Iris biometrics adoption could rise on touchless transactions and masks, researchers say
Biometrics research spurred by COVID-19 could lead to wider adoption of iris and periocular recognition, according to a new report from a team of European researchers.
The team of ten researchers affiliated with Hochschule Ansbach, the Biometrics and Internet Security Research Group at Hochschule Darmstadt, and Fraunhofer Institute for Computer Graphics Research IDG in Germany and France’s EURECOM surveyed the research efforts into solving challenges to biometric authentication and security posed by the pandemic.
The pandemic has had a high impact on facial recognition and touch-based fingerprint systems, according to the paper ‘Biometrics in the Era of COVID-19: Challenges and Opportunities,’ which has not been peer reviewed. Voice biometrics have been moderately impacted, while the impact of COVID-19 on other modalities has been low.
More stringent hand hygiene practices and the ubiquitous use of facial recognition have presented particular challenges for biometrics systems in use, the researchers point out, and pushed research into different directions, such as non-constrained iris or periocular recognition, touchless fingerprint and vein authentication, and biometrics tracking for illness detection.
The study refers to the findings of NIST and the DHS Biometric Rally on masks and facial recognition, and finds the research so far raises questions about additional occlusions like glasses, identification (or one-to-many searches), demographic differentials and other issues, without providing many answers thus far.
In voice biometrics, masks can degrade the reliability of standard recognition systems, but also interfere with presentation attack detection (PAD) systems by attenuating the ‘POP’ noise commonly caused by proximity to a microphone. These kinds of affects can be compensated for over time, however, in part by training algorithms with voices from people wearing masks.
Iris recognition research has increasingly focused on less constrained image-capturing approaches, sometimes with image resolution-improvement techniques. The researchers find these same techniques can also be applied to the periocular region, and that whether applies to irides or the region around the eye, are likely to increase in adoption.
In addition to various modalities, the report also considers remote and mobile biometric recognition, such as for remote education or work-from-home. Touchless hand-based systems could become a viable alternative to touch-based systems, the researchers note, but continue to be an active field of research with varying challenges related to a lack of control over hand position, background and illumination.
In noting several research projects exploring the detection of the COVID-19 virus through biometrics and other digital sensing technologies, the researchers observe that the European Union Aviation Safety Agency (EASA) says between 1 percent and 20 percent of people with fevers will be missed by thermal scanners.
The report concludes by calling for more research to bring the security level of COVID-era biometric systems up to the level of assurance provided by biometrics in other, unmasked circumstances.
accuracy | biometric identification | biometrics | biometrics research | contactless biometrics | facial recognition | iris biometrics | iris recognition | periocular biometrics | voice biometrics