New Technology Allows Identification Through a Mask

By HospiMedica International staff writers
Posted on 07 Apr 2020

Following the worldwide outbreak of coronavirus, an improved facial recognition solution can correctly identify people who wear facial masks.

The Herta (Barcelona, Spain) facial recognition algorithms are based on deep learning (DL) technology, providing very high identification rates, especially in verification tasks that involve partial occlusions issues in crowded environments and when automatic passenger identification systems are used, such as border passport control, transportation, healthcare, entertainment venues, and sports stadiums. Thanks to the new algorithm, it will not be necessary for the person to remove the mask, avoiding possible contagion or long waiting times.

Image: AI enhance facial recognition algorithms can identify people wearing masks (Photo courtesy of Herta)

The DL algorithms involved apply banks of convolutional and non-linear filters over an original image. Each layer of application processes the image and extracts higher-order information. After many layers of these filter banks (typically between tens and hundreds), the faces are encoded directly into small templates which are very fast to compare. All face alignment, frontalization, visual features, localization of regions of interest, etc., are done internally by the algorithm itself. It is worth noting that the most differential part of the human face is in the eye region.

“The company had been working on the issue of partial occlusions for some time and, following the worldwide outbreak of CoVid19, development has been accelerated to launch a version of the software that helps provide an accurate identification under these conditions,” said the company in a statement. “Herta expects that the impact of this new technology in the market will be very important worldwide and that it will be used massively in environments such as transportation, health, government, events, or in the gaming sector.”

Facial recognition systems use biometrics to map facial features from a photograph or video. The geometry of the face is then analyzed, with key factors including interpapillary distance and the distance from forehead to chin. In all, there are over 65 quantifiable features that can be used to identify a face, generating a unique facial signature.

Related Links:
Herta