“Age is just a number.” Proving this saying right is 16-year-old Kavya Kopparapu from India who at this very young change has taken the innovation world by storm by creating a 3D-printed lens that can give a preliminary diagnosis for people with diabetic retinopathy. Kopparapu was encouraged to work on the lens when her own grandfather experienced symptoms of diabetic retinopathy.
For the uninitiated, diabetic retinopathy is considered as the most common cause of vision loss among diabetic patients. Its diagnosis involves a full-fledged two-hour exam and a specialised camera taking a photo of the retina of the eye. The main challenge of this disease is not its treatment but its lack of diagnosis, especially in a country like India with the second-largest population on Earth and not enough focus on healthcare. Though the country has several programs in motion where doctors are sent to rural areas, but as of today, the patient and ophthalmologists ratio is very discouraging.
Considering lack of diagnosis for the disease and the tiring nature of the current diagnosis exam, Kopparapu, along with her team set on a mission to train an AI system to identify diabetic retinopathy in photos instead. The 3D-printed lens invented by them when fitted onto a smartphone and used with its app, can give a preliminary diagnosis for people with diabetic retinopathy, considerably cutting down the current time it takes for a diagnosis.
At the young age of 16, Kopparapu is the founder and CEO of Girls Computing League, a nonprofit working to empower underrepresented groups in technology by fostering the interests of girls in computer science and technology. The Thomas Jefferson High School for Science and Technology student aims to remove the misconception that computer science is not a career for women and minorities by redefining what it means to be a computer scientist. Her team includes her brother and a high school classmate.
While developing the app and the 3-D lens, Kopparapu ensured that she took into account the expert opinion of ophthalmologists, neuroscientists, and experts in machine learning in order to come with the best diagnostic system possible. For the AI of the diagnostic system, the team decided to make use of a convolutional neural network, which is primarily a type of machine learning that has yielded successful results when applied to analysing images. The team employed a Microsoft developed network model, and used 34,000 retinal scans from the National Institutes of Health’s database in order to train the AI to accurately identify the types of eye diseases.
According to Kopparapu, the diagnostic system is capable of identifying diabetic retinopathy almost with the same accuracy of a doctor. In order to test the accuracy of their app, Eyeagnosis, the team joined hands with Aditya Jyot Eye Hospital in Mumbai last October. The app, which not only detects diabetic retinopathy in images, but also blood vessels and other details that would usually require a fluorescent dye injection, has already been tested on five patients at the hospital, and has rendered a 100 per cent result.
Kopparapu’s innovation came to centerstage when she recently presented the Eyeagnosis system at the O’Reilly Artificial Intelligence conference in New York in June. While Kopparapu and her team haven’t specified their expansion plans as of yet, but experts consider that the app and the lens system have real potential and can bring a revolution in diabetic retinopathy diagnosis.