Glaucoma Diagnosis and Prediction Based on Multi-Scale Feature Coding

Digital health has attracted a lot of attention as a result of the development of artificial intelligence (AI), big data, and fifth-generation (5G) telecommunications networks. AI, the Internet of Things (IoT), e-health, telemedicine, and big data analysis and application are all overlapping subfields of digital health. The World Health Organization (WHO) has a wealth of innovative opportunities in the field of digital medicine, and earlier this year issued a series of guidelines advising on how to leverage technology to provide evidence-based treatments for patients in practical situations. In order to promote the combination of artificial intelligence technology and practical application industry, especially clinical medicine, this paper conducts a series of studies on the diagnosis and prediction of glaucoma through deep convolutional neural network. The final results have certain clinical application value. Using AlexNet and LeNet two full convolutional neural network models, the Batch Normalization layer is added to the original model for normalization processing to speed up the convergence speed of the training model. The obtained glaucoma fundus retinal images are trained and compared. At the data set level, the original data is enhanced and expanded, and the model effects before and after expansion are compared again. The final experimental results show that the LeNet model has a better processing effect on simple images. The correct rate of the training set is 98 %, and the correct rate of the verification set is close to 100 %.