Aug 4, 2017 | By Benedict
16-year-old student Kavya Kopparapu has designed a 3D printed eye diagnosis device. “Eyeagnosis” uses a camera to detect symptoms of diabetic retinopathy, a condition that affects one third of diabetes sufferers worldwide.
Eyeagnosis is a low-cost 3D printed tool for diagnosing diabetic retinopathy
Diabetes, a disease that prevents the body from producing insulin properly, affects 415 million people around the world. Of those 415 million, one third will develop diabetic retinopathy, a complication that damages blood vessels in the retina and which can ultimately lead to blindness. It’s scary stuff, and it needs to be caught to be stopped.
Unfortunately, care for diabetes around the world is drastically insufficient. Of the patients suffering with retinopathy, half will go undiagnosed, and those with the severest forms of the complication will eventually go blind.
The typical diagnostic procedure for diabetic retinopathy is a two-hour exam that requires an expensive retinal imager to make a thorough check of the patient’s eyes. And while this exam is effective at spotting the symptoms of retinopathy, it’s far from easily available.
That’s why Kavya Kopparapu, a 16-year-old student whose grandfather has retinopathy, was so keen to do something to help diabetes sufferers.
And help she has: the wunderkind, along with her 15-year-old brother Neeyanth and classmate Justin Zhang, has used 3D printing and a precocious knowledge of the sciences to create “Eyeagnosis,” a cheap tool for diagnosing retinopathy.
The youngsters, who call their fledgling organization “Ocular,” study at Thomas Jefferson High School for Science and Technology in Alexandria, Virginia, but have already taken their idea far and wide.
Diabetic retinopathy affects one third of those with diabetes
Ocular’s innovative device consists of a smartphone app and a 3D printed lens, and can be used to take close-up photographs of a patient’s eyes. An artificial intelligence system in the app can recognize visual symptoms of retinopathy in the photographs, offering a preliminary diagnosis that may otherwise be unavailable to patients.
“The lack of diagnosis is the biggest challenge,” Kopparapu told IEEE Spectrum. “In India, there are programs that send doctors into villages and slums, but there are a lot of patients and only so many ophthalmologists.”
Eyeagnosis came about after Kopparapu got in contact with a number of ophthalmologists, computational pathologists, biochemists, epidemiologists, neuroscientists, physicists, and machine learning experts via email.
After putting the information she received in order, the talented youngster decided to use a machine-learning architecture called a convolutional neural network (CNN) as a means of classifying the eye images taken by the camera.
In the CNN, information passes through layers of called nodes, and the network recognizes more particular features of the image with each layer.
Use of this kind of system is not without a level of irony: “It’s kind of funny that we’re using a system based on how the retinal system works to diagnose a retinal disease,” Kopparapu said.
The network used in Eyeagnosis is an off-the-shelf model developed by Microsoft called ResNet-50, supplemented with 34,000 retinal scans from the EyeGene database of the National Institutes of Health (NIH).
16-year-old Kavya Kopparapu created the diagnosis device
Interestingly, the low resolution of many of these images actually worked in Eyeagnosis’ favor, since the blurry pictures are representative of how patients’ own smartphone-captured images might turn out.
One year ago, Kopparapu finalized the ResNet-50 model so that it could spot the signs of diabetic retinopathy with the accuracy of a human pathologist. And by October 2016, she was in discussions with Aditya Jyot Eye Hospital in Mumbai to test out the app in a clinical setting.
In November, the first 3D printed prototype of the Eyeagnosis lens, which focuses the phone’s off-center flash to illuminate the retina as the photo is being taken, was sent to the hospital.
Five patients have already been accurately diagnosed using the Eyeagnosis device.
Much more testing will be needed before the 3D printed device can be proved fully reliable, but the early indications are highly positive for Kopparapu and her creation.
Kopparapu presented the 3D printed Eyeagnosis at the O’Reilly Artificial Intelligence conference in New York City last month.
“The device is ideal for making screening much more efficient and available to a broader population,” commented J. Fielding Hejtmancik, a visual diseases specialist at NIH.
Posted in 3D Printing Application
Maybe you also like:
- 3D scanning and printing play critical role in development of Arrinera Hussarya supercar
- New photoinitiator improves water-based DLP 3D printing, says Hebrew University of Jerusalem
- Three-storey 3D printed model of Dubai's 'The Tower' skyscraper erected in Dubai Mall
- Meet Surena Mini, Iran's new 3D printed humanoid robot that can dance and do karate
- 3D printed safe-cracking robot steals show at DEF CON, taking just 30 mins to crack SentrySafe
- Disney Research animates 3D printed models with cable-driven mechanisms
- The 'Better Call Saul' special effects team is using 3D printing to make incredible props
3Ders.org provides the latest news about 3D printing technology and 3D printers. We are now seven years old and have around 1.5 million unique visitors per month.
