backtop


Print

New advances in Deep Learning help solving many important medical and healthcare problems.

By Otto Mergen

Diabetic retinopathy (DT), also known as diabetic eye disease, is a diabetes complication that affects the eyes. It's caused by damage to the blood vessels of the retina, the light-sensitive tissue at the back of the eye. It is a serious disease that can lead to irreversible blindness and affects up to 80 percent of all patients who have had diabetes for 20 years or more. Considering that there are 415 million diabetic patients worldwide, it poses serious risks.

However, proper treatment and monitoring of the eyes can substantially reduce the occurrence of DT. Unfortunately, medical specialists capable of detecting the disease are not available in many parts of the world where diabetes is prevalent. Usually specialists detect this disease by examining pictures of the back of the eye and rate them for disease presence and severity. This option is both costly, as it requires a highly-trained specialist to interpret images and in a case of many underdeveloped world is out of reach for many patients.

Now researchers at Google are trying to get around this problem by developing a deep learning algorithm capable of interpreting signs of DR in retinal photographs, potentially helping doctors screen more patients in settings with limited resources. 

https://4.bp.blogspot.com/-3E-DEWAF_fM/WD2ffT8I2RI/AAAAAAAABbo/rvuDPgSGEMs_7pZVhV-wL3svrOeqF05owCLcB/s640/image01.png
Source: Google Research

Researchers in collaboration with doctors trained their algorithms using a dataset of 128000 images each evaluated by 3-7 ophthalmologists from a panel of 54 doctors. Deep neural network algorithm then was tested on two separate clinical validation sets of about 12000 images with several ophthalmologists serving as the reference standard. Initial results are very exciting as algorithm’s performance is on-par with that of ophthalmologists. Although there are still ways to improve the software performance the preliminary results are quite encouraging.

Soon automated DR screening methods would be available to assist doctors in evaluating more patients and quickly refers those with signs of disease to a specialist. Currently, Google works with FDA to further evaluate their technology in clinical studies.
 




"This is from the DailyTech.com. It's a science website." -- Rush Limbaugh



Most Popular ArticlesSuper Hi- Vision Will Amaze the World
January 16, 2017, 9:53 AM
Samsung Chromebook Plus – Coming in February 2017
January 17, 2017, 12:01 AM
Samsung 2017 Handset’s Updates
January 17, 2017, 12:01 AM
Teclast Tbook X5 Pro - Coming January 23
January 16, 2017, 9:21 AM
Gionee Marathon M5 Plus – China’s Flagship Smartphone
January 15, 2017, 2:02 AM

Latest Blog Posts
Some new News
Saimin Nidarson - Jan 23, 2017, 8:59 AM
What is new?
Saimin Nidarson - Jan 22, 2017, 7:00 AM
News
Saimin Nidarson - Jan 20, 2017, 7:00 AM
News of the World
Saimin Nidarson - Jan 19, 2017, 7:00 AM
Some tips
Saimin Nidarson - Jan 17, 2017, 12:16 AM
News of the Day
DailyTech Staff - Jan 16, 2017, 12:10 PM
Tech News
Saimin Nidarson - Jan 15, 2017, 12:32 AM
Here is Some News
Saimin Nidarson - Jan 14, 2017, 12:39 AM
News around the world
Saimin Nidarson - Jan 12, 2017, 12:01 AM
Rumors and Announcements
Saimin Nidarson - Jan 11, 2017, 12:01 AM
Some news of Day
Saimin Nidarson - Jan 7, 2017, 12:01 AM
News 2017 CES
Saimin Nidarson - Jan 6, 2017, 12:01 AM






botimage
Copyright 2017 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki