Biometric facial recognition tools for improved infection control in COVID-19 response

In the absence of a vaccine or robust treatment protocol, effective case management is one of the best tools to combat COVID-19\. Case management is the process of monitoring a patient's care, including screening, testing, treatment, and discharge. Reliable case management helps enable better patient outcomes, stop the spread of the virus, and ensure that accurate data to combat the disease are available to programme staff, policy makers, and researchers.

In high-income countries, formal IDs underpin individuals' health records, providing the basis for case management. However, according to the World Bank (2019), over 1.1 billion people worldwide (1 in 7 people globally) lack formal identification. Those living without identity are predominantly in lower-income countries, where the lack of identity combined with weaker health infrastructure makes case management and suppression of COVID-19 incredibly challenging.

In settings where formal IDs are not prevalent, [patient ID cards and barcodes][0] are commonly used for case management. However, physical ID cards can be a source of viral transmission and other methods of identifying patients are unreliable. [Names are not sufficiently unique][1] and are easily misspelled; [dates of birth may not be known][2]; and data entry errors are commonplace.

Although countries with poor health infrastructure are rapidly establishing makeshift testing, quarantine, isolation and treatment centres, without unique identifiers, managing confirmed cases remains a critical challenge. The lack of unique identifiers makes it difficult to: (1) [conduct contact tracing][3], (2) link individuals consistently to their test and treatment records, and (3) prevent duplicate records. Without a unique identifier, valuable time and resources may be wasted due to inaccurate patient records and unreliable data, compromising quality of care and costing lives. In addition, corruption enabled by weak monitoring mechanisms wastes billions every year. It is estimated that up to 29% of global development assistance is lost to corruption (Center for Global Development 2017). When total yearly investment in development assistance is as high as $153B (OECD 2018), this means up to $44B is lost every year.

Simprints is a tech company from the University of Cambridge that pioneers innovative solutions to verify coverage and impact, by building and deploying biometric identification systems in the hardest environments in the world. During the past few years, Simprints successfully developed a variety of biometric identification solutions for different applications, including an identification system based on fingerprint scanning that has been commercialised across 12 countries, and an image based biometric identification system (facial recognition) which is now at the start of commercialisation.

Currently the challenge for Simprints is to further facilitate the implementation and use of identification services at scale for COVID-19 response in developing countries, with systems that solve the current needs in this market. In particular, in this project Simprints proposes to research, develop, and validate :

* periocular facial recognition (which will work effectively when the nose and mouth are covered by a mask)
* Liveness detection (to prevent the fraudulent enrolment of 'ghost' beneficiaries from photographs)

[0]: https://journals.lww.com/jaids/fulltext/2009/11011/_Talkin__About_a_Revolution___How_Electronic.16.aspx
[1]: http://ebooks.iospress.nl/publication/13402
[2]: https://www.sciencedirect.com/science/article/pii/S138650560000068X?via%3Dihub
[3]: https://linkinghub.elsevier.com/retrieve/pii/S1201971215002593