Google teaches how to detect AI in time blindness from diabetes


We know that Google believes strongly in ‘artificial intelligence used in the medical field. A few days ago we reported DeepMind division of the alliance with the British health service, but the commitment does not stop there. On his blog dedicated to the research, Google announced had turned their attention to blindness caused by diabetes .

Diabetes is the fastest growing cause of blindness in the world. The Diabetic retinopathy can be treated and prevented if detected in time, but doctors can do that are not found anywhere in the world; here is where comes in artificial intelligence.

The first symptoms Diabetic retinopathy can be seen by analyzing the picture of the patient’s retina. L ‘ algorithm developed by Google, the details of which have been published JAMA Network it is able to recognize them individually, without the presence / assistance of a physician ; Google has worked closely with doctors in both India and the United States, and has trained artificial intelligence with 128,000 images in which damaged areas of the retina were highlighted. When tested in the field, with 12,000 images, the artificial intelligence has returned similar diagnosis to those of ophthalmologists.

The team defined the exciting results, but at the same time , it acknowledged that the road ahead is still long . The intention is to expand the system’s capabilities by analyzing even the most complex 3D images obtained by OCT ( Optical coherence tomography ), in addition to the standard picture obtained from ophthalmoscope or slit lamps with special lenses. The other objective is to automate the diagnostic process, so as to be more effective for patients in remote locations that would have trouble getting doctors in the flesh.


[wpdreams_rpp id=0]