The facial recognition algorithm just has to focus on the other parts of the face!

The use of facial recognition has been impacted by the wearing of face masks, and the technology behind this biometric authentication method has had to be adapted for this global phenomenon.

Recently there has been some success in recognizing faces even with various types of masks in place. Face recognition operates by extracting feature points such as the position, shape, and size of a subject’s eyes, nose, and mouth, and matching and identifying detected faces. In the case of faces that are covered by a mask, the software focuses on areas surrounding an individual’s eyes that are not covered.

mask certification photo

Once the software detects the presence of a masked face, the most appropriate facial recognition algorithm is used to extract and check the characteristics of each individual using the exposed parts of the face, achieving a high recognition accuracy of over 99.9% in 1:1 verification even when masks of various colors and patterns were worn.

NEC Corporation, the company behind the ‘Enhanced Video Analytics’ technology plans to begin worldwide sales of products incorporating the software by November. The domestic market will offer a version that utilizes a variety of face and other biometric information to provide multi-modal biometrics solutions.

These products and services are the part of the company’s platform for integrating biometrics, video analysis, AI, and security technologies while providing comprehensive services from applications to networks.

From end September 2020, their Japan headquarters will showcase a system that links face recognition and thermal cameras at walkthrough gates that are used for monitoring entrances and exits.