Diabetes is a worldwide chronic metabolic disease. The complications of long-term hyperglycemia include cardiovascular disease, stroke, diabetic retinopathy and more. According to the American Academy of Ophthalmology, up to 45% of diabetic patients have some degree of "diabetic retinopathy," which is dysfunctional angiogenesis in the eye, which is the major cause of blindness in working-age adults the reason.
One of the challenges in screening for such eye diseases is the lack of trained professionals who can properly evaluate retinal images. If you can use machine analysis, it will greatly speed up the diagnosis and treatment. Recently, researchers from Singapore tried to use artificial intelligence to accomplish this task. Researchers collected 494,661 retinal images from China, India, Malaysia, and from the United States, belonging to Hispanic, African-American and white people, Learning Systems (DLS) were trained to identify and detect possible diabetic retinopathy, glaucoma and age-related macular degeneration (AMD) pictures. At the end of identification, the researchers compared the recognition of artificial intelligence and human evaluators. The results show that in these serious eye diseases, the artificial intelligence system has a high accuracy. The researchers believe that more research is needed to assess how DLS can be used in health-care settings to improve the outcome of vision care.
In medical treatment area, Acarbose , Sitagliptin Phosphate, Vildagliptin intermediates are all the effective drug for type II diabetes patients. Besides, Liraglutide is still under R&D. Recently, the U.S. FDA approved Sanofi's short-acting insulin injection Admelog (insulin lispro injection) for improving glycemic control in adults and children over 3 years of age. Hopefully, with the development of drug research and training the professional doctor who can properly evaluate retinal images, they will bright the diabetes patients the better lives.