Diabetic eyes spotted in just one click

Top News | Jane Cheung 7 Nov 2018

A start-up at the Hong Kong Science Park has developed technology that uses artificial intelligence to identify a "diabetic eye" in two seconds.

The smart health-care technology from Visiona MedTech International can detect diabetic retinopathy, a diabetes complication that affects eyes.

It is caused by damage to the blood vessels of the retina and can cause permanent blindness in serious cases.

The diagnostic process is similar to how people look into a machine to have their eyes checked for glasses. The scan is then uploaded to the analysis service platform.

It takes two seconds for the platform to analyze the images and determine if a person suffers from symptoms attributed to diabetic eye.

Reny Ng Ho-woon, chief executive of Visiona, said the AI in the system has studied hundreds of thousands of images of diabetic retinopathy.

"It memorizes all of the images it studies and is able to identify symptoms of a person with diabetic eye," he said.

Between April and June, the company cooperated with the Housing Society to help some 2,000 elderly residents at 20 public housing estates check their eyes.

More than 100 were found to have diabetic retinopathy, among which 92 percent did not know they suffered from the condition, while 46 percent did not even know they had diabetes.

Ng said once elderly people were found with symptoms of diabetic retinopathy, they were referred to receive a more comprehensive check by ophthalmology specialists.

Asked about the accuracy of the AI system, Ng said a third-party assessment showed it to be 96.8 percent accurate and safe to be used in community programs.

Ophthalmology specialist Jeffrey Pong Chiu-fai said there are about 700,000 Hongkongers suffering from diabetes - or 10 percent of the population. "There are 200 patients every year who have become blind because of diabetes," he said.

Pong said a diabetic patient has a 25 percent chance of being diagnosed with diabetic retinopathy after five years, and it rises to 80 percent after 15 years.

Ng said: "We hope to help citizens find out about the condition, so they can receive treatment as soon as possible."

Meanwhile, Growgreen, a science park incubatee, has developed a smart grower which allows people to grow their own vegetables at home.

The machine, the size of a microwave, has smart sensors to adjust light at the optimized wave-length for growing specific plants. It comes equipped with an automated water system.

Growgreen general manager Yvonne Chan said the machine is connected to a mobile app that gives users advice for growing plants and recipes.

She said under the optimal environment in the machine, vegetables grow faster than on a farm.


Search Archive

Advanced Search
July 2021

Today's Standard