Exploring The Eye in AI

October 17, 2019

Exploring The Eye in AI

World Sight Day is an annual occurrence that aims to increase global awareness surrounding vision impairment and blindness. According to the World Health Organisation, 80% of visual impairment is avoidable, preventable and/or treatable and currently 2.2 Billion people across the globe are visually impaired or blind. Although these statistics are troublesome, technological advancements focussing on Artificial Intelligence and Machine Learning are rapidly changing the playing field. 

Early Detection & Prevention

Diabetic retinopathy, a disease that damages the retina, is one of the leading causes of visual impairment and blindness across the globe. According to the European Society of Cardiology, there are currently over 425 million people living with diabetes, and by 2045 the number is expected to increase by 48%. Individuals that suffer from diabetes need to be screened regularly by an ophthalmologist, since early diagnosis can mitigate retinal damage. Unfortunately, it proves nearly impossible to screen each diabetic individual regularly, due to the international shortage of ophthalmologists. 

In an effort to detect diabetic retinopathy earlier, researchers at Deepmind and Moorfields Eye Hospital built an AI algorithm that can identify 50 eye conditions and make referral decisions on its own. The Algorithm is able to analyse an OCT retinal scan, identify signs of underlying eye conditions and determine the urgency of the condition at hand. The AI algorithm would be able to reduce the patient load for ophthalmologist and streamline earlier detection, but clinical trials are still being conducted to determine the accuracy of the algorithm’s diagnoses.

Noteworthy Apps & Projects

AI and Machine Learning are also used to assist the visually impaired with day to day tasks. In 2018, Google announced that they are working on an app called Lookout that is able to identify objects, read labels and signs, identify currencies and scan barcodes on behalf of the visually impaired. Lookout relies on AI technology to interpret visual stimuli, based on what your phone’s camera sees. In 2017, Microsoft launched a similar application called Seeing AI which transforms “the visual world into an audible experience”. Microsoft has also taken Seeing AI a step further by recently incorporating haptic feedback into the app’s design, allowing users to sense the edges of various objects within images on their mobile devices.

Project Torino is another noteworthy endeavour, which leverages AI and Machine Learning capabilities to teach visually impaired children how to code. The system involves pods and a programming language which assists in helping children between the ages of 7 and 11 learn the concepts of coding. Microsoft launched Project Torino so that individuals who face mental and physical challenges can familiarise themselves with computational thinking and to ensure that these individuals also have the opportunity to pursue a career that incorporates computer science. 

Seamless Integration using AWS

Amazon Web Services, one of Swipe’s partners, has also been utilised to enhance the lives of the visually impaired. Aira, a wearable smart glass technology that connects the visually impaired with Aira agents through AWS elements,  actively contributes towards making daily tasks more achievable for those who can’t rely on their vision. A combination of ML, IoT, AI and Augmented Context, allows for seamless integration and real-time assistance. 

So Where to From Here? 

When regarding the rapid advancement in technology over the past decade and the impressive breakthroughs achieved by means of AI and Machine Learning, one can only imagine what the future might hold in store.

Do you have a visionary idea that could add value to the lives of visually impaired individuals? Leave a comment down below!

Hendri Lategan


Swipe iX Newsletter

Subscribe to our email newsletter for useful tips and valuable resources, sent out monthly.