I do not get tired, nor am I going to, to continue counting the solutions and applications that use artificial intelligence (AI), especially those that mean a clear benefit for people. And in this case even more, since children with visual problems will have a greater opportunity with Track AI to improve their vision by detecting their anomaly in advance.
The World Health Organization estimates that there are 19 million children in the world with vision problems, of which 70 to 80 percent can be prevented or cured. In most cases, these children remain undiagnosed for years, which can have more serious consequences for their vision, their general development, as well as their educational and social opportunities. The less severe but easily preventable types of visual impairment currently affect 12.8 million children.
Track AI is the name of the application miracle? that uses Huawei’s artificial intelligence, and that it has developed together with IIS Aragón and DIVE Medical. The system, based as we have said on the Chinese manufacturer’s AI, has been designed to detect and diagnose early signs of visual impairment in children, and the purpose of DIVE, according to Victoria Pueyo, pediatric ophthalmologist of this organization, “is to implement this technology worldwide and, with a multi-ethnic approach, be able to attend to all types of visual impairment ”.
To detect visual impairments, DIVE (Devices for Comprehensive Visual Examination) software is used on a laptop (Huawei Matebook E) to monitor and track the gaze of each eye while the patient is searching for stimuli designed to test different aspects of function. visual. The data collected from the gaze pattern of patients is processed on the Huawei P30 smartphone using the HUAWEI HiAI algorithm, in order to identify possible indicators of visual problems.
In the past, smartphones were not powerful enough to process complex AI-based algorithms, and data had to be sent to the cloud for processing. Communication to and from the cloud makes AI processing slow and useless without a network. The new Track AI technology combines the best in ‘machine learning’ and ‘on-device’ artificial intelligence processing technology with concrete research and knowledge from IIS Aragon and DIVE Medical.
How does Track AI work? Children with visual impairment have different gaze patterns compared to children with normal vision. DIVE can accurately collect gaze position data during a scientifically designed visual stimulus. Interpreting these data can be challenging for unskilled healthcare workers. However, using AI, computers can be trained to identify abnormal gaze patterns for easier screening and assisted diagnostics. In this way, it is possible to speed up the process, from diagnosis to problem management.
Track AI builds on the capabilities of Google’s TensorFlow, an open source platform, and Huawei HiAi, to create a machine learning system on a smartphone that uses Kirin 980, the latest chip from the Chinese manufacturer that incorporates two dedicated processing units. neural (dual NPU) that are compatible with the artificial intelligence of the device, which means that the AI functions are processed in the smart mobile itself, which significantly accelerates the activity in addition to preserving the privacy of the user. This dual neural processing unit also helps make the smartphone’s AI features smarter over time, as well as increasing daily performance and efficiency.
The TrackAI device on the smartphone is portable, does not require WI-FI and receives real-time data to speed up diagnoses, so it will soon be usable around the world, from an ophthalmologist in a UK hospital to a doctor from a village in a remote part of Africa. However, developing an AI-based system first requires the collection of data from thousands of children, and research centers are currently collecting data in China, Mexico, the United Arab Emirates, Spain and the United Kingdom. Once a large amount of data has been entered, the neural networks will be trained and a prototype device will be tested later this year. The goal is to implement it in 2020, which will help healthcare professionals diagnose and treat eye conditions at an early stage. Let’s see it!
