Dr. Google might be your next ophthalmologist.
The search giant said Tuesday it had trained a deep-learning algorithm to spot signs of diabetic retinopathy in patients. It does this in a manner similar to eye specialists by checking pictures of the back of the eye for signs of the disease.
Google worked with a team of doctors in India and the US to create a dataset of 128,000 images and used them to train the to train a deep neural network -- computers wired like the brain -- to detect diabetic retinopathy. The results, published Tuesday in the Journal of the American Medical Association, were on par with a panel of board-certified ophthalmologists.
The algorithm could help patients get in front of doctors more quickly in the future, Google said, adding that more work was necessary before it could be used widely.
"Automated, highly accurate screening methods have the potential to assist doctors in evaluating more patients and quickly routing those who need help to a specialist," Dr. Lily Peng said in a blog post. "We hope this study will be one of many examples to come demonstrating the ability of machine learning to help solve important problems in healthcare."
Diabetic retinopathy afflicts people with diabetes and can lead to blindness. It affects blood vessels in the retina and is the most common cause of vision loss among diabetics, according to the National Institutes of Health.
This isn't Google's first foray into eye care. In 2014, the company teamed up with Switzerland's Novartis to develop smart contact lenses.