Researchers use facial recognition to identify patients in Kenya, develop custom 3D-printed masks

An open source biometric facial recognition system described in a study by Associate Professor of Biomedical Informatics and Medicine at Vanderbilt University Martin Were, MD, MS, and colleagues could accurately identify patients in under-developed countries where patient misidentification is very common, the university announced.

The team tested a biometric system dubbed OpenFace which was embedded in the national electronic health record system in Western Kenya. According to their study published in the International Journal of Medical Informatics, the system was used to enroll and match 103 patients at clinics across the region. Patients gave their consent to sit and have some 13 pictures taken to build a training set. A second round of photographs followed and the system tried to match these against the initial set taken.

Patients often forget to bring their IDs, making it very common to not properly identify people who need medical services. The algorithm in the information system uses demographic data to match these people with their medical records, but in 30 percent to 50 percent of cases it does not provide accurate results.

The technology was evaluated based on Sensitivity, False Acceptance Rate (FAR), False Rejection Rate (FRR), Failure to Capture Rate (FTC) and Failure to Enroll Rate (FTE). The test revealed a sensitivity of 99.0 percent, FAR <1 percent, FRR 0.00, FTC 0.00 and FTE 0.00. Performance was not affected if people wore glasses. The system successfully identified 102 out of 103 patients. The one misidentified patient was later identified during a second try.

“We were very pleased by the accuracy of this open-source system but are aware that such systems have to be implemented with great attention to patient confidentiality,” said in a prepared statement Were, associate professor of Biomedical Informatics at Vanderbilt University Medical Center and a member of the Vanderbilt Institute for Global Health.

“For low- and middle-income countries, where patient identification may involve various additional challenges, automated facial recognition modules attached to electronic health records could provide a vast improvement at low cost. This study adds to the ongoing conversation about the costs and benefits of these systems in health care.”

Researchers Sight Ampamya of Moi University in Kenya and John Kitayimbwa of Uganda Christian University also participated in the study, financed by the Norwegian Program for Capacity Development in Higher Education and Research for Development.

University of Connecticut enters licensing deal to develop 3D-printed mask frames

Under the brand Secure Fit, recently founded Connecticut Biotech has started manufacturing and distributing 3D-printed mask frames to help stop the spread of COVID-19, following a licensing deal with the University of Connecticut (UConn), the institution announced.

“This is an important technology that can help a lot of people by providing a specific way to make regular surgical masks more protective,” said in a prepared statement Dr. Cato T. Laurencin, CEO of the Connecticut Convergence Institute for Translation in Regenerative Engineering. “It’s wonderful to see technology that started here in the state of Connecticut being developed by a Connecticut company.”

Together with a group of scientists, he designed mask add-ons to make them more effective through a filtration system and positioning.

“The novel aspect is using 3D facial recognition technology on two-dimensional images to decipher the topography of the face and create a 3D mask frame,” said in a prepared statement Amit Kumar, director of licensing and business development in UConn’s Office of the Vice President for Research.

“We are excited about this public-private partnership with UConn and the opportunity to use and build upon our vast experience to rapidly commercialize this much-needed device,” said in a prepared statement Connecticut Biotech CEO Donald J. Vacarro. “We applaud Dr. Cato Laurencin’s team for developing a custom mask frame that secures filtration and positioning of most mask types.”

To place an order, those interested can go to the company’s website and submit a photo from the front and one from the side of their face.

UConn submitted the patent for the concept in April, as part of its efforts to contribute to COVID-19 relief efforts.

Article Topics

biometrics  |  biometrics research  |  facial recognition  |  healthcare  |  patient identification  |  research and development