Coronavirus rapid eye scan tests envisaged by German firm

Semic RF developed its scanning app with colleagues in the US and is currently pending approval from regulators. The firm hopes to roll it out in Germany by the end of next month, according to Semic's managing director, Wolfgang Gruber.

The app, called Semic EyeScan, uses a photo of the eye taken with a smartphone and identifies the virus through a symptomatic inflammation called “pink eye”. The firm has reportedly isolated Covid-19 from over two million different shades of pink, Gruber said.

He explained that the app - already tested on over 70,000 individuals - can process up to a million scans per second, with the option to expand that capacity further. The app could also potentially allow crowds back into mass-attendance events, such as concerts and football matches, with easy testing for spectators.

“You take your app, take a picture of both eyes, send it for evaluation and then you can have the evaluated result stored as a QR code on the tested person’s smartphone,” Gruber said.

Semic RF based the app on its 'Cradle' project, 'Digital Photography of Eyes', which started in 2014 with a university clinic in the US. The project evaluated conventional leukocoria (white pupillary reflex) examinations.

The firm incorporated their experience from the Cradle project into the EyeScan app and developed it further to investigate respiratory diseases in people of all ages.

In early 2020, the American Academy of Ophthalmology (AAO) recognised conjunctivitis (pink eye) in SARS-CoV-2, the virus strain that causes Covid-19.

With this knowledge, Semic RF examined data from 450 corona-infected patients with its deep learning system to determine if they were infected by Covid-19. The firm checked this through the app by scanning the eyes in video mode, to see whether the solution could reliably detect the Covid-19 infected conjunctiva.

In April 2020, the firm extended the software of the EyeScan app by adding 'YUV' colour model technology, a colour encoding system that encodes a colour image or video taking human perception into account. The firm also integrated a gamma compression method, or gamma correction, which controls the overall brightness of an image.

With the additional use of the YUV colour model technology and integration of the gamma compression method, the firm said it made further progress after evaluating the data of 70,000 patients. Here, they said the app successfully isolated diseases such as SARS-CoV-2 and diabetes with a hit rate of over 95 per cent.

The app is initially being targeted at companies and commercial users, at a cost of €480 (£415) per month, with a later rollout planned to private individuals.