Google uses neural network algorithm to detect diabetic retinopathy
On November 29th (local time in the United States), Google researchers published a paper in the Journal of the American Medical Association (JAMA), which showed that Google’s deep learning algorithm, after training through a large number of fundus image data, It is possible to detect diabetic retinopathy with an accuracy of more than 90%.
Google has been making big moves in diabetes. In September, Verily, an American life sciences company owned by Google's parent company Alphabet, teamed up with the well-known French pharmaceutical company Sanofi to invest about $500 million to build a diabetes joint venture, Onduo, which combines equipment and services. Verily's experience in microelectronics, analytics and consumer software development, Sanofi's clinical expertise and experience in bringing innovative therapies to diabetes, hopes to help people better make solutions that are good for everyday health, including From improving drug management, early prevention to habit improvement. It seems that in the prevention and treatment of diabetes, Google has made a lot of effort. This time, what surprise did it bring us?
An example of a retinal photograph for diabetic retinopathy (DR). A healthy retina can be seen on the left, the retina on the right has damage, and there is bleeding and fluid leakage in the eye.
More than 128,000 fundus image analyses, sensitivity and specificity analysis are over 90%
Diabetic retinopathy is an eye disease affecting diabetic patients and is the fastest cause of blindness. About 415 million diabetic patients worldwide are at risk of disease. If found early, the disease can be treated, otherwise it is highly likely to cause irreversible blindness.
One of the most common methods to detect diabetic eye disease is to the rear of the eye examined by a specialist photos, and determine if there are signs of disease, and if so, to identify the severity of the medical part of intuition and experience in the medical category. Although it is recommended that all diabetic patients should be screened annually, under normal conditions, many people do not receive good specialist care. This means that millions of people do not get the care they need to prevent vision loss.
A few years ago (specifically not leaking), a Google research team began to investigate whether machine learning can be used to screen for diabetic retinopathy (DR). The core is a deep learning algorithm that can detect signs of retinal photo lesions and may help doctors. Screen for more patients, especially in communities with limited resources and services.
4-Layer KN95 Protective Face Mask,Folding KN95 Face Mask,Dust Protect Kn95 Face Mask,3D KN95 Face Cover
Dongguan Keyutai Mask Co., Ltd. , https://www.maskkytai.com