In 2017, a complete screening tool was developed using Google’s deep machine learning algorithm. The actual work on the tool began after researchers at Google published a paper in the US stating that their algorithms have achieved an accuracy of 98.6% in detecting diabetic retinopathy, on a par with the performance of ophthalmologists and retinal specialists. Google created a database of 1,28,000 images from different sight centers around the world. The database was created by grading each image, marking abnormal spots, lesions and indications of bleeding. Further, the data was assembled into a set of instructions built into a computer program to identify eye complications arising from diabetes. By analyzing millions of retinal scans showing signs of diabetic blindness, a neural network learns to identify the condition on its own. The machine predicts retina condition and advise on next steps of action in seconds.

This tool is being used by 71 vision centers across rural Tamil Nadu which have been established by Aravind Eye Hospital. These centers are supervised by trained technicians who take pictures of patients’ eye with retinal cameras and send the digital reports to doctors at Aravind Hospital. One of the challenges faced by the technicians and ophthalmologists is varying quality of retina images from different retina cameras. The hospital is working on minimizing the effects of this challenge before its final deployment. The Hospital has also started supplementing its Diabetic Retinopathy (DR) grading process with Google AI in 10 of its rural tele-consultation centers. In addition, the Hospital is in process of deploying AI-powered solutions for other diseases such as glaucoma and diabetic macular edema.

So far, the software only works for Swiss-German sign language. But our research suggests that the “architecture” of the system wouldn’t need to change to deal with other languages. It would just need more video recordings of each language to act as data to train it with.

An area of research we would like to explore is how we could use what the AI already knows to help it learn new languages. We’d also like to see how we can add other aspects of communication while using sign language, such as facial expressions.

At the moment, the software works best in a simple environment such as a classroom. But if we can develop it to tolerate more variation in the background of the video footage it is assessing it could become like many popular apps that allow you to learn a language wherever you are without the help of an expert. With this sort of technology being developed, it will soon be possible to make learning sign languages just as accessible to everyone as learning their spoken siblings.