TECH4ALL: Helping Children See the Future with TrackAI

There are around 19 million visually disabled children in the world.

But it is not only a visual issue: Take a moment to imagine how tough it is for a young child to struggle to run into the playground, catch up his or her friends, or even follow teachers’ instructions.

And so it is for a baby, who is starting to explore the world and learn from it. We need vision to figure out how life works, to interact with others, and even to understand their feelings.

Visual impairment interferes with children’s general development and social lives. Many of them will be excluded from educational opportunities, and in certain settings, up to 60% of children becoming blind will die during the following two years.

The Missed Potential for Prevention & Cure

And here’s the most striking fact: 70% of these 19 million children could have a normal vision if their problem had been detected early enough.

The problem is that congenital or progressive low vision can go unnoticed in very young children. Babies get used to it and are able to adapt to almost any circumstances. So, how can we tell whether a baby is seeing properly or not? With current tools only highly experienced ophthalmologists can assess visual functioning in very young children.

I’m an ophthalmologist, and I’ll tell you how we do it: We show children different stimuli (specific objects or pictures) and observe how they react. Based on their behavior, we assess whether they can see properly or not. This is far from ideal and as a result, many visual problems remain undetected until it’s too late.

We aim to improve that.

Visual screening programs are the best tool to decrease the rate of undiagnosed visual disorders and, therefore, the rate of visual impairment in childhood. We need to empower non-trained people from the community to identify children who may be at risk of developing visual problems and refer them to a specialist who can diagnose and treat them.

As a first step, we need an easy-to-use, portable and affordable device to identify these children with visual disorders as soon as possible, ideally when they are only babies. This is the main goal of our TrackAI project.

Track AI

We know that children with visual difficulties look in a different way than children with good vision. We can sit a child in front of a screen and present him or her with different visual stimuli.

A child with normal vision looks at the stimuli like this:

The red dots correspond to his eyes moving over the screen. However, the gaze of a child with visual disorders looks completely different, as you can see here:

So we need to perform an automatic visual assessment of children at very young ages, who cannot communicate with us. To do that, we are developing a system composed of a Huawei P30 smartphone and a device called DIVE – our Device for an Integral Visual Examination.

The doctor will use the P30 to control the whole system, while data collection will be done through the DIVE device.

DIVE + HiAI

The DIVE device presents different stimuli on a high-resolution screen or a Huawei MateBook and captures the gaze of the patient using eye tracking technology.

But analysing gaze data is a challenging problem, and that’s where AI comes in: Starting from the DIVE platform and its gaze analysis software, we’re now incorporating artificial intelligence powered by the Huawei open platform HiAI 2.0.

The HiAI 2.0 open platform has enabled us to access the capabilities of the Kirin chipset and has helped our engineers develop our AI model efficiently.

We are now able to find gaze patterns in the way children look at visual stimuli, and relate them to visual development and to certain specific disorders.

A Close Eye on the Data

Running our AI model on the P30 enables the accurate and reliable detection of visual impairments in real time. This combination unleashes the potential not only to measure visual function, but also to estimate the probability for the patient to have a certain pathology.

Developing a system based on AI requires information from thousands of patients.

During this unique collaboration between Huawei and our DIVE Medical startup, ophthalmologists are collecting data from children with normal visual development and children with a wide range of visual disorders in five different countries.

The information gathered during this gaze data collection phase is being used to train the neural network. We have already examined 1500 children of different ethnicities and ages and a first prototype of the neural network is undergoing testing.

We are also building the final system, which will use DIVE and the P30 together to generate a diagnosis probability for the patient, and we also have a first prototype in testing phase.

The unprecedented capabilities of Huawei’s P30 smartphone allow DIVE technology to reach out and be useful across the world, from an ophthalmologist in a hospital in UK to a family doctor in a remote area in Mexico.

The UN’s Sustainable Development Goals include a target to achieve universal health coverage by 2030, putting a spotlight on the importance of improving eye care as it’s so vital for many aspects of well-being.

Another goal is to ensure healthy lives and promote well-being for everyone at any age by providing more efficient funding for health systems.

A portable and affordable device like DIVE run by non-trained people from the community could be the perfect solution for ensuring universal and sustainable eye care for children.

Now that we are close to having a functioning and validated system that is able to run accurate visual screening in babies as young as six months old, we are thinking of the next steps. These steps involve collaborating with different NGOs and organizations to bring this technology to many places excluded from healthcare. We aim to start some of these projects in 2020.

We are facing a high-impact challenge and the result of this project will indeed change the lives of children all around the world. As researchers, we need to partner with tech companies to develop our projects. Huawei has enabled us to make a global impact and bring DIVE to every corner of the world.

Track AI is part of Huawei’s TECH4ALL initiative. Click the links to read more about TECH4ALL and to visit the Huawei TECH4ALL minisite.

_________

This is a guest post by Dr. Victoria Puyeo

Victoria is a Pediatric Ophthalmologist at Miguel Servet University Hospital and an Associate Professor at the University of Zaragoza.

She is the Principal Investigator of the Vision, Imagen, and Neurodevelopment Research Group, from the Aragon Institute of Health Research Foundation (IIS Aragon). Victoria co-founded DIVE Medical, a solution that emerged from the difficulties experienced every day by ophthalmologists and paediatricians when examining the visual function of young patients.

---

Disclaimer: Any views and/or opinions expressed in this post by individual authors or contributors are their personal views and/or opinions and do not necessarily reflect the views and/or opinions of Huawei Technologies.