Predikta Station

The COVID-19 pandemics revealed that the world was not prepared to deal with global pandemics. Not only government failed establishing assertive policies to contain the outbreaks, but also the lack of mass screening technologies led to further aggravation of contamination scenario, exposing the population to an unpredictable virus. In this context, we developed an automatized Totem, based on Infrared thermal scanning that provides, in about 30 seconds, an accurate screening test for COVID-19. Using infrared facial recognition and machine learning modeling we were able to provide a fast and accessible turnaround for COVID-19 screening. Furthermore, we developed a health pass app, that allows tracking the user history and suggesting actions in the case of disease suggestion. The social implications of our system go beyond the current pandemics, once this technology may be used for several other diseases, providing a safer environment for users through an accessible and affordable e-health tool.

It is unequivocally true that, during global pandemics, providing means to control infection spreading is fundamental to fight back the outbreak. As the mass usage of molecular tests is limited to their price and has relatively slow turnaround, we need efficient and quick screening procedures to back up the health system, providing ways to spent their limited resources more rationally. Thus, we look back to the reality of most countries during the COVID-19 pandemics: (1) not enough resources to test everyone, (2) vaccination was still in early development and availability issues were certain and (3) government and population were reluctant to adopt restrictive measures due to economic consequences. It is certainly a chaotic scenario to control a pandemic. 

In the other front, we saw an increase in the usage of infrared sensors to assess fever, albeit inefficient due to poor implementation and equipment limitations. In this context, we looked at how we could use the clinical value of infrared thermal technology to create an accessible screening tool that could test as much people as possible in a short period of time, using affordable technology to provide a safer environment for users and help to ease the impact of the pandemics.

Our solution is a Totem, based on Predikta Technology, coupled with a thermal camera able to analyze infrared images of the face and detect significant alterations in the thermal profile, suggestive of COVID-19. Differently from other IRT instruments., that usually only measures fever, our application analyzes thermal profiles using several face landmarks. It is equipped with machine learning algorithms capable of recognize thermal profiles related to covid-19 infection with good accuracy and face recognition software for infrared images.

The totem is versatile, easily updatable and affordable, each unit can test hundreds of people by day. It offers an user interface that can be accessed locally in the unit or through mobile. All data is integrated in the cloud infrastructure, but the tests can be done without internet. We also count with a mobile app, where the user can input personal data related to their health status and get suggestions on how to proceed in case of positive test result. The app also generates a health pass, base on our test output, making it easier to track and control infection hotspots. These are 3 pillars of Predikta Technology: Infrared scanning unit – machine learning modeling – mobile and user integration.

When we first idealized our infrared thermal scanning, we centered it on the accessibility, affordability and accuracy.  We intend that our technology reaches as much people as possible. There are virtually no barriers that prevent a person to use our test, anyone who is able to stand in front of the equipment is eligible to be tested. From children in a school, students in a university, patients in hospitals, consumers in a shopping center or supermarket to people leaving in remote and/or poor communities. One of the driven cores of Predikta technology is the accessibility.

The Predikta mobile app also plays an important role in our system, once it allows the user to input personal health information that they feel comfortable to share with us, then we can track their health status and suggest actions to treat any disorder detected early. Based on our scan output, a health pass can or not be provided to the user, and our system will suggest measures to be taken in case of covid-19 suggestion (e.g. directing him to the nearest hospital or laboratory to perform a molecular test or simply suggesting him to stay home and to repeat the test next day). We believe that our system works as a double layer protection, at same time protecting the user and providing a safer environment for everybody else in the place that our totem is installed. Our goal, in the middle and long term, is to provide a powerful and accessible instrument of preventive telemedicine applying the Predikta Technology to other clinical areas where thermography can be meaningful (e.g.  flu syndrome, ophthalmology, orthopedics, cancer, metabolism disorders, etc.), we strongly believe in the positive impact of our technology in the social context.

Our application is powered by Predikta Technology and, in this phase of
implementation, it consists in an automatized screening device to detect early signs of COVID-19 using infrared thermal image of the face and machine learning modelling. Then, its main goal is to track infected people in the early stages of infection, not allowing them to spread the disease and assuring them the best treatment as possible. In this context, our application falls within the scope of the challenge in the category warning predictive systems, aiming to slow down the infection rate and strengthen COVID-19 surveillance.

In the present time, we are at the beta version of our device which are installed and collecting data for validation in several places, thus we considered it in the initial phases of pilot implementation. We firmed partnership with at least 4 hospitals and laboratories to collect data to improve machine learning modelling and general functionalities of our devices. At the moment, we have collected data of more than 15.000 patients and the product is on the verge of being validated by the Brazilian regulatory health department. We count with a cloud infrastructure that is still being improved and automatized to provide more reliable data for modelling. We still have a lot of work to do to fine tune our product, however, we already passed the prototype phase. The company structure is well defined, as well the business model.

The solution is innovative due to the ability to observe in an image of the person's body the physiology and correlate with various healthy physiological or pathological processes using artificial intelligence to predict, the thermal image has always been much more than just temperature and should be defined as balance of several physiological processes obtained at the time of image capture. In addition, it allows for the assessment and delivery of diagnostic support, allowing it to be almost instantaneous, non-invasive and without any harm to health. In addition, it will allow us to assess symptoms and inflammatory processes far beyond fever and actually use these infrared sensors to take the most potential out of them in the observation of thermodynamics and human or animal physiology.