Certainly, many of us know how dangerous Diabetes is, since ulcers associated with the disease can lead to limb amputation, while Diabetic retinopathy (DR) can make people blind. In order to avoid these two factors caused by the disease, the Google decided to study the possibilities for detecting diabetic retinopathy before it becomes a problem, and without the help of an on-site doctor.
Fortunately, these studies are having satisfactory results, since the disease is more easily diagnosed when examining a fundus image, the Google team has spent the last few years developing a data set of 128.000 individual images, examined by 3 to 7 ophthalmologists out of a total of 54. Marking the damaged areas of the eye - micro aneurysms, hemorrhages and the like, then these data were placed in a machine learning system (Artificial Intelligence), with this data, Google managed to build a highly reliable diagnostic tool. When tested with 12.000 images, the diagnosis of the system was “on par with that of ophthalmologists”.
With these results, Google hopes to expand the scope of this system to be able to diagnose the disease from more complex 3D images (generated from Optical Coherence Tomography). The team is still considering the possibility of automating the diagnostic process, to better serve patients in remote locations that might otherwise not have access to trained specialists.
Well there's Google by initiative.
