Summary

Problems solved by technology

After the fundus image data collection is completed, professional ophthalmologists will diagnose this set of image data. However, due to the lack of professional ophthalmology departments in many maternity hospitals, it is difficult to obtain clinical opinions from experienced ophthalmologists in time even if fundus examinations can be performed.

Images